· adjuvant chemotherapy: which patient? what regimen? natalie turner, mbbs, laura biganzoli, md,...

BREAST CANCER

Adjuvant Chemotherapy of BreastCancer: Improving PatientSelection and UnderstandingBenefits and Risks

CHAIRAngelo Di Leo, MDIstituto Toscano TumoriPrato, Italy

SPEAKERSJoseph Sparano, MDAlbert Einstein College of Medicine, Montefiore Medical CenterBronx, NY

Erica Mayer, MD, MPHDana-Farber Cancer InstituteBoston, MA

Adjuvant Chemotherapy: Which Patient? What Regimen?Natalie Turner, MBBS, Laura Biganzoli, MD, PhD, Luca Malorni, MD, PhD, Ilenia Migliaccio, MD, PhD,Erica Moretti, MD, Marta Pestrin, MD, PhD, Giuseppina Sanna, MD, Olimpia Siclari, MD, and Angelo Di Leo, MD, PhD

OVERVIEW

In the past, treatment decisions regarding adjuvant chemotherapy in early breast cancer (EBC) were made solely based on clinico-pathologic factors. However, with increased awareness of the importance of underlying tumor biology, we are now able to use genomicanalyses to determine molecular breast cancer subtype and thus identify patients with tumors that are chemotherapy resistant andunlikely to benefit from the addition of chemotherapy. Although genomics has allowed some patients to avoid chemotherapy—specifically those with luminal A–like breast cancer—these assays do not indicate which regimen is most appropriate. For this,consideration must be given to the combination of underlying tumor biology, tumor stage, and patient characteristics, such as age andtolerability of side effects.

Adjuvant chemotherapy improves outcomes in EBC.However, as our understanding of breast cancer biology

increases, and the choice of chemotherapy regimen broad-ens, two major questions are raised concerning adjuvantchemotherapy treatment decisions, namely: should my pa-tient be receive chemotherapy? And if so, which is the bestregimen?

WHICH PATIENT?Decisions regarding chemotherapy have previously beenmade based solely on clinicopathologic factors, such as tu-mor stage and node status. However, in recent years there isincreasing acknowledgment of the importance of tumor bi-ology in terms of both prognosis and variable response tochemotherapy.1 In particular, as was initially demonstratedin pioneering work from the International Breast CancerStudy Group (IBCSG) and Cancer and Leukemia Group B(CALGB), some hormone receptor–positive (HR�) breastcancers do not benefıt from adjuvant chemotherapy, are rel-atively chemotherapy resistant, and have an otherwise excel-lent prognosis with endocrine therapy alone.2,3

The Role of Genomic AnalysisToday, rather than considering all HR� tumors as chemo-therapy resistant, the concept of high- and low-risk HR�disease has emerged, primarily because of genomic pro-fıling, with various gene signatures available. The 21-generecurrence score (RS)—based on expression of 16 cancer-related genes and fıve reference genes—was shown in the

initial validation study, using data from the randomizedphase III National Surgical Adjuvant Breast and Bowel Proj-ect (NSABP) B-14 trial of tamoxifen compared with no adju-vant treatment in node-negative, HR� breast cancer toeffectively separate patients with node-negative, HR� EBCinto low, intermediate, and high risk of disease recurrence.For patients with high-risk disease, 10-year disease-free sur-vival (DFS) was signifıcantly less than that in the low-riskgroup (69.5% vs. 93.2% for high vs. low risk respectively, p�0.001).4 The RS, now marketed as Oncotype DX, has alsobeen validated as a risk prediction tool in HR�, node-negative EBC treated with adjuvant cyclophosphamide/methotrexate/fluorouracil (CMF); HR�, node-positive EBCtreated with cyclophosphamide/doxorubicin/fluorouracil(CAF); and HR�, node-positive or node-negative EBC inpostmenopausal women treated with either tamoxifen oraromatase inhibition.5,6,7Other gene signatures—such as the 70-gene signature

(marketed as MammaPrint), PAM50 intrinsic subtype clas-sifıer and the Gene Expression Grade Index (GGI)—havealso been retrospectively validated as prognostic markers inEBC, although only the RS utilized randomized data for itsvalidation.8,9The merit of genomics is that it allows us to identify a sub-

set of breast cancers, namely luminal A–like tumors, whereproportional benefıt from adjuvant chemotherapy is mini-mal independent of disease stage. As such, we recommendavoidance of adjuvant chemotherapy for node-negative, lu-minal A–like EBC. In the setting of nodal involvement, al-though the proportional benefıt of chemotherapy is low, the

From the Sandro Pitigliani Medical Oncology Department, Hospital of Prato, Istituto Toscano Tumori, Prato, Italy; Translational Research Unit, Department of Medical Oncology, Hospital of Prato,Istituto Toscano Tumori, Prato, Italy.

Authors’ disclosures of potential conflicts of interest are found at the end of this article.

Corresponding author: Angelo Di Leo, PhD, Sandro Pitigliani Medical Oncology Department, Piazza dell’Ospedale 2, Prato, Italy, 59100; email: [email protected].

© 2013 by American Society of Clinical Oncology.

ADJUVANT CHEMOTHERAPY AND BREAST CANCER

asco.org/edbook | 2013 ASCO EDUCATIONAL BOOK 3

absolute benefıt is uncertain, hence the addition of chemo-therapy to endocrine therapy may be discussed with the pa-tient in this setting.

Pending Issues for Selecting Which Patients ShouldReceive Adjuvant ChemotherapyImmunohistochemical (IHC) definition of luminal breast can-cers. The use of molecular classifıers in routine clinical prac-tice may be limited by their relative cost and complexity.Alternatively, luminal A and B tumors may be defıned byIHC determination of estrogen receptor (ER), progesteronereceptor (PgR), HER2, and Ki-67. Various criteria have beensuggested, though, as yet, there is no consensus on the bestdefınition to employ.10-12 In one approach, distinction be-tween luminal A andHER2-negative-luminal B tumors relieson the semi-quantitative expression of Ki-67, with a putativecut point of 14%.11,13 This 14% cutoff, fırst identifıed byCheang et al, was able to distinguishmolecular-defıned lumi-nal A and luminal B subtypes with a sensitivity and a speci-fıcity of 77% and 78%, respectively.11 Furthermore, in a largecohort of patients with EBC, IHC-defıned luminal B tumors(either HER2 positive or Ki-67 � 14%) were associated witha worse outcome compared with IHC-defıned luminal A tu-mors.11 Yet the accuracy of this method compared with mo-lecular subtype determination by genomics has been shownto be less than 80%, indicating that further refınements andvalidation are needed.We recommend that, where genomic assays are available

without additional costs to patients, this approach be used todetermine tumor subtype. However, where issues of costand/or access to genomic assays exist, IHC can be utilized asthe fırst step in breast cancer subtype determination, with agenomic assay reserved cases that cannot be clearly catego-rized as luminal A or B using IHC. It is critical that, if relyingon IHCassessment, quality assurance checks aremade in col-

laborationwith your pathologist, to avoid commonproblemsrelated to IHC assessment, such as inaccurate fıxation time,and variable quality of antibodies, which can lead to false-positive and -negative results.

Late relapses in luminal breast cancers. The majority of lu-minal A tumors have an excellent prognosis with endocrinetherapy alone; however, luminal cancers, particularly lumi-nal A, harbor risk of late relapse after disease-free intervalsof fıve to 15 or more years. Luminal A and B breast cancersare extremely molecularly heterogeneous. A recent inte-grated analysis of copy number changes and gene expressionprofıling of nearly 2,000 EBCs dissected luminal cancersin different prognostic subgroups, some with particularlyunfavorable outcomes, such as the IntClust2 group.14 Themolecular determinants of late relapses are still under inves-tigation. In the future, should it be possible to identify pa-tients at higher risk of late relapse, it is anticipated that thesepatients would benefıt from extended endocrine therapy,rather than the addition of adjuvant chemotherapy.

Defining treatment for patients with intermediate-risk dis-ease. The approach to luminal A tumors with nodal involve-ment is uncertain. Classically nodal involvement implied theneed for adjuvant chemotherapy; however, the generallygood outcomes with luminal A disease with endocrine ther-apy alone, combined with relative chemoresistance, hasraised the possibility that tumors should preferentially betreated based on biology, rather than stage. Just as a tumorwith aggressive biologic characteristics will often be treatedwith chemotherapy despite small size or no nodal involve-ment, it may be that chemotherapy provides no additionalbenefıt over endocrine therapy in less-aggressive breast can-cers. This clinical question is currently being investigated inthe RxPONDER trial (NCT01272037), in which patientswith low-risk ER� breast cancer, defıned by RS of 25 or less,but with involvement of one to three axillary nodes are ran-domly assigned to endocrine therapy alone or chemotherapyplus endocrine therapy. Results from this trial should help toanswer whether favorable tumor biology is more importantthanunfavorable tumor stagewhenmaking adjuvant therapydecisions.Two current prospective trials should help refıne the

approach to these borderline or intermediate-risk patients.The TAILORx trial (NCT00310180) randomly assigned pa-tients with intermediate risk on RS to receive treatment withendocrine therapy alone or endocrine therapy and chemo-therapy, while the MINDACT trial (NCT00433589) ran-domly assigned patients with EBC with 3 or fewer positivenodes and discordant recurrence risk assessment based onthe 70-gene signature or clinical assessment to chemotherapyor not using the result of one or the other. Until results fromthese trials become available, we recommend treatment ofintermediate-risk disease based on standard clinicopatho-logic risk factors. Chemotherapy could be avoided in patientswith small node-negative tumors, while features such asnodal involvement, large T size, histology, high prolifera-

KEY POINTS

� Identification of breast cancer molecular subtypes and useof genomics has revolutionized the approach to theselection of patients requiring adjuvant chemotherapy.

� A subgroup of hormone receptor–positive breast cancers(luminal A–like tumors) derive little, if any, benefit fromthe addition of chemotherapy to endocrine therapy and cantherefore be spared chemotherapy-associated toxicities.

� Pending issues regarding the selection of patients foradjuvant chemotherapy include the need for an accurateimmunohistochemical (IHC) surrogate for defining molecularsubtypes, how to best approach patients with intermediaterelapse risk on current assays, and the need to ideallyincorporate peripheral markers of micrometastases intorelapse risk assessment.

� Genomics cannot provide guidance on the selection ofspecific adjuvant chemotherapy regimens, and thus thechoice of what regimen to utilize should be made based onboth tumor biology and patient characteristics.

DI LEO ET AL

4 2013 ASCO EDUCATIONAL BOOK | asco.org/edbook

tion, or lymphovascular invasion would favor chemotherapy.Additional patient factors to be considered include age at di-agnosis, performance status, and presence or absence of co-morbidities.

Not all patients with high-risk disease benefit from chemo-therapy. From retrospective data from the NSABP B-20 trialused in the validation of the RS, a clear advantage for chemo-therapy in patients with disease classifıed as high risk for re-currence was seen. However, after 10 years’ median follow-up, approximately 60% of patients in this high-risk categorywho only received tamoxifen remained disease free.5 Thus, asignifıcant proportion of patients with high-risk disease donot relapse despite not receiving chemotherapy. Current riskprediction tools are therefore over-sensitive, and refınementsthat allow more accurate stratifıcation of high-risk diseasemay allow some patients to avoid chemotherapy and the as-sociated toxicities.However, until such refınements aremadeor new tools are defıned, we recommend treatment of high-risk disease with adjuvant chemotherapy, acknowledgingthat some patients, who are currently not identifıable, are notreceiving benefıt.

Considerations for the FutureAll currently utilized prognostic factors for EBC are derivedfrom the excised primary tumor. Yet there is now increasingawareness of the role of peripheral factors—such as markersof micrometastatic disease—in predicting tumor behavior,disease course, and patient outcomes. Assessment of theseperipheral factors should enable a more holistic approach todetermining the risk of relapse in EBC.Detection of micrometastases may be possible through

analysis of disseminated tumors cells (DTC; e.g., on bonemarrow assessment), circulating tumor cells (CTC), cell-freeDNA (cf-DNA), plasma microRNAs, or metabolomics.15-18A large number of studies showed that the presence of DTCin bone marrow has prognostic effect for patients with EBC,suggesting that DTC could be used as a marker for diseaserecurrence. Similarly, CTC assessmentmay predict poor sur-vival after neoadjuvant treatment or primary breast cancersurgery, whereas cf-DNA concentration and miRNA havebeen associated with EBC disease relapse.15-18 These noveltechniques may thus have utility in predicting disease recur-rence, improving early detection of relapse, and enhancingpatient outcomes; although in each case, further validation inlarger cohorts is needed before routine use in the EBC setting.In the future, these tools could be combined with tumor bi-ologic and genomic characteristics to provide a more com-plete picture of a patient’s relapse risk and thus help informtreatment decisions regarding the need for adjuvant chemo-therapy.

WHAT REGIMEN?Gene expression tests, such as the RS, are extremely promis-ing in identifyingwho should benefıt fromchemotherapy but

not informative on the type of chemotherapy to be given. Forthis, consideration must be given to the combination of tu-mor subtype, tumor stage, and patient characteristics.

Endocrine-Responsive Breast CancerLuminal A. As discussed above, the general recommendationis avoidance of chemotherapy, particularly in node-negativedisease and instead to treat solely with endocrine therapy.One of the challenges in the implementation of genomic as-says into adjuvant chemotherapy treatment decisions relatesto the fact that, regardless of the specifıc assay employed,gene expression is initially reported as a continuous variable,with arbitrary cut-off points, then assigned to separate thelower-risk tumors from the higher-risk tumors. These cut-offpoints have been carefully calculated to maximize the accu-racy of the assay to discriminate between better and pooreroutcomes. Nonetheless, use of arbitrary cut-off points whenevaluating a continuous variable invariably leads to some tu-mors residing close to that cut-off point, where interpreta-tion of relapse risk for these borderline cases is diffıcult.

Luminal B. As luminal B tumors are characterized by highproliferation rates, increased aggressiveness, and increasedrisk of relapse, chemotherapy in addition to endocrine ther-apy is recommended. However, there are no data to deter-mine whether chemotherapy should be a standard adjuvantregimen as would be used for HR- disease, or if less intensivechemotherapy is as effective. Choice of approach shouldtherefore bemade based on clinical assessment of relapse risk.For luminal B tumors with additional high-risk fea-

tures (pT � 1 and/or pN � 0), one may reasonably optfor sequential anthracycline/taxane chemotherapy. In ab-sence of higher-risk features, less intense chemotherapy—such as docetaxel/cyclophosphamide (TC), doxorubicin/cyclophosphamide (AC), or CMF—may be considered,sparing toxicity without compromising effıcacy.19-21

HER2� Breast CancerThe standard adjuvant treatment of HER2� EBC is trastu-zumab plus chemotherapy, preferably using an anthracycline-based regimen. However, only a small percentage of patientsreceiving adjuvant anthracyclines actually benefıt, yet all areexposed to the associated toxicities. In particular, the risk ofcardiotoxicity is increased with sequential trastuzumab ther-apy.22 Thus the major consideration in HER2� EBC iswhether or not to include an anthracycline in the treatmentregimen.The observed increased sensitivity of HER2� tumors to

anthracyclines may be related to coexpression of TOP2A. Inthe BCIRG-006 trial, both trastuzumab-containing arms(AC plus docetaxel/trastuzumab [ACTH] and doxcetaxel/carboplatin/trastuzumab [TCH]) were superior to the non-trastuzumab arm of AC plus docetaxel (ACT), with the studynot powered to detect a difference between ACTH andTCH.22 Treatment response based on TOP2A status wasalso assessed, with benefıt from anthracyclines evident inpatients with TOP2A amplifıcation but not with TOP2A-



nonamplifıed gene status. Numerous studies have also inves-tigated the predictive role of TOP2A, with some evidence ofutility, but as yet data are inconclusive.22-24Until the issue regarding the need for anthracyclines is re-

solved, choice of adjuvant chemotherapy regimen forHER2� EBC should be based on clinical assessment of re-lapse risk. In the setting of high-risk features (pT � 1 and/orpN� 0), an anthracycline-based regimen is reasonable. Con-versely, lower-risk tumors could be treated with TCH. Alter-native chemotherapy backbones towhich anti-HER2 therapymay be added include TC or weekly paclitaxel; however, asthese regimens have not yet been formally validated in ran-domized adjuvant trials, they should be used cautiously.25

Triple-Negative Breast Cancer (TNBC)TNBC with good prognosis. TNBC is typically associated withpoor prognosis; however, emerging molecular, clinical, andpathologic data indicate that the TNBC group is heteroge-neous. In particular, histologic special types of breast can-cer—some of which are known to have excellent prognosis—are classifıed as TNBC. Of the 18 different histologic specialtypes of breast cancer, at least four rare types—medullary,apocrine, metaplastic and adenoid cystic carcinoma—arepredominantly negative for ER, PgR, and HER2.26 Metaplas-tic carcinomas show worse prognosis compared with TNBCof no special type. Conversely, adenoid cystic, andmedullarycarcinomas, despite the high proliferation index of the latter,have been consistently reported to have excellent prognoses.However, data on the role of adjuvant chemotherapy for spe-cial types of breast cancer are lacking,mainly because of theirlow frequencies.We recommend that, in the case of a good prognosis

TNBC, the diagnosis fırst be confırmed by a pathologist withexperience in rare tumor subtypes. Then, in the absence ofnodal involvement, adjuvant chemotherapy may be avoided.However, for node-positive disease, even of a good prognosissubtype, adjuvant chemotherapy is recommended, as thesafety of not treating with chemotherapy in this setting is un-known because of an absence of data.

Other TNBC. Excluding the rare, good prognosis types, TNBCshould be treated with chemotherapy, usually a sequentialanthracycline/taxane regimen, especially in patients withnode-positive disease. There is no data comparing a less in-tensive chemotherapy regimen with standard anthracyclineplus taxane for lower-risk (pT1, pN0) TNBC; although con-siderationmay be given to regimens such as TC orCMF, par-ticularly if there was specifıc concern regarding treatmenttoxicities.With TNBC and BRCA-associated breast cancers sharing

many biologic characteristics, the use of platinums has beenconsidered in TNBC. Although excellent responses to cispla-tin have been seen in patients with triple-negative, BRCA1-associated breast cancer, with pathologic complete response(pCR) rates of over 80%, pCR rates with cisplatin for spor-adic TNBC are considerably lower, around 20%.27,28 Thus,

platinum-based regimens are not recommended over stan-dard adjuvant regimens in sporadic TNBC outside of a clin-ical trial.

Chemotherapy in the Very Young PatientBreast cancer in young patients (younger than age 35) is typ-ically characterized by aggressive disease, including higherincidence of hormone-insensitive, undifferentiated, andHER2� tumors, andmay be associated with unique biologicfeatures compared with older women, while young age itselfis an independent risk factor for poor prognosis.29Athough luminal A–like EBC occurs less commonly in

younger women than older women, these tumors can betreated with endocrine therapy alone (usually combined es-trogen blockade), with excellent outcomes expected. How-ever, because young age itself is an poor prognostic factor, forluminal B, HER2�, or TNBC, we recommend use of adju-vant polychemotherapy with an anthracycline plus taxane-based regimen, even for node-negative disease.Importantly, with all adjuvant breast cancer chemotherapy

regimens being at least moderately gonadotoxic, young pa-tients with EBC should have early referral to a reproductivespecialist.

Chemotherapy in the Older PatientWhen considering adjuvant chemotherapy in older patients(age 65 or older) with breast cancer, the potential benefıtsfrom effective treatmentmust be balanced against risk of tox-icities, functional decline, and decreased quality of life. Olderwomen with biologically aggressive EBC may gain as muchbenefıt from adjuvant chemotherapy as younger women;however, this specifıcally applies to fıt elderly patients, withfew, if any, comorbidities.30 Even in this select group, in-creased toxicities from chemotherapy may be seen.31 Theutility of adjuvant chemotherapy in older patients who areless fıt is unknown.Determining which patients meet the defınition of “fıt” is

also problematic because of a lack of suitable frailty screeningtools. The comprehensive geriatric assessment remains thegold standard for detecting functional defıcits, although itsintegration into oncology practice is challenging.Ideally older patients should undergo some form of geriat-

ric assessment to determine their fıtness. We recommendconsideration of adjuvant chemotherapy for fıt older patientswith node-positive, HR- EBC, whereas chemotherapy mayalso be considered for HR�, luminal-B disease and node-negative, HR- disease. Based on superior outcomes withpolychemotherapy, combination treatment is preferable tosingle agent therapy.32 Conversely, in patients who are frailand less fıt, adjuvant chemotherapy outside of a clinical trialsetting is typically not recommended.

CONCLUSIONThe selection of both patients and regimens for adjuvant che-motherapy in EBC is an evolving fıeld. Ongoing research

DI LEO ET AL


should ideally be directed at better defıning which patientsmay avoid chemotherapy, which patients require chemo-therapy because of the presence of peripheral markers of

micrometastases, and which specifıc regimen is most effec-tive and safe based on the underlying tumor subtype and in-dividual patient characteristics.

Disclosures of Potential Conflicts of Interest

Relationships are considered self-held and compensated unless otherwise noted. Relationships marked “L” indicate leadership positions. Relationships marked “I” are those held by an immediatefamily member; those marked “B” are held by the author and an immediate family member. Relationships marked “U” are uncompensated.

Employment or Leadership Position: None. Consultant or Advisory Role: Angelo Di Leo, AstraZeneca; Novartis; Pfizer; Roche. Stock Ownership: None.Honoraria: Angelo Di Leo, AstraZeneca; Novartis; Pfizer; Roche. Research Funding: Angelo Di Leo, Pfizer. Expert Testimony: None. Other Remuneration:None.

References

1. Sorlie T, Perou CM, Tibshirani R, et al. Gene expression patterns ofbreast carcinomas distinguish tumor subclasses with clinical implica-tions. Proc Natl Acad Sci U S A. 2001;98:10869-10874.

2. International Breast Cancer Study Group (IBCSG). Endocrine respon-siveness and tailoring adjuvant therapy for postmenopausal lymphnode-negative breast cancer: a randomized trial. J Natl Cancer Inst.2002;94:1054-1065.

3. Berry DA, Cirrincione C, Henderson IC, et al. Estrogen-receptor statusand outcomes ofmodern chemotherapy for patients with node-positivebreast cancer. JAMA. 2006;295:1658-1667.

4. Paik S, Shak S, Tang G, et al. A multigene assay to predict recurrence oftamoxifen-treated, node-negative breast cancer.NEngl JMed. 2004;351:2817-2826.

5. Paik S, TangG, Shak S, et al. Gene expression and benefıt of chemother-apy in women with node-negative, estrogen receptor-positive breastcancer. J Clin Oncol. 2006;24:3726-3734.

6. Albain KS, Paik S, van’t Veer L. Prediction of adjuvant chemotherapybenefıt in endocrine responsive, early breast cancer using multigene as-says. Breast. 2009;18 Suppl 3:S141-S145.

7. Dowsett M, Cuzick J, Wale C, et al. Prediction of risk of distant recur-rence using the 21-gene recurrence score in node-negative and node-positive postmenopausal patients with breast cancer treated withanastrozole or tamoxifen: a TransATAC study. J Clin Oncol. 2010;28:1829-1834.

8. van ’t Veer LJ, Dai H, van de Vijver MJ, et al. Gene expression profılingpredicts clinical outcome of breast cancer. Nature. 2002;415:530-536.

9. Sotiriou C, Wirapati P, Loi S, et al. Gene expression profıling in breastcancer: understanding the molecular basis of histologic grade to im-prove prognosis. J Natl Cancer Inst. 2006;98:262-272.

10. Bhargava R, Beriwal S, DabbsDJ, et al. Immunohistochemical surrogatemarkers of breast cancer molecular classes predicts response to neoad-juvant chemotherapy: a single institutional experience with 359 cases.Cancer. 2010;116:1431-1439.

11. Cheang MC, Chia SK, Voduc D, et al. Ki67 index, HER2 status, andprognosis of patients with luminal B breast cancer. J Natl Cancer Inst.2009;101:736-750.

12. Dawood S, Hu R, Homes MD, et al. Defıning breast cancer prognosisbased on molecular phenotypes: Results from a large cohort study.Breast Cancer Res Treat. 2011;126:185-192.

13. Goldhirsch A, Wood WC, Coates AS, et al. Strategies for subtypes—dealing with the diversity of breast cancer: highlights of the St. GallenInternational Expert Consensus on the Primary Therapy of Early BreastCancer 2011. Ann Oncol. 2001;22:1736-1747.

14. Curtis C, Shah SP, Chin SF, et al. The genomic and transcriptomic ar-chitecture of 2,000 breast tumours reveals novel subgroups. Nature.2012;486:346-352.

15. Pierga JY, Bidard FC, Mathiot C, et al. Circulating tumor cell detectionpredicts early metastatic relapse after neoadjuvant chemotherapy inlarge operable and locally advanced breast cancer in a phase II random-ized trial. Clin Cancer Res. 2008;14:7004-7010.

16. Rack BK, Schindlbeck C, Andergassen A, et al. Use of circulating tumorcells (CTC) in peripheral blood of breast cancer patients before and afteradjuvant chemotherapy to predict risk of relapse: The SUCCESS trial.J Clin Oncol. 2010;28:15s (suppl; abstr 1003).

17. Huang ZH, Li LH, Hua D. Quantitative analysis of plasma circulatingDNA at diagnosis and during follow-up of breast cancer patients. Can-cer Lett. 2006;243:64-70.

18. WuX, Somlo G, Yu Y, et al. De novo sequencing of circulatingmiRNAsidentifıes novelmarkers predicting clinical outcome of locally advancedbreast cancer. J Transl Med. 2012;10:42.

19. Jones SE, Savin MA, Holmes FA, et al. Phase III trial comparing doxo-rubicin plus cyclophosphamide with docetaxel plus cyclophosphamideas adjuvant therapy for operable breast cancer. J Clin Oncol. 2006;24:5381-5387.

20. Fisher B, Brown AM, Dimitrov NV, et al. Two months ofdoxorubicin-cyclophosphamide with and without interval reinduc-tion therapy compared with 6 months of cyclophosphamide, meth-otrexate, and fluorouracil in positive-node breast cancer patientswith tamoxifen-nonresponsive tumors: results from the NationalSurgical Adjuvant Breast and Bowel Project B-15. J Clin Oncol. 1990;8:1483-1496.

21. Shulman LN, Cirrincione CT, Berry DA, et al. Six cycles of doxorubicinand cyclophosphamide or paclitaxel are not superior to four cycles asadjuvant chemotherapy for breast cancer in women with zero to threepositive axillary nodes: Cancer and Leukemia Group B 40101. J ClinOncol. 2012;30:4071-4076.

22. Slamon D, Eiermann W, Robert N, et al. Adjuvant trastuzumab inHER2-positive breast cancer. N Engl J Med. 2011;365:1273-1283.

23. Desmedt C, Di Leo A, de Azambuja E, et al. Multifactorial approach topredicting resistance to anthracyclines. J Clin Oncol. 2011;29:1578-1586.

24. Di Leo A, Desmedt C, Bartlett JM, et al. HER2 and TOP2A as predictivemarkers for anthracycline-containing chemotherapy regimens as adju-vant treatment of breast cancer: a meta-analysis of individual patientdata. Lancet Oncol. 2011;12:1134-1142.

25. Jones S, Collea RP, Oratz R, et al. Preliminary toxicity results of a phase



II trial of adjuvant docetaxel/cyclophosphamide plus trastuzumabin HER2� early stage breast cancer. Cancer Res. 2009;69 (abstr2111).

26. Tavassoli FA, Devilee P (eds). Pathology & Genetics of Tumours of theBreast and Female Genital Organs. Lyon; International Agency for Re-search on Cancer Press; 2003.

27. Byrski T, Gronwald J, Huzarski T, et al. Pathologic complete responserates in young womenwith BRCA1-positive breast cancers after neoad-juvant chemotherapy. J Clin Oncol. 2010;28:375-379.

28. Silver DP, Richardson AL, Eklund AC, et al. Effıcacy of neoadjuvantcisplatin in triple-negative breast cancer. J Clin Oncol. 2010;28:1145-1153.

29. Anders CK, Hsu DS, Broadwater G, et al. Young age at diagnosis corre-lates with worse prognosis and defınes a subset of breast cancers withshared patterns of gene expression. J Clin Oncol. 2008;26:3324-3330.

30. Muss HB. Adjuvant treatment of elderly breast cancer patients. Breast16 Suppl. 2007;2:S159-S165.

31. Muss HB, Berry DA, Cirrincione C, et al. Toxicity of older and youngerpatients treated with adjuvant chemotherapy for node-positive breastcancer: the Cancer and Leukemia Group B Experience. J Clin Oncol.2007;25:3699-3704.

32. Muss HB, Berry DA, Cirrincione CT, et al. Adjuvant chemotherapy inolder women with early-stage breast cancer. N Engl J Med. 2009;360:2055-2065.

DI LEO ET AL


Early and Late Long-Term Effects of Adjuvant ChemotherapyErica L. Mayer, MD, MPH

OVERVIEW

Adjuvant chemotherapy continues to play an important role in breast cancer management. Exposure to chemotherapy can lead to avariety of early and late long-term toxicities, including ovarian failure (with resultant infertility and sexual dysfunction), bone loss,weight gain, neurotoxicity, neurocognitive changes, cardiac toxicity and secondary malignancy. Although chemotherapy effects mayvary in medical severity, all effects have the potential to lead to a decrease in quality of life and a decrement on overall health status.Improved understanding of the etiology and management of chemotherapy-related toxicity may allow optimization of patient selectionfor treatment and ameliorate the concerns of patients who are considering embarking on a chemotherapy program. This article presentsan overview of relevant early and late long-term toxicities, with a focus on recent advances and clinical management.

With the advent of genomic testing, rates of use of adju-vant chemotherapy for breast cancer have been in de-

cline,1 although chemotherapy continues to play a signifıcantrole in the adjuvant therapy of patients with triple-negativeand HER2-positive disease. Administration of chemother-apy can reduce risk of breast cancer recurrence and is gen-erally considered tolerable with the use of contemporarysupportive care, however, exposure to typical regimens maylead to both early and late toxicities. Given the estimated 3million breast cancer survivors in the United States, manywomen are at risk of chemotherapy-related effects. Im-proved understanding of the etiology and management ofchemotherapy-related toxicity may allow optimization ofpatient selection for chemotherapy and ameliorate the con-cerns of patients who are considering embarking on a chemo-therapy program.

OVARIAN FAILURE: PREMATURE MENOPAUSE,INFERTILITY, SEXUAL DYSFUNCTIONChemotherapy-induced premature ovarian failure is a note-worthy toxicity for premenopausal women receiving adjuvantchemotherapy. Although the majority of women may experi-ence temporary amenorrhea, in many, ovarian function willreturn in the months following completion of treatment.However, a subsetwill experience permanent chemotherapy-induced ovarian failure, with the risk of permanent meno-pause increasing with age and modulated by chemotherapytype and duration.2,3An implied additional cost of infertility exists for patients

who sustain permanent amenorrhea. For patients with hor-mone receptor–positive cancers, infertility can also occur

secondary to natural ovarian aging during long-term endo-crine therapy. Even for patients who regain menses afterchemotherapy, there may be an increased likelihood of pre-mature menopause (under age 45) in patients who receivechemotherapy at a younger age.4 Concerns over risks ofinfertility can be substantial for young patients with breastcancer and may influence treatment selection.5,6 Pretherapyreferral to a fertility specialist for a discussion of fertilitypreservation may help alleviate concerns and is endorsed byASCO guidelines.7,8 Assisted reproductive technologies maybe considered for patients facing infertility after treatment,although there are theoretical concerns about ovarian stim-ulation in the setting of hormone receptor–positive cancer.Specialty consultation to discuss newer techniques, includingovarian stimulation with aromatase inhibitors, is recom-mended.9 Available data to date do not suggest adverse out-comes in patients who become pregnant after a breast cancerdiagnosis, regardless of hormone receptor status of the tumor.9One of the most prominent early effects of ovarian failure

can be frequent and disruptive hot flashes, which may be ac-centuated by concomitant use of adjuvant endocrine therapy.Multiple management techniques have been evaluated10,11;commonly used strategies include behavior modifıcation, anti-depressants, clonidine, gabapentin, vitamin E, and acupunc-ture.12-14 Data to date do not conclusively suggest inferioroutcomes after concurrent use of CYP2D6-inhibiting anti-depressants and tamoxifen; however, recent prescribing trendshave supported a shift toward greater use of weak CYP2D6-inhibitingantidepressants inbreast cancer survivors.15Atrophicvaginitis is another common and signifıcant side effectwhich can lead to dysuria, frequent urinary tract infections,pruritus, and dyspareunia. Nonhormone lubricants (both

From the Breast Oncology Center, Dana-Farber Cancer Institute, and Harvard Medical School, Boston, MA.

Author’s disclosures of potential conflicts of interest are found at the end of this article.

Corresponding author: Erica L. Mayer, MD, MPH, Dana-Farber Cancer Institute, 450 Brookline Ave., Boston, MA 02468; email: [email protected].


LONG-TERM EFFECTS OF ADJUVANT CHEMOTHERAPY


water-based and silicone) are typically recommended, par-ticularly in a patientwith a history of hormone receptor–pos-itive breast cancer. Topical vaginal estrogen therapymay be amore effective approach16; although estrogen levels havebeen noted to rise with local estrogen delivery, no signal ofincreased risk of cancer recurrence has been noted.17,18 Ad-ditional recommended lifestyle interventions include in-creased sexual activity and Kegel exercises to improve pelvicmuscle tone.Sexual dysfunction is a common sequelae of treatment-

related ovarian failure, and can be manifest by decreased li-bido, dyspareunia, and diffıculty with orgasm; it is estimatedto affect up to 50% of all breast cancer survivors, particularlythose who experience treatment-related ovarian failure.19,20The etiology of sexual dysfunction is multifactorial, likely re-flecting not only reduced levels of estrogen, but also the psy-chologic trauma of cancer diagnosis and treatment for ayoungwoman. Specifıc interventionsmay include vaginal lu-bricants, avoidance of concomitant medications which de-crease libido, and counseling with a sexual health specialist.

WEIGHT GAINIt has long been recognized that women who receive adju-vant chemotherapy for breast cancer gain weight, on averageabout 10 lb. Reasons for weight gain are likely multifactorial,including changes in activity level, menopausal status, endo-crine manipulation, diet, metabolism, and mood. Results ofobservational studies describing the effect of weight gain af-ter a breast cancer diagnosis on future disease recurrencehave been variable,21,22 although analyses from the Nurses’Health Study and other studies have suggested weight gainmay increase risk of disease recurrence.23 Diet and exerciseinterventions have demonstrated success in preventingweight gain or stimulating weight loss,24,25 and currentguidelines for exercise training in early cancer survivors havebeen published.26 All efforts should be made to encouragebreast cancer survivors to “avoid inactivity” and pursue amore active lifestyle.

BONE LOSSLoss of bone density is a common occurrence in both pre-and postmenopausal breast cancer survivors. Proposed etiol-ogies include premature menopause as well as the effect ofbreast cancer therapies, specifıcally aromatase inhibitors, oncirculating estrogen levels.27 The development of decreasedbone densitymay have signifıcant health consequences. In ananalysis from theWomen’s Health Initiative, increased frac-ture risk was seen in postmenopausal breast cancer survivorscompared with participants without breast cancer.28 A vari-ety of guidelines for bone health screening and managementof osteopenia and osteoporosis have been published.29,30 Ingeneral, it is recommended to screen at-risk individuals forosteoporosis by dual-energy x-ray absorptiometry every 1 to2 years, and consider initiation of bisphosphonate therapyfor scores defıning osteoporosis. Othermaneuvers, includingpursuing weight-bearing exercise and adequate calcium andvitamin D, are generally recommended for breast cancersurvivors.

NEUROPATHYPeripheral sensory or motor neuropathy can occur after ex-posure to microtubule inhibitors such as taxanes. Sensoryneuropathy can be characterized by both paresthesias andpain, which can signifıcantly affect a patient’s quality of life.31Rates and severity of taxane-related neuropathy vary and re-flect agent selection, dose, schedule, and comorbidities. Sup-portive management during treatment typically includesdose modifıcation and treatment delay; supplemental pre-ventative agents including glutathione, acetyl-L-carnitine,and alpha-lipoic acid are under investigation.32 Althoughsome patients will experience gradual improvement in neu-ropathy, many are left with residual and potentially disablingsymptoms. Multiple management techniques have beenevaluated, including use of gabapentin or venlafaxine, forwhich there is supportive data from a randomized trial.33CALGB 170601, a randomized phase III trial of duloxetinefor painful chemotherapy-induced neuropathy, demon-strated reductions in pain scores with duloxetine comparedwith placebo.35 Investigation of additional nonpharmaco-logic modalities is ongoing.Ultimately, best management of peripheral neuropathy

would be avoidance of the toxicity through identifıcation ofindividuals at greatest risk.One path of investigation has sug-gested higher paclitaxel acute pain syndrome (P-APS) scoreswith fırst dose of chemotherapy may correlate with periph-eral neuropathy.35,36 Pharmacogenomic analysis has identi-fıed single nucleotide polymorphisms (SNPs) associatedwiththe development of moderate to severe peripheral neuropa-thy after paclitaxel exposure.37-39 Reassuringly, the develop-ment of neuropathy does not appear to be correlated withchemotherapy effıcacy,40 therefore supporting the safety offurther development of predictive biomarkers for neurotox-icity, including SNPs or P-APS scores, which may allowmodifıcation of adjuvant chemotherapy selection to reducethe risk of permanent disabling neuropathy.

KEY POINTS

� Adjuvant chemotherapy can lead to early and late long-term side effects for breast cancer survivors.

� Given the number of breast cancer survivors, many womenare at risk of experiencing toxicity.

� Effects of chemotherapy vary in severity, but can oftennegatively affect quality of life and overall health status.

� Management of long-term chemotherapy-related toxicityinvolves screening for symptoms, use of supportivemedication, and referral for specialty consultation asneeded.

� Ongoing research is evaluating both the etiology of toxicityas well as effective interventions.

ERICA L. MAYER


CARDIAC DYSFUNCTIONThe most common chemotherapy-related cardiotoxicity ob-served in breast cancer survivors is left ventricular dysfunc-tion.41 Although cyclophosphamide and taxanes have beenassociated with a variety of cardiac complications, the mostcommon agents implicated in adjuvant therapy after chemo-therapy are anthracyclines and trastuzumab.Anthracycline-mediated cardiotoxicity can occur in an

acute or subacute fashion, however, the majority of casesare late-onset, occurring at least 1 year after completion oftherapy.42 Late-onset cardiotoxicity tends to be irreversible,is related to cumulative anthracycline dose, and may re-flect a variety of intracellularmechanisms including free rad-ical formation.43 Although absolute rates of anthracycline-related cardiac dysfunction in clinical trials generally havebeen low, analysis of nontrial older populations through theSurveillance, Epidemiology, and End Results (SEER) data-base suggests the incidencemay be higher, with a rate of CHFof 38.4% in patients with breast cancer ages 66 to 70 treatedwith adjuvant anthracycline-based therapy compared with29% in those who did not get chemotherapy.44Exposure to trastuzumab, the highly effective monoclonal

antibody for HER2-positive breast cancer, may lead to cardiacdysfunction following a different pattern than anthracycline-mediated toxicity: typically occurring during time of medi-cation administration, and generally reversible with a hold intherapy and use of cardiac medication.45 It has been pro-posed that the differential patterns of cardiotoxicity betweenthe two agents reflect divergent mechanisms of action, withanthracyclines causing permanent cardiomyocyte apoptosisand necrosis, and trastuzumab leading to temporary cellularchanges.41 Rates of symptomatic or severe cardiac dysfunc-tion in the major trials of adjuvant trastuzumab have variedfrom a maximum of 4.1% with a standard anthracycline-containing regimen,46 to 0.6% to 1.87% in regimens withoutconcurrent chemotherapy and trastuzumab or without anthra-cyclines.47-49 Rates of asymptomatic drop in cardiac functionare higher, in the range of 17% to 19% after anthracyclineexposure in themajor adjuvant trials.36,37 Reassuringly, long-term follow-up from the adjuvant trastuzumab trials havesuggested late cardiotoxicity events related to trastuzumabappear to be rare.50,51 Routine evaluation of cardiac functionis recommended for patients receiving ongoing trastuzumab;however, following completion of therapy, further cardiacevaluation is symptom-driven only.Prechemotherapy identifıcation of individuals at risk of

cardiac toxicity would improve selection of chemotherapyregimens. Clinical risk factors predictive of cardiotoxicityhave included older age, pre-existing hypertension, low base-line left ventricular ejection fraction, and elevated bodymassindex. Long-term follow-up data from NSABP B-31 havebeen used to construct a Cardiac Risk Score, incorporatingage and left ventricular ejection fraction (LVEF), to predictrisk of a cardiac event.51 Novel imaging, including echocar-diographic techniques with greater sensitivity to discern sub-

clinical cardiac dysfunction, and potentially concurrent useof cardiac biomarkers, may have even more promise as pre-dictive tools.52 There is no clear role for routine screening forleft ventricular dysfunction in breast cancer survivors in theabsence of clinical symptoms, however, for survivors withsuspected cardiac toxicity, imaging with echocardiographyand referral to cardiology are strongly recommended.

NEUROCOGNITIVE DYSFUNCTIONNeurocognitive changes after chemotherapy exposure,termed “chemo-brain,” are a source of serious concern andanxiety for breast cancer survivors.53,54 It is estimated up to75% of women who receive chemotherapy will report achange in cognitive function in the 2 years after treatment.55Patients typically describe problemswith attention,memory,and concentration.Studies to date have been complicated by small sample size,

diverse defınitions of cognitive impairment, and variablepre-exposure assessments of cognitive function. Addition-ally, studies evaluating cognitive dysfunction during che-motherapy present divergent results from those using apostchemotherapy time-point. Early cross-sectional studiesdemonstrated increased cognitive impairment in patientswith breast cancer during and after chemotherapy whencompared with healthy matched controls.56,57 Subsequently,prospective longitudinal studies have suggested evidence ofpostchemotherapy change in cognitive dysfunction, at leastin a subset of patients.54 A recent meta-analysis has analyzed17 studies evaluating cognitive functioning in the post-chemotherapy period, confırming small defıcits in verbalability and visuospatial control in patients at least 6 months

Table 1. Summary of Complications with AdjuvantChemotherapy

ToxicityRole for Screening/RecommendedInterventions

Ovarian failure Behavioral/medical management hotflashes, reproductiveendocrinology consultation forfertility counseling

Sexual dysfunction Non-hormonal lubricants, counseling

Weight gain Exercise and diet interventions

Neuropathy Consider venlafaxine, duloxetine forpainful symptoms, glutamate,vitamin B6

Cardiotoxicity Echocardiogram for symptoms.Referral to cardiology as needed

Cognitive dysfunction Referral to neuropsychologist,cognitive behavioral therapy

Osteoporosis Screening bone densityexamination. Use of calcium,vitamin D, bisphosphonates

Secondary hematologicmalignancies

Evaluation of peripheral bloodcounts in setting of symptoms



out from chemotherapy compared with testing prechemo-therapy or in healthy participants.58 Studies evaluating pre-and post chemotherapy imaging have also attempted toinvestigate anatomic correlates of cognitive defıcits; smallprospective trials using MRI have suggested both structuralchanges and changes in brain activation after chemotherapyexposure.59,60It has been proposed, however, that the etiology of cogni-

tive dysfunction is likelymultifactorial; although exposure tochemotherapy may contribute, other factors are likely rele-vant as well, including other modalities of treatment (sur-gery, radiotherapy, endocrine therapy), supportive caremedications,menopausal symptoms, anxiety, depression, fa-tigue, or other comorbid conditions.61 An International Cog-nition and Cancer Task Force has suggested a subgroup ofpatients may be especially sensitive to neurocognitive effectsof chemotherapy, including those of advanced age or lowercognitive reserve.55,62 Furthermore, genetic polymorphismsin susceptibility genes, including apolipoprotein E (APOE)and catechol-O-methyltransferase (COMT)may identify in-dividuals with vulnerability to cognitive dysfunction afterchemotherapy.63,64 Variability in age and polymorphism dis-tribution may explain the variability in results in some of theprospective cognitive studies to date. There is interest inwhether prophylactic or therapeutic interventions, includingthe psychostimulant modafınil,65 fluoxetine,66 or structuredcognitive behavioral therapy,67 may improve symptoms.Further work is necessary to confırm if individuals at partic-ularly high risk of cognitive change can be identifıed in ad-vance to protect from exposures or treat with preventativemedication, while providing reassurance to others who maybe unlikely to experience this toxicity.

SECONDARY MALIGNANCYOne of the rarest yet most feared long-term toxicities of ad-juvant chemotherapy is hematologicmalignancy, specifıcallymyelodysplasia (MDS) or acute myeloid leukemia (AML).The etiology of this complication is thought to reflect expo-sure to topoisomerase II–targeted agents (anthracyclines) oralkylating agents (cyclophosphamide), which are frequentlyincluded in adjuvant chemotherapy regimens. Topoisomer-ase II–related myeloid malignancies may present within 5years of exposure, may not be preceded by MDS, and mayhave abnormal cytogenetics of 11q23. Alkylator-related ma-lignancies present after a longer duration of time,may be pre-ceded by MDS, and may have cytogenetic abnormalities ofchromosomes 5 and 7.68 The risk of secondary malignancyappears to reflect increased cumulative dose exposure, al-though most series report population rates of less than

1%.69,70 Concurrent use of growth factors to support “dose-dense” scheduling, or radiotherapy, may contribute a slightadditional risk of myeloid malignancy, although these topicsremain under investigation.71,72Older patients are thought to be at increased risk of my-

eloid complications, although they have typically been ex-cluded from analysis of younger cohorts who participate inclinical trials. A SEER-Medicare population-based analysisof over 64,000 older patients with breast cancer (median age76) suggested a small but signifıcant increase in the risk ofdeveloping AML after chemotherapy, with a hazard ratio of1.53 (p � 0.005), although absolute risks at 10 years remainsmall (1.8%with chemotherapy vs. 1.2%without chemother-apy).73 Analysis of over 21,000 patients in the National Com-prehensive Cancer Network (NCCN) suggested a 10-yearrate of MDS/AML of 0.27%, with a slight additional increasein those who had received adjuvant chemotherapy.74Methods to reduce exposure to potentially bone marrow–

toxic chemotherapy may help reduce the risk of secondarymyeloid malignancies. Adjuvant regimens which replace an-thracycline with a taxane have demonstrated slightly de-creased rates of secondary myeloid malignancies,47,75 andcould be considered in a situation where risk of myeloid ma-lignancy is a signifıcant concern. Specifıc predictors of futureMDS/AML are not known, however, continued use ofgenomic tools to better identify chemotherapy-sensitive tu-mors would help improve the risk/benefıt ratio for this tox-icity by reducing chemotherapy exposure in patients unlikelyto receive anticancer benefıt.

CONCLUSIONDespite the gains in reduction of cancer recurrence with ad-ministration of adjuvant chemotherapy, breast cancer survi-vors may be exposed to risks of both short- and long-termtoxicities after chemotherapy. Although chemotherapy ef-fects may vary in medical severity, all effects have the poten-tial to lead to a decrease in quality of life and a decrement onoverall health status. Improved management of the effects ofchemotherapy will require better understanding of manage-ment strategies, and larger prospective trials evaluating a va-riety of interventions are underway. Additionally, for the fırsttime, the 2013 NCCN Guidelines will include guidelines forscreening breast cancer survivors for many of the commonshort- and long-term effects from therapy (Ligibel L.,personal communication, 2013). After decades of recogni-tion of many of the detrimental effects of treatment, it ishoped that future clinical and research activities will defıni-tively reduce adverse therapy-related outcomes for breastcancer survivors.


The author(s) indicated no potential conflicts of interest.

ERICA L. MAYER


References

1. Hassett MJ, Silver SM, Hughes ME, et al. Adoption of gene expressionprofıle testing and associationwith use of chemotherapy amongwomenwith breast cancer. J Clin Oncol. 2012;30:2218-2226.

2. Walshe JM, Denduluri N, Swain SM. Amenorrhea in premenopausalwomen after adjuvant chemotherapy for breast cancer. J Clin Oncol.2006;24:5769-5779.

3. Ruddy KJ, Partridge AH. Fertility (male and female) and menopause.J Clin Oncol. 2012;30:3705-3711.

4. Letourneau JM, Ebbel EE, Katz PP, et al. Acute ovarian failure underes-timates age-specifıc reproductive impairment for young women under-going chemotherapy for cancer. Cancer. 2012;118:1933-1939.

5. PartridgeAH,Gelber S, Peppercorn J, et al.Web-based survey of fertilityissues in young women with breast cancer. J Clin Oncol. 2004;22:4174-4183.

6. Peate M, Meiser B, Friedlander M, et al. It’s now or never: fertility-related knowledge, decision-making preferences, and treatment inten-tions in young women with breast cancer-an Australian fertilitydecision aid collaborative group study. J ClinOncol. 2011;29:1670-1677.

7. Lee SJ, Schover LR, Partridge AH, et al. American Society of ClinicalOncology recommendations on fertility preservation in cancer patients.J Clin Oncol. 2006;24:2917-2931.

8. Letourneau JM, Ebbel EE, Katz PP, et al. Pretreatment fertility counsel-ing and fertility preservation improve quality of life in reproductive agewomen with cancer. Cancer. 2012;118:1710-1717.

9. Azim AA, Costantini-FerrandoM, Oktay K. Safety of fertility preserva-tion by ovarian stimulationwith letrozole and gonadotropins in patientswith breast cancer: a prospective controlled study. J ClinOncol. 2008;26:2630-2635.

10. Morrow PK, Mattair DN, Hortobagyi GN. Hot flashes: a review ofpathophysiology and treatment modalities. Oncologist. 2011;16:1658-1664.

11. HickeyM, Saunders C, PartridgeA, et al. Practical clinical guidelines forassessing andmanagingmenopausal symptoms after breast cancer.AnnOncol. 2008;19:1669-1680.

12. Loprinzi CL, Sloan J, Stearns V, et al. Newer antidepressants and gaba-pentin for hot flashes: an individual patient pooled analysis. J ClinOncol.2009;27:2831-2837.

13. Barton DL, Loprinzi CL, Quella SK, et al. Prospective evaluation of vi-tamin E for hot flashes in breast cancer survivors. J Clin Oncol. 1998;16:495-500.

14. Bokmand S, Flyger H. Acupuncture relieves menopausal discomfort inbreast cancer patients: A prospective, double blinded, randomizedstudy. Breast. Epub 2012 Aug 17.

15. Nachtigall LE. Comparative study: Replens versus local estrogen inmenopausal women. Fertil Steril. 1994;61:178-180.

16. Dusetzina SB, Alexander GC, Freedman RA, et al. Trends in co-prescribing of antidepressants and tamoxifen among women withbreast cancer, 2004-2010. Breast Cancer Res Treat. 2013 Jan;137:285-296.

17. Wills S, Ravipati A, Venuturumilli P, et al. Effects of vaginal estrogenson serumestradiol levels in postmenopausal breast cancer survivors andwomen at risk of breast cancer taking an aromatase inhibitor or a selec-tive estrogen receptor modulator. J Oncol Pract. 2012;8:144-148.

18. Le Ray I, Dell’Aniello S, Bonnetain F, et al. Local estrogen therapy andrisk of breast cancer recurrence among hormone-treated patients: anested case-control study. Breast Cancer Res Treat. 2012;135:603-609.

19. Burwell SR,Case LD,KaelinC, et al. Sexual problems in youngerwomenafter breast cancer surgery. J Clin Oncol. 2006;24:2815-2821.

20. Ganz PA, Rowland JH, Desmond K, et al. Life after breast cancer: un-

derstanding women’s health-related quality of life and sexual function-ing. J Clin Oncol. 1998;16:501-514.

21. Thivat E, Therondel S, Lapirot O, et al. Weight change during chemo-therapy changes the prognosis in non metastatic breast cancer for theworse. BMC Cancer. 2010;10:648.

22. Caan BJ, Emond JA, Natarajan L, et al. Post-diagnosis weight gain andbreast cancer recurrence inwomenwith early stage breast cancer.BreastCancer Res Treat. 2006;99:47-57.

23. Kroenke CH, Chen WY, Rosner B, et al. Weight, weight gain, and sur-vival after breast cancer diagnosis. J Clin Oncol. 2005;23:1370-1378.

24. VillariniA, Pasanisi P, RaimondiM, et al. Preventingweight gain duringadjuvant chemotherapy for breast cancer: a dietary intervention study.Breast Cancer Res Treat. 2012;135:581-589.

25. SegalR,PondGR,VallisM,et al.Randomized trialof a lifestyle interventionfor women with early-stage breast cancer (BC) receiving adjuvant hor-mone therapy: Initial results. J Clin Oncol. 2011;29 (suppl; abstr 512).

26. Schmitz KH, Courneya KS, Matthews C, et al. American College ofSportsMedicine roundtable on exercise guidelines for cancer survivors.Med Sci Sports Exerc. 2010;42:1409-1426.

27. Guise TA. Bone loss and fracture risk associated with cancer therapy.Oncologist. 2006;11:1121-1131.

28. ChenZ,MaricicM,BassfordTL, et al. Fracture risk amongbreast cancersurvivors: results from the Women’s Health Initiative ObservationalStudy. Arch Intern Med. 2005;165:552-558.

29. Hadji P, Aapro MS, Body JJ, et al. Management of aromatase inhibitor-associated bone loss in postmenopausal women with breast cancer:practical guidance for prevention and treatment. Ann Oncol. 2011;22:2546-2555.

30. LustbergMB, Reinbolt RE, Shapiro CL. Bone health in adult cancer sur-vivorship. J Clin Oncol. 2012;30:3665-3674.

31. Swain SM, Arezzo JC. Neuropathy associated with microtubule inhibi-tors: diagnosis, incidence, and management. Clin Adv Hematol Oncol.2008;6:455-467.

32. Pachman DR, Barton DL, Swetz KM, et al. Troublesome symptoms incancer survivors: fatigue, insomnia, neuropathy, and pain. J Clin Oncol.2012;30:3687-3696.

33. Durand JP, Deplanque G, Montheil V, et al. Effıcacy of venlafaxine forthe prevention and relief of oxaliplatin-induced acute neurotoxicity: re-sults of EFFOX, a randomized, double-blind, placebo-controlled phaseIII trial. Ann Oncol. 2012;23:200-205.

34. Lavoie Smith EM, PangH, Cirrincione C, et al. CALGB 170601: A phaseIII double blind trial of duloxetine to treat painful chemotherapy-induced peripheral neuropathy (CIPN). J Clin Oncol. 2012;30 (suppl;abstr CRA9013).

35. Loprinzi CL, Reeves BN, Dakhil SR, et al. Natural history of paclitaxel-associated acute pain syndrome: prospective cohort study NCCTGN08C1. J Clin Oncol. 2011;29:1472-1478.

36. Reeves BN, Dakhil SR, Sloan JA, et al. Further data supporting thatpaclitaxel-associated acute pain syndrome is associated with develop-ment of peripheral neuropathy: North Central Cancer TreatmentGroup trial N08C1. Cancer. 2012;118:5171-5178.

37. Schneider BP, Li L, Miller K, et al. Genetic associations with taxane-induced neuropathy by a genome-wide association study (GWAS) inE5103. J Clin Oncol. 2011;29 (suppl; abstr 1000).

38. Sucheston LE, Zhao H, Yao S, et al. Genetic predictors of taxane-induced neurotoxicity in a SWOG phase III intergroup adjuvant breastcancer treatment trial (S0221). Breast Cancer Res Treat. 2011;130:993-1002.

39. Baldwin RM, Owzar K, Zembutsu H, et al. A genome-wide association



study identifıes novel loci for paclitaxel-induced sensory peripheralneuropathy in CALGB 40101. Clin Cancer Res. 2012;18:5099-109.

40. Schneider BP, Zhao F,WangM, et al. Neuropathy is not associated withclinical outcomes in patients receiving adjuvant taxane-containingtherapy for operable breast cancer. J Clin Oncol. 2012;30:3051-3057.

41. Curigliano G, Mayer EL, Burstein HJ, et al. Cardiac toxicity from sys-temic cancer therapy: a comprehensive review. Prog Cardiovasc Dis.2010;53:94-104.

42. Von Hoff DD, Layard MW, Basa P, et al. Risk factors for doxorubicin-induced congestive heart failure. Ann Intern Med. 1979;91:710-717.

43. Gianni L, Herman EH, Lipshultz SE, et al. Anthracycline cardiotoxicity:from bench to bedside. J Clin Oncol. 2008;26:3777-3784.

44. PinderMC, Duan Z, Goodwin JS, et al. Congestive heart failure in olderwomen treated with adjuvant anthracycline chemotherapy for breastcancer. J Clin Oncol. 2007;25:3808-3815.

45. de Azambuja E, Bedard PL, Suter T, et al. Cardiac toxicity with anti-HER-2 therapies: what have we learned so far? Target Oncol. 2009;4:77-88.

46. Tan-Chiu E, Yothers G, Romond E, et al. Assessment of cardiac dys-function in a randomized trial comparing doxorubicin and cyclophos-phamide followed by paclitaxel, with or without trastuzumab asadjuvant therapy in node-positive, human epidermal growth factor re-ceptor 2-overexpressing breast cancer: NSABP B-31. J Clin Oncol. 2005;23:7811-7819.

47. Slamon D, Eiermann W, Robert N, et al. Adjuvant trastuzumab inHER2-positive breast cancer. N Engl J Med. 2011;365:1273-1283.

48. Piccart-GebhartMJ, ProcterM, Leyland-Jones B, et al. Trastuzumab af-ter adjuvant chemotherapy in HER2-positive breast cancer. N EnglJ Med. 2005;353:1659-1672.

49. Suter TM, ProcterM, vanVeldhuisenDJ, et al. Trastuzumab-associatedcardiac adverse effects in the herceptin adjuvant trial. J ClinOncol. 2007;25:3859-3865.

50. Procter M, Suter TM, de Azambuja E, et al. Longer-term assessment oftrastuzumab-related cardiac adverse events in the Herceptin Adjuvant(HERA) trial. J Clin Oncol. 2010;28:3422-3428.

51. Romond EH, Jeong JH, Rastogi P, et al. Seven-year follow-up assess-ment of cardiac function inNSABPB-31, a randomized trial comparingdoxorubicin and cyclophosphamide followed by paclitaxel (ACP) withACP plus trastuzumab as adjuvant therapy for patients with node-positive, human epidermal growth factor receptor 2-positive breast can-cer. J Clin Oncol. 2012;30:3792-799.

52. Sawaya H, Sebag IA, Plana JC, et al. Assessment of echocardiographyand biomarkers for the extended prediction of cardiotoxicity in patientstreated with anthracyclines, taxanes, and trastuzumab. Circ CardiovascImaging. 2012;5:596-603.

53. Janelsins MC, Kohli S, Mohile SG, et al. An update on cancer- andchemotherapy-related cognitive dysfunction: current status. Semin On-col. 2011;38:431-438.

54. Ahles TA, Root JC, Ryan EL. Cancer- and cancer treatment-associatedcognitive change: an update on the state of the science. J Clin Oncol.2012;30:3675-3686.

55. Wefel JS, Vardy J, Ahles T, et al. International Cognition and CancerTask Force recommendations to harmonise studies of cognitive func-tion in patients with cancer. Lancet Oncol. 2011;12:703-708.

56. Brezden CB, Phillips KA, Abdolell M, et al. Cognitive function in breastcancer patients receiving adjuvant chemotherapy. J Clin Oncol. 2000;18:2695-2701.

57. Tchen N, Juffs HG, Downie FP, et al. Cognitive function, fatigue, andmenopausal symptoms in women receiving adjuvant chemotherapy forbreast cancer. J Clin Oncol. 2003;21:4175-4183.

58. JimHS, Phillips KM, Chait S, et al. Meta-analysis of cognitive function-

ing in breast cancer survivors previously treated with standard-dosechemotherapy. J Clin Oncol. 2012;30:3578-3587.

59. Deprez S, Amant F, Smeets A, et al. Longitudinal assessment ofchemotherapy-induced structural changes in cerebral white matter andits correlation with impaired cognitive functioning. J Clin Oncol. 2012;30:274-281.

60. McDonald BC, Conroy SK, Ahles TA, et al. Alterations in brain activa-tion during working memory processing associated with breast cancerand treatment: a prospective functional magnetic resonance imagingstudy. J Clin Oncol. 2012;30:2500-2508.

61. Hurria A, Somlo G, Ahles T. Renaming “chemobrain”. Cancer Invest.2007;25:373-377.

62. Ahles TA, Saykin AJ, McDonald BC, et al. Longitudinal assessment ofcognitive changes associated with adjuvant treatment for breast cancer:impact of age and cognitive reserve. J Clin Oncol. 2010;28:4434-4440.

63. Ahles TA, Saykin AJ, Noll WW, et al. The relationship of APOE geno-type to neuropsychological performance in long-term cancer survivorstreated with standard dose chemotherapy. Psychooncology. 2003;12:612-619.

64. Small BJ, Rawson KS, Walsh E, et al. Catechol-O-methyltransferase ge-notype modulates cancer treatment-related cognitive defıcits in breastcancer survivors. Cancer. 2011;117:1369-1376.

65. Kohli S, Fisher SG, Tra Y, et al. The effect of modafınil on cognitivefunction in breast cancer survivors. Cancer. 2009;115:2605-2616.

66. Lyons L, ElBeltagyM,BennettG, et al. Fluoxetine counteracts the cognitiveand cellular effects of 5-fluorouracil in the rat hippocampus by a mecha-nism of prevention rather than recovery. PLoS One. 2012;7:e30010.

67. FergusonRJ,McDonald BC, RocqueMA, et al. Development of CBT forchemotherapy-related cognitive change: results of a waitlist controltrial. Psychooncology. 2012;21:176-186.

68. Pedersen-Bjergaard J, Pedersen M, Roulston D, et al. Different geneticpathways in leukemogenesis for patients presenting with therapy-related myelodysplasia and therapy-related acute myeloid leukemia.Blood. 1995;86:3542-3552.

69. Smith RE, Bryant J, DeCillis A, et al. Acute myeloid leukemia and my-elodysplastic syndrome after doxorubicin-cyclophosphamide adjuvanttherapy for operable breast cancer: the National Surgical AdjuvantBreast and Bowel Project Experience. J Clin Oncol. 2003;21:1195-1204.

70. Praga C, Bergh J, Bliss J, et al. Risk of acute myeloid leukemia and my-elodysplastic syndrome in trials of adjuvant epirubicin for early breastcancer: correlation with doses of epirubicin and cyclophosphamide.J Clin Oncol. 2005;23:4179-4191.

71. Hershman D, Neugut AI, Jacobson JS, et al. Acute myeloid leukemia ormyelodysplastic syndrome following use of granulocyte colony-stimulating factors during breast cancer adjuvant chemotherapy. J NatlCancer Inst. 2007;99:196-205.

72. Fisher B, RocketteH, Fisher ER, et al. Leukemia in breast cancer patientsfollowing adjuvant chemotherapy or postoperative radiation: theNSABP experience. J Clin Oncol. 1985;3:1640-1658.

73. Patt DA, Duan Z, Fang S, et al. Acute myeloid leukemia after adjuvantbreast cancer therapy in older women: understanding risk. J Clin Oncol.2007;25:3871-3876.

74. Karp J, Blackford A, Visvanathan K, et al. Myelodysplastic syndromeand/or acute myelogenous leukemia (MDS and/or AML) after a breastcancer diagnosis: the National Comprehensive Cancer Network(NCCN) experience. Cancer Res. 2012;72(suppl):S3-5.

75. Jones S, Holmes FA, O’Shaughnessy J, et al. Docetaxel with cyclophos-phamide is associated with an overall survival benefıt compared withdoxorubicin and cyclophosphamide: 7-year follow-up of US Oncologyresearch trial 9735. J Clin Oncol. 2009;27:1177-1183.

ERICA L. MAYER


Translating Genomic Research into Clinical Practice: Promiseand PitfallsJoseph A. Sparano, MD, Harry Ostrer, MD, and Paraic A. Kenny, PhD

OVERVIEW

Breast cancer is a heterogeneous disease associated with variable clinical outcomes despite standard local therapy for the primarytumor and systemic adjuvant therapy to prevent distant recurrence. Management decisions are typically made using classical prognosticand predictive clinicopathologic factors, and more recently gene expression profiling assays are commonly used in practice. Recentadvances in genomic sequencing—often referred to collectively as next-generation sequencing (NGS)—have facilitated more in-depthevaluation of the cancer genome than could be afforded by the initial generation of gene expression studies, including DNA singlenucleotide variants, small insertions and deletions, structural alterations, and copy number alterations (CNAs). In addition, thisinformation has been integrated with other molecular profiling methods of processes that affect gene transcription (e.g., epigenetic,microRNA) and protein expression—the ultimate readout of the genetic code. Although NGS has provided new insights on theclassification of breast cancer and identified potential predictive biomarkers and novel targets, there are formidable logistical andscientific obstacles that must be addressed before the promise of this technology is fully realized.

Genomics is defıned as the study of all of the nucleotidesequences in an organism. The original sequencing

methods relied on Sanger sequencing using synthesis withDNA polymerase and termination with dideoxynucleo-tides described in 1975, then modifıed in 1977 to be morerapid and accurate.1,2 These methods were used to sequencethe fırst human genomes that were reported in 2001.3,4Over the past decade, newer methods allow sequencingto be done more quickly, accurately, and cheaply, which areoften referred to as NGS.5 Such assays are now being used inthe clinic not only for research but also commercially forclinical use. Techniques are also available that allow high-throughput evaluation of the epigenome, microRNAs, andproteins and analytic approaches that integrate informationfrom multiple profıling methods. This review will addresshow NGS is being used for discovery-based research and itspotential clinical application in breast cancer.

FIRST GENERATION: GENE EXPRESSION PROFILINGUntil recently genomic profıling focused on the evaluation ofgene expression or the translation of information encoded ingenomic DNA into an RNA transcript.6 RNA transcripts in-clude mRNAs that are translated into proteins and variousother RNAs (e.g., transfer RNA, ribosomal RNA, microRNA, noncoding RNA) that have important biologic func-

tions. Perou et al fırst identifıed “intrinsic” breast cancer sub-types by evaluating variation in gene expression patternsusing hierarchical clustering in a set of 65 breast cancers from42 individuals using complementary DNA microarrays rep-resenting 8,102 human genes.7 The subtypes were recapitu-lated in other datasets and shown to be clinically relevantwith distinct clinical outcomes.8 It was hypothesized thatthese subtypes had distinctive gene expression profıles be-cause they originated from different cell types, including lu-minal epithelial cells (the cells that line the duct and giverise to the majority of breast cancers) and basal epithelialcells of the normalmammary gland (characterized by expres-sion of cytokeratins 5/6 and 17), hence the terms “luminal”and “basal” subtypes.More recent data indicate that the basaltumors arise from luminal progenitor cells rather than thebasal myoepithelial cells of the mammary gland.9,10 The in-trinsic gene panel was subsequently reduced to a panel of 50genes (called the PAM50) detectable by qRT-PCR, with 10genes selected for each centroid used to defıne four intrinsicsubtypes, including luminal A, luminal B, basal, and HER2-enriched (plus a “normal” subtype, which reflects an inade-quate biopsy specimen containing predominantly normalbreast tissue).11 The PAM50 and several multiparametergene expression assays have been approved for clinical use,including some that have been recommended by expert pan-els for clinical decision making.12

From the Departments of Medicine, Obstetrics, Gynecology and Women’s Health, Pathology, Genetics, Pediatrics, and Developmental & Molecular Biology, Albert Einstein College of Medicine,Bronx, NY; Montefiore Medical Center, Bronx, NY.


Corresponding author: Joseph A. Sparano, MD, Department of Oncology, Montefiore Medical Center, Albert Einstein College of Medicine, 1825 Eastchester Road, Bronx, NY 10461; email:[email protected].


NEXT-GENERATION SEQUENCING IN BREAST CANCER


NEXT GENERATION: BEYOND GENE EXPRESSIONNGS relies on the use of high-throughput, massively parallelsequencing and bioinformatic approaches to analyzemassivedatasets, including a variety of approaches for genome se-quencing and gene expression.When applied to RNA (RNA-Seq), NGS not only provides absolute expression levels (asthe precise number of transcripts can be counted) but alsoallows the identifıcation of alternatively spliced isoforms,mutant transcripts, and novel transcripts arising from fusiongenes. Terms commonly used in NGS studies are summa-rized in Table 1, and the processes typically used are illus-trated in Fig. 1.The technology platforms for NGS are developing at a

rapid rate, which is inexorably leading to faster, cheaper, andmore accurate sequencing data. Currently, the fıeld is domi-nated by machines from Roche, Illumina (HiSeq2000), andABI (SOLiD).5 Each of these platforms is capable of perform-ing whole genome sequencing, whole exome sequencing,RNA-Seq, and methylation analysis on a time scale betweenapproximately 10 hours (Roche) and 11 days (Illumina).These platforms are being supplemented by smaller ma-chines from the same manufacturers that may be run morequickly but handle smaller amounts of sequencing (i.e., notsuffıcient for whole genome analysis). Of the possible ap-proaches, whole genome sequencing that sequences all of theDNA base pairs in the genome is the most intensive. Wholeexome sequencing, in which libraries are generated from thetranscribed exons of the genome (encoding proteins, mi-croRNAs, and other RNAs), offers a less intensive approach,which still likely captures the majority of the interpretableinformation. RNA-Seq analysis sequences cDNA copies ofthe sample RNA component, providing detailed informationon gene expression levels, splicing, and mutations. DNAmethylation can also be analyzed using Methyl-Seq, provid-ing detailed coverage of the distribution of methylated CpGislands throughout the genome, which play key roles in con-trolling gene expression.Although the precise technical details of conducting these

assays vary between platforms, the principles are quite simi-lar in each case. Short, single-stranded DNA sequences aregenerated from the starting material, which can be genomicDNA (whole genome), hybridization-captured DNA exons(exome sequencing), cDNA (RNA-Seq) or bisulfıte-treatedDNA (Methyl-Seq). Small DNA adapters are ligated to eachstrand of the library, and the samples are then amplifıed toprovide enough representations of each individual strand forsequencing. Amplifıcation can be performed on beads (Illu-mina and SOLiD) or on a glass slide (Illumina). Each strandcan then be sequenced by sequential addition of the fourDNA bases (A, C, G, and T).In all cases, the output of the experiment is an extremely

large amount of rawDNA sequence reads (of between 60 and400 bp, depending on the platform), which must be pro-cessed further to be interpreted. This computationally inten-sive step requires mapping of each individual read to thecorrect location on the reference genome. A variety of com-mercial and open-source software packages are available toperform this function and subsequent steps of the analysis. Inthe case of cancer, the goal is usually to identify sequencevariants between the tumor sample and the reference ge-nome or nontumor DNA from the same patient, if available.Comparing normal and tumor DNA from the same individ-ual is particularly advantageous, as it excludes the influenceof the large number of single nucleotide polymorphisms,which differ between individuals in a species. Sequence alter-ations—which can include point mutations, insertions,deletions, or translocations—can be identifıed using theanalysis packages. Comparison of differences in mutationpatterns, amplifıcation/deletion, and translocation in indi-vidual tumor samples may be visually represented by Circosplots.13Because tumors have a high mutation rate, any sequencing

experiment will identify a large number of sequence varia-tions—only some of which make a contribution to the diseasepathology.Distinguishingcausative (driver) fromrandom(pas-senger)mutations canbechallenging forvariants andgenes thathave not been previously reported as playing a role in cancer.For suchnovelmutations, some insight canbegained fromsoft-ware tools such as SIFT, which can predict whether a particularchange in aprotein sequence is likely tohave a function-alteringeffect. Both the sequencing technology platforms and the rele-vant analysis strategies have beendescribed in detail elsewhere.5Whole genome and exome sequencing and the analysis of

the data generated are important research tools, but the sheervolume of data generated makes it diffıcult to effıciently de-ploy these approaches in a routine clinical setting. Instead, itis likely that targeted resequencing of small portions of thegenome, known to be frequently altered in cancer is likely toprovemore useful. Approaches such asAmpliSeq (Life Tech-nologies), which uses ultra-highmultiplex PCR to amplify allof the exons of 400 cancer-relevant genes from smallamounts of starting material, will generate clinically action-able data in the near future.

KEY POINTS

� Next-generation sequencing (NGS) refers to not onlyevaluation of gene expression, but also DNA singlenucleotide variants, small insertions and deletions,structural alterations, and copy number alterations.

� When applied to RNA, NGS provides information onexpression levels, alternatively spliced isoforms, mutanttranscripts, and novel transcripts arising from fusion genes.

� NGS has facilitated more refined disease classification andtarget discovery.

� PIK3CA and TP53 are the most commonly mutated genesthat vary in frequency by intrinsic subtype.

� Clinical application of NGS for therapeutic targeting iscurrently limited by a dearth of available drugs targetingactionable alterations.

SPARANO, OSTRER, AND KENNY


NGS IN BREAST CANCERAll cancers carry somatic mutations in their genome that aredominated by point mutations, amplifıcations, transloca-tions, and complex rearrangements.14 The genomic landscapeof breast cancer is characterized by a handful of commonlymutated gene “mountains” and a much larger number ofgene “hills” that aremutated at a low frequency.15 Drivermu-tations have been defıned in at least 40 genes in breast cancer,and somatic genomic rearrangements resulting in oncogenicfusion transcripts are also common—some of which havebeen postulated to play a key role in disease progression.16-18The results of selected NGS studies in breast cancer are sum-marized inTable 2 anddescribed below,which includes stud-ies correlating genomics with clinical outcomes for thepurpose of classifıcation or prediction, with specifıc breastcancer phenotypes for target discovery, or with genomics in-tegratedwith epigenomic,microRNA, and proteomic data togain a deeper understanding of how specifıc genes may beregulated and influence protein expression.

USE OF NGS FOR DISEASE CLASSIFICATION:METABRICThe Molecular Taxonomy of Breast Cancer InternationalConsortium (METABRIC) study was a joint effort by re-search teams from Canada and the United Kingdom togenomically classify breast tumors. Approximately 2,000clinically annotated, fresh-frozen breast cancer specimensfrom patients with operable breast cancer were assembledfrom tumor banks in the United Kingdom and Canada; thisincluded 997 tumors analyzed in a discovery group and 995in a validation group.19 Nearly all patients with estrogen re-ceptor (ER)-positive and/or lymph node–negative diseasedid not receive adjuvant chemotherapy, whereas patientswith either lymph node–positive or ER-negative disease did.Matched DNA and RNAwere extracted from each specimen

and subject to copy number and genotype analysis on the Af-fymetrix SNP 6.0 platform, transcriptional profıling on theIllumina HT-12 v3 platform, and TP53mutations by Sangersequencing in a cohort of 820 patients. This allowed evalua-tion of somatically acquired CNAs, germline copy numbervariants, and whether genomic variants acted in cis (impactsits own expression) or in trans (impacts expression of othergenes in genome, defıned as outside at least a contiguousthree megabase window).There were several important observations in this study.

First, a number of known driver mutations (e.g., ZNF704,PTEN,MYC, CCND1,MDM2, ERBB2, CCNE1) and putativedriver mutations (e.g., MDM1, MDM4, CDK3, CDK4,CAMK1D, PI4KB, NCOR1) were identifıed. There were alsoimportant deletions identifıed, including known deletions(e.g., PTEN) and novel deletions (e.g., PPP2R2A,MTAP, andMAP2K4). PPP2R2A deletions were noted in luminal B can-cers and have likewise been reported in clear cell and ovarianand endometrioid cancers. MTAP is often codeleted withCDKN2A and CDKN2B tumor suppressor genes in a varietyof cancers.MAP2K4mutationswere noted in ER-positive tu-mors and felt to be consistent with a tumor suppressor gene.A deletion event on chromosome 5 was also noted in basal-like tumors. Second, the effect of specifıc genes acting in transby influencing expression of distant sites of the genome iden-tifıed known aberration hot spots that could be grouped intopathwaymodules (e.g., ERBB2,MYC) and novel loci, includ-ing T-cell receptor loci on chromosome 7 (TRG) and 14 (TRA).These upregulate mRNAs and are highly enriched from T-cellactivationandproliferation,dendritic cell presentation, and leu-kocyte activation, thereby indicating an adaptive immune re-sponse associated with tumor-infıltrating lymphocytes. Third,joint clustering of copy number and gene expression data re-vealed that 10 integrative clusters characterized by well-defınedcopy number aberrations that split many of the intrinsic sub-types were associated with variable clinical outcomes and were

TABLE 1. Glossary of Terms Commonly Used in Next-Generation Sequencing Studies (Listed Alphabetically)

Term Definition

Circos plot A circular ideogram figure that visually represents differences in genomic structure between individual specimens

Codon A series of three adjacent bases in one polynucleotide chain of a DNA or RNA molecule, which codes for a specificamino acid

Copy number alterations Alterations in gene copy number, which are typically somatically acquired

Copy number variants Germ-line variation in the number of copies of a particular gene, which varies between individuals

Exome The part of the genome formed by exons; the exome of the human genome consists of roughly 180,000 exons constitutingapproximately 1% of the total genome or about 30 megabases of DNA

Exon Any nucleotide sequence encoded by a gene that remains present within the final mature RNA product of that gene afterintrons have been removed by RNA splicing; these include exons that are translated into protein and the untranslated regionflanking them, as well as exons encoding microRNAs and noncoding RNAs.

Indel Mutation resulting in a co-localized insertion or deletion and a net gain or loss in nucleotides

Intron Any nucleotide sequence within DNA or RNA that is not encoded into protein; introns are removed by RNA splicing while thefinal mature RNA product of a gene is being generated

Library A collection of DNA fragments prepared from a more complex sample

RNA-Seq Use of high-throughput sequencing technologies to sequence cDNA to get information about a sample’s RNA content

Single nucleotide variant DNA sequence variation when a single nucleotide (A, C, T, G) differs between members of a biological species



consistent in discovery and validation sets. This indicates het-erogeneity even within “intrinsic subtypes” revealed by inte-gration of information regarding somatic CNAs, includingsubgroupsofER-positive tumors thatwerehigh risk (character-ized by 11q13/14 cis-acting alterations), favorable risk (charac-terized by a paucity of copy number and cis-acting alterations),and intermediate risk (characterizedby17q23/20qand8p12cis-acting subgroups). In addition, this report demonstrated for thefırst time that genomic copy number loss at T-cell receptor locidrives a trans-acting immune response in an otherwise ge-nomically quiescent subgroup of ER-positive and ER-negative

tumors associatedwith a goodprognosis,whereas copynumberloss at 5q in a group of basal-like cancers drives a trans-actingtranscriptional control of genomicandchromosomal instabilityassociated with a poor prognosis.

USE OF NGS TO IDENTIFY PREDICTORS OFSENSISTIVITY AND RESISTANCE TO ENDOCRINETHERAPYAromatase inhibitors (AIs) are commonly used as adjuvantendocrine therapy in postmenopausal women with ER-

FIG 1. Summary of the steps involved in various next-generation sequencing approaches. (A-C) A tumor sampleis acquired and processed to purify either DNA (for genome/exome/methylation analysis) or RNA (for geneexpression analysis. (D) A library of short DNA fragments is prepared from the sample. For RNA-Seq analysis,the RNA from the sample is converted to cDNA. Some fragments may contain mutations (circles). (E) Sequencingreads are mapped to the reference human genome. Here, part of the reference sequence of the PIK3CA gene isshown, with several aligned reads. Some reads have an A to G point mutation at position 3140 of the PIK3CAcoding sequence. This mutant encodes the oncogenic H1047R PIK3CA mutation. (F) RNA-Seq analysis reads aremapped similarly to E. This approach provides data on gene expression level as well as gene mutations andfusions. (G) Sample output from mutationassessor.org showing an analysis of three epidermal growth factorreceptor mutations. Two of the mutations are predicted to have a neutral effect on protein function, while theG719S mutation is predicted to have a high effect.



positive disease. Response to preoperative AI therapy hasbeen shown to be a short-term surrogate reflecting favorableprognosis with AI therapy alone. Response may be assessedby an algorithm that reflects the extent of residual disease af-ter a 16-week course of therapy and Ki67 expression, or Ki67expression alone after a 2-week course of therapy.20 Ellis et alperformed whole genome analysis of 46 samples and exomesequencing in 31 additional samples from postmenopausalwomen with ER-positive breast cancers treated with pre-operative AI therapy, including 29 samples with Ki67 levelsabove 10% after a 16-week course of AI therapy (and thusconsidered to have AI-resistant disease) and 48 with KI67levels of 10%or less (indicatingAI-sensitive disease).21 Therewere several important observations from this study. First,the background mutation rate was about twofold higher forAI-resistant tumors than for sensitive tumors. Second, 18signifıcantly mutated genes were identifıed, including genespreviously identifıed in breast cancer (e.g., PIK3CA, TP53,GATA3, CDH1, RB1, MLL3, MAP3K1, CDKN1B) and novelgenes not previously observed, including fıve previouslyseen in hematopoietic cancers (e.g., RUNX1, CBFB, MYH9,MLL3, and SF3B1). Third, certain mutations were associatedwith specifıc subtypes, including an association betweenMAP3K1mutations and luminal A tumors and TP53muta-tions with luminal B tumors. Fourth, GATA3 mutationscorrelated with AI-induced suppression of proliferation andhence sensitivity to AI therapy.

INTEGRATING GENOMIC, EPIGENOMIC, ANDPROTIEN DATA: THE CANCER GENOME ATLAS(TCGA)TCGA included tumor and germ-line DNA samples ob-tained from 825 patients, which included evaluation by

Affymetrix 6.0 SNP arrays (733 patients) and Agilent mRNAexpression microarrays (547 patients), plus Illumina In-fınium DNA methylation chips (802 patients), miRNA se-quencing (697 patients), whole exome sequencing (507patients), and reverse-phase protein arrays (403 patients).22This included 466 tumors evaluated for gene expression,methylation, miRNA, and exome sequencing, and 348 whoalso had protein arrays. Correlation was not performed withclinical outcomes because of the short follow-up time (me-dian 17 months) and resultant small number of events.Nearly all genes previously implicated in breast cancer

were identifıed (PIK3CA, PTEN, AKT1, TP53, GATA3,CDH1, RB1, MLL2, MAP3K1, CDKN1B), plus a number ofnovel genes recently identifıed in other studies (TXB1,RUNX1, CBFB, AFF2, PIK3R1, PTPN22, PTPRD, NF1,SF3B1, CCND3). The overall mutation rate was lowest in theluminal A subtype and highest in the basal-like and HER2-enriched intrinsic subtypes. The most commonly mutatedgenes included TP53 (37%), PIK3CA (36%), GATA3 (11%),MAP3K1 (8%), MLL3 (7%), and CHD1 (7%), with 17 othermutations occurring in 1% to 4% (Fig. 2A). The distributionof mutations varied by subtype (Fig. 2B), with PIK3CA mu-tations occurring more commonly in luminal A/B andHER2-enriched than in basal subtypes (45%/29% and 39%vs. 9%), and TP53mutations dominating in basal (80%) andHER2-enriched subtype (72%) compared with luminal B(29%) and luminal A (12%) subtypes. This pattern is similarto other reports, with PIK3CA and TP53mutations predom-inating and consistently varying in frequency by subtype andwith othermutations relatively uncommon.The types ofmu-tations also differed by intrinsic subtypes, including differ-ences in TP53 mutations between basal-like (nonsense andframe shift) and luminal tumors (missense). Approximately9% of 507 cases evaluated revealed germ-line predisposing

TABLE 2. Summary of Selected Next-Generation Sequencing Studies in Breast Cancer

ReferenceNo.Patients

PatientPopulation

ClinicalOutcomes Profiling Methods Key Findings

Curtis et al22 1,992 All types Yes CNV 10 subtypes identified that correlate with clinical outcomes

CNA

Gene transcription Deletions in PPR2A, MTAP, MAP2K4 common

Ellis et al25 77 ER-positive Yes Whole genome sequencing 18 significantly mutated genes identified

Whole exome sequencing MAP3K1 associated with luminal A subtype

GATA3 mutations associated with response to neoadjuvantaromatase inhibitor therapy

TCGA26 825 All types No DNA copy number and CNA TP53 and PIK3CA mutations most common, with otherssubstantially less common

DNA methylation Luminal A subtype characterized by mutations in PIK3CA,GATA3, and MAPK3K1

mRNA expression Basal-like subtype similar to serous ovarian cancer

microRNA expression

Protein arrays

Shah et al28 104 Triple-negative No RNA-Seq Wide and continuous spectrum of genomic evolution

TP53, PIK3CA, PTEN mutations clonally dominant

Abbreviations: CNV, copy number variants; CNA, copy number alterations; ER, estrogen receptor; TCGA, The Cancer Genome Atlas.



variants (e.g., ATM, BRCA1, BRCA2, BRIP1, CHEK2, NBN,PTEN, RAD51C, TP53). Similar to other reports, copy num-ber changes correlated with some intrinsic subtypes, includ-ing loss of 5q and gain of 10p in basal-like cancers and gain of1q and/or 16q loss in luminal tumors.A unique feature of TCGA was the comprehensive nature

of the molecular profıling, which included evaluation of

miRNA, methylation, and proteins. Clustering analysis re-vealed seven subtypes by miRNA that did not correlate withmutational status or PAM50, with the exception of twomiRNA groups that showed overlap with basal-like sub-type and TP53 mutations. Five distinct DNA methylationgroups were identifıed, including at the extremes a hyper-methylated phenotype enriched for luminal B subtype and a

FIG 2. Distribution of genomic alterations observed in The Cancer Genome Atlas study in breast cancer,included all tumors (2a) and by intrinsic subtype the seven most common mutations occurring in at least 5%(2b).



hypomethylated phenotype that overlapped with basal-liketumors enriched for TP53 mutations. Protein analysis iden-tifıed seven subtypes that were highly concordant with themRNA intrinsic subtypes, especially basal-like and HER2-enriched subtypes. A multiplatform data matrix analysis re-vealed that the information content from copy numberaberrations, miRNAs, and methylation is captured at thelevel of gene expression and protein expression and activity.TCGA permitted a detailed analysis of each of the intrinsic

mRNA-defıned subtypes. Luminal tumors exhibited themost heterogeneous gene expression, mutational spectrum,and copy number changes and were associated with severalcharacteristics. First, althoughPIK3CAmutationswere com-mon in luminal tumors, markers typical of an activated PI3Kpathway (e.g., pAKT, pS6, p4EBP1 protein) were not present.In contrast, PIK3CAmutations were usually associated withpathway activation in basal and HER2-enriched tumors.Second, luminal tumors also were frequently associated withMAP3K1 and MAP2K4 mutations—two contiguous stepsin the p38-JNK pathway—whereas they occurred infre-quently in other subtypes. Third, in comparison to luminal Btumors, luminal A tumors were associated with higher fre-quency of intact TP53 and Rb1 tumor suppressor genes.HER2-enriched tumors showed high aneuploidy, somaticmutation rate, and DNA amplifıcation of other potential tar-gets (e.g., FGFR, EGFR, CDK4, cyclin D1). The basal subtypeshowed a high degree ofTP53mutations andhighPI3Kpath-way activity despite a low PI3K mutation rate (because ofPTEN and INPP4B loss and/or amplifıcation of PIK3CA).Similar to serous ovarian carcinoma, DNA repair defıcits(ATM mutations, BRCA1 and 2 inactivation, RB1 loss, andcycle E activation), genomic instability, and increased activ-ity of the HIF1-alpha/ARNT, MYC and FOXM1 pathwayswere also common.

GENOTYPIC ANALYSIS OF PHENOTYPICALLYDEFINED TRIPLE-NEGATIVE BREAST CANCERTheMETABRIC andTCGAstudies focused onusingNGS tosub-classify intrinsic breast cancer subtypes. However, phe-notypic classifıcation based on patterns of ER, progesteronereceptor, and HER2/neu expression is typically used in clin-ical practice for clinical decision making. Triple-negativebreast cancer (TNBC) is a phenotypic subset that accountsfor approximately 15% of all breast cancers, occurs morecommonly in youngerwomen andblack orHispanicwomen,and is characterized by a higher risk of relapse, earlier time torelapse, proclivity for recurrence in visceral organs and thecentral nervous system, and absence of specifıc targeted ther-apy.23 Shah et al described an analysis of 104 patients withTNBC subjected to RNA-Seq and deep resequencing mea-surements of allelic abundance for more than 2,400 somaticmutations.24 Approximately 20% of tumors had potentiallyclinically actionable somatic aberrations, including BRAFV600E, high-level EGFR amplifıcations, and ERBB2 andERBB3 mutations. The distribution of somatic mutation

abundance varied in a continuous distribution and wasunrelated to CNA or tumor cellularity. In another report,Banerji et al identifıed a recurrent MAGI3-AKT3 fusionin TNBC that led to constitutive activation of AKT kinasethat was abolished by a competitive AKT small-moleculeinhibitor.25

CLINICAL TRIALS INTEGRATING HIGH-THROUGHPUTGENOMIC ANALYSIS INTO CLINICAL CAREAndre et al tested the ability of array comparative genomichybridization (CGH) and Sanger sequencing to provide im-proved therapeutic direction in the SAFIR1 trial.26 Patientswithmetastatic breast cancer underwent biopsy ofmetastaticsites for genomic analysis, followed by genotype-directedtherapy after progression on standard therapy. At the timeof their initial preliminary report, of the 423 patients who

TABLE 3. Summary of Key Findings from Next-Generation Sequencing Studies in Breast Cancer

Disease Classification, Prognosis, and Prediction

Breast cancer is a heterogeneous disease with variable genomiccomplexity.

Gene expression classifies disease into “intrinsic” subtypes (luminal A,luminal B, HER2-enriched, basal).

Evaluation of somatically acquired CNAs and germ-line CNVs permit furthersub-classification.

Basal-like breast cancers are genotypically similar to serous ovariancarcinoma.

Some mutations are predictive of response to aromatase inhibitors(GATA3).

Target Discovery

Somatic mutations are common and dominated by point mutations,duplications, translocations, and rearrangements.

Rearrangements commonly result in fusion genes, some of which produceoncogenic proteins.

Most common currently targetable mutations are in the PI3K-mTOR-AKTpathway.

Not all PI3KCA mutations lead to pathway activation (especially in luminaldisease).

Promise

Next-generation sequencing is available for clinical use due to improvedmethodology and declining costs.

Refined disease classification may assist in guiding standard therapy.

Up to approximately 20% of genomic alterations may be potentiallyactionable targets.

Pitfalls

Few breast cancers are “addicted” to driver alterations.

Inactivation of tumor suppressor genes (e.g., TP53, RB1) are difficult totherapeutically target.

Preliminary clinical trials show that fewer than about 30% of screenedpatients receive a genomically directed therapy.

Logistical barriers include need for representative biopsy and currentdearth of available targeted agents.

Scientific barriers include tumor heterogeneity, intrinsic and acquiredresistance, and need for combinatorial therapy.

Abbreviations: can, copy number alterations; CNV, copy number variants.



consented, 59% eventually had their tumor sequenced, re-sulting in 26 patients (6% of consented group) who receiveda genome-directed treatment and eight patients (2% of con-sented group) having evidence of clinical benefıt from thetreatment. It is anticipated that up to 30% may eventuallyreceive a genome-directed therapy in this ongoing trial. Inthe SAFIR2 trial, patients with HER2-negative metastaticbreast cancer will undergo biopsy of a metastatic disease siteand if, after completing six to eight cycles of standard chemo-therapy, they are responding or showing stable disease with apotentially actionable genomic alteration, they will be ran-domly assigned to receive either standard care or a treatmenttargeted to the alteration.In addition to matching the right drug or combinations of

drugs with the right patient, a signifıcant limitation of usingNGS to direct therapy may be tumor heterogeneity from themutator phenotype and other factors contributing to heter-ogeneity and the emergence of drug resistance.27-28 For ex-ample, evaluation of the primary and metastatic sites in twopatients with renal cell carcinoma revealed that biopsy of asingle metastatic site was not representative of the genomiclandscape of all tumor sites and that tumor adaptation con-tributed to therapeutic failure.29

CONCLUSIONKey points about the knowledge gained thus far from apply-ing NGS to clinical breast cancer specimens are summarizedin Table 3. Potentially promising applications of NGS in-clude target discovery, refıned and more accurate diseaseclassifıcation, and improved therapeutic direction. However,there remain signifıcant pitfalls including logistical obstaclesto having an effective drug available to target every oncogenicalteration, and a regulatory and cancer care delivery systemthat would allow matching the right patient with the rightdrug, as exemplifıed by the SAFIR1 trial. In addition to thelogistical obstacles, other pitfalls include a dearth of tumorsthat are “addicted” to oncogenic mutations, innate tumorheterogeneity at presentation resulting in the need for com-binatorial therapy directed at multiple aberrant pathways,and the rapid emergence of resistance to therapy even whenappropriately applied.

ACKNOWLEDGMENTSupported in part by grants from the National Institutes ofHealth (P30–13330), Susan G. Komen for the Cure(KG100888), and the American Cancer Society (RSG-Tris-borate-EDTA [TBE]-123001).



Employment or Leadership Position: None. Consultant or Advisory Role: Joseph A. Sparano, Pfizer. Stock Ownership: None. Honoraria: Joseph A.Sparano, Bristol-Myers Squibb. Research Funding: Joseph A. Sparano, Merck. Expert Testimony: None. Other Remuneration: None.

References

1. Sanger F, Coulson AR. A rapid method for determining sequences inDNA by primed synthesis with DNA polymerase. J Mol Biol. 1975;94:441-448.

2. Sanger F, Nicklen S, Coulson AR. DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci U S A. 1977;74:5463-5467.

3. Lander ES, Linton LM, Birren B, et al. Initial sequencing and analysis ofthe human genome. Nature. 2001;409:860-921.

4. Venter JC, Adams MD, Myers EW, et al. The sequence of the humangenome. Science. 2001;291:1304-1351.

5. SoonWW,HariharanM, SnyderMP. High-throughput sequencing forbiology and medicine.Mol Syst Biol. 2013;9:640

6. Sparano JA, Fazzari M, Kenny PA. Clinical application of gene ex-pression profıling in breast cancer. Surg Oncol Clin N Am. 2010;19:581-606.

7. Perou CM, Sorlie T, Eisen MB, et al. Molecular portraits of humanbreast tumours. Nature. 2000;406:747-752.

8. Sorlie T, Tibshirani R, Parker J, et al. Repeated observation of breasttumor subtypes in independent gene expression data sets. Proc NatlAcad Sci U S A. 2003;100:8418-8423.

9. Gastaldi S, Sassi F, Accornero P, et al. Met signaling regulates growth,repopulating potential and basal cell-fate commitment of mammary

luminal progenitors: implications for basal-like breast cancer. Onco-gene. Epub 2012 May 7.

10. Molyneux G, Geyer FC,Magnay FA, et al. BRCA1 basal-like breast can-cers originate from luminal epithelial progenitors and not from basalstem cells. Cell Stem Cell. 2010;7:403-417.

11. Parker JS, Mullins M, Cheang MC, et al. Supervised risk predictor ofbreast cancer based on intrinsic subtypes. J Clin Oncol. 2009;27:1160-1167.

12. Harris L, Fritsche H, Mennel R, et al. American Society of Clinical On-cology 2007update of recommendations for the use of tumormarkers inbreast cancer. J Clin Oncol. 2007;25:5287-5312.

13. Krzywinski M, Schein J, Birol I, et al. Circos: an information aestheticfor comparative genomics. Genome Res. 2009;19:1639-1645.

14. Hillmer AM, Yao F, Inaki K, et al. Comprehensive long-span paired-end-tag mapping reveals characteristic patterns of structural variationsin epithelial cancer genomes. Genome Res. 2011;21:665-675.

15. Wood LD, Parsons DW, Jones S, et al. The genomic landscapes of hu-man breast and colorectal cancers. Science. 2007;318:1108-1113.

16. Stephens PJ, Tarpey PS, Davies H, et al. The landscape of cancergenes and mutational processes in breast cancer. Nature. 2012;486:400-404.



17. Stephens PJ, McBride DJ, LinML, et al. Complex landscapes of somaticrearrangement in human breast cancer genomes. Nature. 2009;462:1005-1010.

18. Inaki K, Hillmer AM, Ukil L, et al. Transcriptional consequences ofgenomic structural aberrations in breast cancer. Genome Res. 2011;21:676-687.

19. Curtis C, Shah SP, Chin SF, et al. The genomic and transcriptomic ar-chitecture of 2,000 breast tumours reveals novel subgroups. Nature.2012;486:346-352.

20. Ellis MJ, Suman VJ, Hoog J, et al. Randomized phase II neoadjuvantcomparison between letrozole, anastrozole, and exemestane for post-menopausal women with estrogen receptor-rich stage 2 to 3 breast can-cer: clinical and biomarker outcomes and predictive value of thebaseline PAM50-based intrinsic subtype-ACOSOGZ1031. J ClinOncol.2011;29:2342-2349.

21. Ellis MJ, Ding L, Shen D, et al. Whole-genome analysis informsbreast cancer response to aromatase inhibition. Nature. 2012;486:353-360.

22. Cancer Genome Atlas Network. Comprehensive molecular portraits ofhuman breast tumours. Nature. 2012;490:61-70.

23. Foulkes WD, Smith IE, Reis-Filho JS. Triple-negative breast cancer.N Engl J Med. 2010;363:1938-1948.

24. Shah SP, Roth A, Goya R, et al. The clonal and mutational evolutionspectrum of primary triple-negative breast cancers. Nature. 2012;486:395-399.

25. Banerji S, Cibulskis K, Rangel-Escareno C, et al. Sequence analysis ofmutations and translocations across breast cancer subtypes. Nature.2012;486:405-409.

26. Andre F, Bachelot T, CamponeM, et al. Array CGH andDNA sequenc-ing to personalize therapy for metastatic breast cancer: a prospectiveNational trial (UNICANCER SAFIR-01). European Society of MedicalOncology. 2012;23(suppl 9):7-30 (abstr 3426).

27. Loeb LA. Human cancers express mutator phenotypes: origin, conse-quences and targeting. Nat Rev Cancer. 2011;11:450-457.

28. NgCK, PembertonHN, Reis-Filho JS. Breast cancer intratumor geneticheterogeneity: Causes and implications. Expert Rev Anticancer Ther.2012;12:1021-1032.

29. GerlingerM, RowanAJ, Horswell S, et al. Intratumor heterogeneity andbranched evolution revealed by multiregion sequencing. N Engl J Med.2012;366:883-892.



BREAST CANCER

Beyond Trastuzumab andLapatinib: New Options for HER2-Positive Breast Cancer

CHAIRDavid Cameron, FRCPThe University of EdinburghEdinburgh, United Kingdom

SPEAKERSIan Krop, MD, PhDDana-Farber Cancer InstituteBoston, MA

Martine Piccart-Gebhart, MD, PhDJules Bordet InstituteBrussels, Belgium

Beyond Trastuzumab and Lapatinib: New Options forHER2-Positive Breast CancerDimitrios Zardavas, MD, David Cameron, MD, Ian Krop, MD, PhD, and Martine Piccart, MD, PhD

OVERVIEW

HER2-positive breast cancer (BC) constitutes a molecular subtype of the disease with an aggressive biologic behavior. Trastuzumabrevolutionized the treatment of this disease, changing its natural history. Lapatinib is active in the metastatic setting, approved forpatients who were pretreated with trastuzumab. However, resistance to anti-HER2 agents is a major clinical issue, occurring in bothearly-stage and advanced disease, and new treatment options are clearly needed. An abundance of HER2-targeted agents are beingclinically developed: monoclonal antibodies, small molecule inhibitors, and antibody drug conjugates (ADC). Combining HER2-targetedagents in regimens of dual HER2 blockade has already reached clinical practice in the metastatic setting, confirming the preclinicalefficacy of enhanced HER2 inhibition. Promising results have been generated in the neoadjuvant setting, and large randomized trialsare seeking evidence for dual HER2 blockade in the adjuvant setting. ADC represent another hope for improved treatment outcomesof HER2-positive BC, as exemplified by the positive results of clinical trials employing trastuzumab-DM1 (trastuzumab emtansine, T-DM1).Moreover, an understanding of the molecular mechanisms mediating resistance to HER2 blockade has opened new therapeutic avenues,with several targeted agents entering clinical trials. This paper presents the clinical data of the HER2-targeted agents underdevelopment, as well as an overview of the biologic rationale for the development of agents aimed at circumventing anti-HER2resistance.

Epidermal growth factor receptor 2 (EGFR2; HER2/neu,ErbB2), a type II transmembrane receptor, constitutes a

well-studied oncogene,mediating a plethora of oncogenic ef-fects, through interlinked activation of PI3K/Akt/mTORandRas/Raf/MEK/ERK signaling pathways.1 Amplifıcation ofthe gene and/or protein overexpression is detected in ap-proximately 20% of patients with BC and confers an aggres-sive biologic phenotype, resulting in poor clinical outcome.2The development of trastuzumab, a humanized monoclo-

nal antibody targeting domain IV of HER2, represents thefırst signifıcant milestone in the development of anti-HER2-targeted agents. Trastuzumab is themainstay ofHER2 block-ade in all stages of HER2-positive disease, changing itsnatural history.3 Lapatinib constitutes another important de-velopment in the arena of HER2-blocking agents, being adual EGFR/HER2 reversible tyrosine kinase inhibitor (TKI),currently approved in combination with either capecitabineor letrozole for patients with HER2-positive metastatic BC(MBC).3Despite the antitumor activity of those two agents, a subset

of patients with early stage HER2-positive BC unfortunatelyrelapses; in the metastatic setting, primary or secondary re-sistance develops inevitably. This clinical reality mandatesnew treatment options, with a plethora of HER2-targeted

agents showing differentmechanisms of action (Table 1) cur-rently under clinical development (Tables 2 and 3). Onepromising approach is dual HER2-blockade. Combined ad-ministration of different HER2-targeted agents showsheightened antitumor activity, based on the complementar-ity of their mechanisms of action, leading to more completeHER2 blockade. This strategy has already reached clinicalpractice in the metastatic setting and has demonstratedhigher rates of pathologic complete response (pCR) in theneoadjuvant setting (Fig. 1). A second approach is the devel-opment of anti-HER2 ADC, as exemplifıed by the develop-ment of trastuzumab-DM1. A third approach, which isearlier in clinical development, derives from the elucidationof the molecular mechanisms mediating resistance to HER2blockade. This strategy combines HER2 blockade with acompound blocking a molecular mediator of anti-HER2 re-sistance, with the goal of “(re)sensitizing” HER2-positive BCcells to HER2 inhibition (Table 4).

TRASTUZUMAB AND LAPATINIBTrastuzumab and lapatinib exhibit complementary mecha-nisms of HER2 blockade (Table 1), and preclinical studieshave shown a synergy between these two agents: increased

From the Institut Jules Bordet, Universite Libre de Bruxelles, Brussels, Belgium; Edinburgh Cancer Research UK Centre, University of Edinburgh, Edinburgh, United Kingdom; and Dana-FarberCancer Institute, Harvard Medical School, Boston, MA.


Corresponding author: Martine Piccart, MD, PhD, Institut Jules Bordet, Boulevard de Waterloo 125B-1000, Brussels, Belgium; email: [email protected].


ZARDAVAS ET AL

e2 2013 ASCO EDUCATIONAL BOOK | asco.org/edbook

antibody-dependent cellular cytotoxicity (ADCC) and en-hanced induction of apoptosis in xenograft tumor modelshave been reported.4,5In the metastatic setting, the combination of trastuzumab

and lapatinib was assessed in the phase III EGF104900 study,which randomly selected patients who were pretreated withtrastuzumab to receive lapatinib alone or lapatinib combinedwith trastuzumab. The dual HER2 blockade resulted in a sig-nifıcant prolongation of median progression-free survival(mPFS; 11.1 week vs. 8.1 week, p � 0.011) and of overall sur-vival (OS; 14 months vs. 9.5 months, p � 0.026), despite sig-nifıcant cross-over (52%).6 The toxicity of the lapatinib/trastuzumab regimen was acceptable, with increased rates ofgrade 1 and 2 diarrhea (p � 0.03) and a higher frequency of20% or greater reduction in the left ventricular ejection frac-tion (LVEF; 5.4 vs. 2.1%), as compared to lapatinib mono-therapy.In the neoadjuvant setting, Neo-Adjuvant Lapatinib

and/or Trastuzumab Treatment Optimization trial (Neo-ALTTO) was a phase III study that randomly selected 455

patients with HER2-positive operable BC to receive lapa-tinib, trastuzumab, or their combination for a total of 6weeks. After this period, the assigned HER2-blocking strat-egy was combined with paclitaxel for further 12 weeks. Thedual HER2-blockade resulted in a signifıcant improvementin the pCR rate, outperforming the trastuzumab and lapa-tinib arms (pCR 51.3% vs. 29.5% vs. 24.7%, respectively; p �0.0001 for the trastuzumab-lapatinib arm versus trastu-zumab alone).7 The pCR rates were more pronounced in thepatients with hormone receptor–negative disease. No majorcardiac dysfunction was noted, and the lapatinib-containingarms were associated withmore grade 3 and 4 adverse events(AEs) and higher rates of treatment discontinuation.Similar results were reported by the Chemotherapy, Her-

ceptin, and Lapatinib in Operable Breast Cancer study(CHER-LOB).8 In this phase II study, 121 patients with op-erable HER2-positive BC were allocated to receive neoadju-vant taxane-anthracycline chemotherapy combined withlapatinib (Arm A), trastuzumab (Arm B), or lapatinib plustrastuzumab (Arm C). The lapatinib/trastuzumab combina-tion was associated with higher pCR rates (48%) than thelapatinib (pCR 28%) and trastuzumab arms (pCR 32%). Athird neoadjuvant study—the Translational Breast CancerResearch Consortium (TBCRC 006)—assessed the lapatinib/trastuzumab combination in a chemotherapy-free regimenadministered for 12 weeks (plus hormone therapy for pa-tients with hormone receptor–positive BC).9 In this phase IIstudy, 64 patients diagnosed with HER2-positive BC with amedian tumor size of 6 cmachieved pCR in 28%overall; onceagain hormone receptor negativity was associated with in-creased pCR rates (42% vs. 21% for patients with estrogenreceptor (ER)–negative and ER-positive BC, respectively).The combined administration of trastuzumab and lapa-

tinib is currently under clinical investigation in the adjuvantsetting. The ALTTO study (NCT00490139)—a phase IIIstudy with a completed accrual exceeding 8,300 patients—compared the activity of lapatinib monotherapy, trastu-zumab monotherapy, trastuzumab followed by lapatinib,and trastuzumab concomitantly with lapatinib in patientswith early-stageHER2-positive BC. The primary endpoint of

KEY POINTS

� Despite the efficacy of trastuzumab in the adjuvant setting,there is a subset of patients developing recurrence of thedisease. In the metastatic setting, resistance inevitablyoccurs.

� Currently, an abundance of HER2-targeted agents areunder clinical development, namely: monoclonal antibodies,small molecule inhibitors, and antibody drug conjugates.

� The concept of dual HER2 blockade has already made itsway into clinical practice with the approval of pertuzumabin combination with trastuzumab and docetaxel in the firstline of metastatic disease.

� Trastuzumab-DM1, an anti-HER2 antibody drug conjugate, isan active agent currently in advanced clinical testing.

� Several molecular components of resistance mechanisms toHER2 blockade are being pursued as therapeutic targetswith experimental compounds under development.

TABLE 1. Mechanisms of Action Anti-HER2 Blocking Agents

Class of Compounds Agent(s) Mechanism of Action

Monoclonal antibody Trastuzumab � Inhibition of ligand-independent HER2/3 dimerization� Inhibition of HER2-mediated intracellular signaling� Induction of ADCC

Pertuzumab � Inhibition of ligand-dependent HER2/3 dimerization� Inhibition of HER2-mediated intracellular signaling� Induction of ADCC

Antibody drug conjugates Trastuzumab-DM1 � Trastuzumab’s action� Selective delivery of the antimicrotubule agent DM1

Small-molecule inhibitors Lapatinib � Kinase activity (reversible) inhibition of full-length HER2� Kinase activity (reversible) inhibition of p95HER2

Afatinib, neratinib � Kinase activity (irreversible) inhibition of EGFR/HER2/HER4� Kinase activity inhibition of mutated HER2

Abbreviations: ADCC, antibody-dependent cellular cytotoxicity; p95HER2, carboxy terminal fragments of HER2; EGFR, epidermal growth factor receptor.

TRASTUZUMAB, LAPATINIB, AND HER2

asco.org/edbook | 2013 ASCO EDUCATIONAL BOOK e3

TABLE 2. Selected Ongoing Trials with T-DM1 and Pertuzumab in HER2-Positive Breast Cancer

Trial (NCT Identifier) Phase Setting Treatment Primary Endpoint Secondary Endpoints

T-DM1

NCT00951665 I Trastuzumab pretreated T-DM1 � pertuzumab � paclitaxel AEs, DLTs, PK ORR, PFS, CBR, duration of response

NCT00679341 II Trastuzumab naive T-DM1 versus trastuzumab � docetaxel PFS OS, ORR, CBR, duration of OR, TTSP

NCT01745965 II Neoadjuvant T-DM1 versus T-DM1 � hormonotherapyversus trastuzumab � hormonotherapy

pCR OS, cardiac safety, toxicity, HRQL

TH3RESA(NCT01419197)

III Trastuzumab retreated T-DM1 versus physician’s choice PFS, OS ORR, duration of OR, AEs, HRQL, PKof T-DM1

NCT01745965 II Neoadjuvant T-DM1 versus T-DM1 � HT versustrastuxumab � HT

pCR OS, toxicity, safety, HRQL

NCT01772472 III Adjuvant T-DM1 versus trastuzumab IDFS DFS, OS, DRFI, safety, PRO

Pertuzumab

NCT00934856 I Trastuzumab pretreated Pertuzumab � T-DM1 � docetaxel DLTs, AEs PFS, ORR, CBR, duration of OR, TTF,PK

NCT01565083 II Trastuzumab naive Pertuzumab � trastuzumab �vinorelbine

ORR TTR, duration of OR, PFS, TTP, OS,cardiac toxicity, HRQL

NCT01491737 II Trastuzumab naive Pertuzumab � Trastuzumab � AIversus trastuzumab � AI

PFS OS, ORR, CBR, duration of OR, TTR,safety, HRQL

PHEREXA(NCT01026142)

II Trastuzumab pretreated Trastuzumab � capecitabine �pertuzumab versus trastuzumab �capecitabine

PFS TTP, TTF, ORR, CBR, duration of OR,safety

PERUSE(NCT01572038)

III Trastuzumab naive Pertuzumab � trastuzumab � taxane Safety PFS, OS, ORR, CBR, duration of OR,TTR, HRQL

APHINITY(NCT01358877)

III Adjuvant Chemotherapy � trastuzumab �pertuzumab versus chemotherapy �trastuzumab

IDFS DFS, OS, RFI, DRFI, cardiac safety,safety, HRQL

Abbreviations: NCT, National Clinical Trial; T-DM1, trastuzumab-DM1; AE, adverse event; DLT, dose-limiting toxicity; PK, pharmacokinetics; ORR, overall response rate; PFS, progression-free survival;CBR, clinical benefit rate; OS, overall survival; OR, objective response; TTSP, time to symptom progression; HRQL, health-related quality; HT, hormonal therapy; IDFS, invasive disease-free survival;pCR, pathologic complete response; DFS, disease-free survival; DRFI, distance recurrence-free interval; PRO, patient-reported outcome; AI, aromatase inhibitor; TTF, time to treatment failure; RFI,recurrence-free interval.

TABLE 3. Selected Ongoing Clinical Trials with HER-Family Small Molecule Inhibitors


Afatinib

NCT00950742 I Trastuzumabpretreated

Afatinib � trastuzumab MTD AEs, PK, ORR, PFS

LUX-Breast 1(NCT01125566)

III Trastuzumabpretreated

Afatinib � vinorelbine versus Trastuzumab� Vinorelbine

PFS OS, TTD, HRQL, Safety, PK

LUX-Breast-2(NCT01271725)

II Trastuzumabpretreated

Afatinib ORR Duration of OR, PFS, safety

LUX-Breast 3(NCT01441596)

II Patients with CNS metstrastuzumab and/orlapatinib pretreated

Afatinib versus afatinib � vinorelbineversus investigator’s choice oftreatment

Patient benefit at12 wk

PFS, OS

DAFNE (NCT01594177) II Neoadjuvant Afatinib � trastuzumab � NAC pCR AEs, CR, TR

Neratinib

NCT01423123 I Trastuzumabpretreated

Neratinib � trastuzumab � paclitaxel Safety, DLTs ORS, PFI

NCT01008150 II Neoadjuvant Trastuzumab � NAC versus neratinib �NAC versus neratinib � trastuzumab �NAC

pCR CCR, RFI, OS, toxicity

NCT01494662 II Patients with CNS mets Neratinib ORR PFS, OS, CNS response, safety,association of CTC and OS

NEFERTT(NCT00915018)

II Trastuzumab naive Neratinib � paclitaxel versus trastuzumab� paclitaxel

PFS OS, ORR, duration of OR, CBR,AEs, HRQL, time to CNSmets

Abbreviations: NCT, National Clinical Trial; MTD, maximum tolerated dose; AE, adverse event; PK, pharmacokinetics; ORR, overall response rate; CBR, clinical benefit rate; PFS, progression-free survival; OS,overall survival; TTD, time to deterioration; HRQL, health-related quality; OR, objective response; CNS, central nervous system; mets, metastases; NAC, neoadjuvant chemotherapy; pCR, pathologic completeresponse; CR, conservation rate; TR, translational research; CCR, clinical complete response; DLT, dose-limiting toxicity; PFI, progression-free interval; RFI, recurrence-free interval; CTC, circulating tumorcell.

ZARDAVAS ET AL


FIG 1. pCR rates in trials assessing dual HER2 blockade.Abbreviations: pCR, pathologic complete response; Anthra, anthracycline; C, carboplatin; D, docetaxel; FEC, 5-fluorouracil, epirubicin, and cyclophosphamide; L, lapatinib; P, pertuzumab; pac,paclitaxel; T, trastuzumab; tax, taxane.

TABLE 4. Selected Ongoing Trials with Targeted Agents in HER2 Breast Cancer


Everolimus

NCT01305941 II CNS Mets Everolimus � trastuzumab �vinorelbine

CNS RR ORR, toxicity

BOLERO-3(NCT01007942)

III Trastuzumab pretreated Everolimus � trastuzumab vinorelbineversus trastuzumab � vinorelbine

PFS OS, ORR, PRO, CBR

BOLERO-1(NCT00876395)

III Trastuzumab naive Everolimus � trastuzumab �paclitaxel versus trastuzumab �paclitaxel

PFS OS, ORR, AEs, TTOR, OS,CBR, PK

Other PI3K-Blocking agents

NCT01285466 I Trastuzumab pretreated BEZ235 or BKM120 � paclitaxel �trastuzumab

DLTs AEs, ORR, PK

NCT01042925 I/II Trastuzumab pretreated XL147 � trastuzumab � (paclitaxel) MTD, OTR PFS, PK, PD

NCT01132664 I Trastuzumab pretreated BKM120 � trastuzumab AEs, DLT ORR

PIKHER2(NCT01589861)

I/II Trastuzumab pretreated(with PI3K activation)

BKM120 � lapatinib MTD, ORR Safety, CBR, PFS, PK

IGF Blocking agents

NCT00684983 II Trastuzumab pretreated Cixutumumab � lapatinib �capecitabine versus lapatinib �capecitabine

PFS OS, TTF, ORR, HRQL, TR

SRC Blocking Agents

NCT01306942 I/II Trastuzumab pretreated Dasatinib � trastuzumab � paclitaxel MTD Safety, CBR, PFS, RD

HER3 Blocking Agents

NCT01512199 I/II Trastuzumab naive U3–1287 � trastuzumab � paclitaxel MTD, PFS PK, DCR

Abbreviations: NCT, National Clinical Trial; CNS, central nervous system; mets, metastases; RR, response rate; ORR, objective response rate; PFS, progression-free survival; OS, overall survival;PRO, patient-reported outcome; CBR, clinical benefit rate; AE, adverse event; TTOR, time to overall response; PK, pharmacokinetics; DLT, dose-limiting toxicity; MTD, maximum tolerated dose; OTR,objective tumor response; PD, pharmacodynamic; TTF, time to treatment failure; HRQL, health-related quality; TR, translational research; RD, response duration; DCR, disease control rate.



this study is the comparison of disease-free survival (DFS)between the trastuzumab-alone and the combination arms.Importantly, the fırst interim analysis of the ALTTO studyresulted in the closure of the lapatinib-alone arm, since it wasdeemed unlikely to cross the boundary of noninferiority.10

PERTUZUMABPertuzumab is a humanized monoclonal antibody, bindingto the dimerization domain II ofHER2, thus blocking ligand-induced HER2/HER3 heterodimerization.11 This dimeriza-tion represents an important molecular event drivingmalignant progression of HER2-positive BC cells through itspotent ability to activate the PI3K/Akt/mTOR signalingpathway. Preclinical evidence showed a synergistic effect ofpertuzumab combined with trastuzumab, so that clinical tri-als assessing this dual HER2 blockade strategy followed.12Several phase I studies have been conducted in the meta-

static setting, assessing pertuzumab either as monotherapyor in combination with cytotoxic agents. A small phase IItrial assessed the combination of pertuzumab and trastu-zumab in 11 patients who were pretreated with trastuzumaband had hadHER2-positiveMBC.13 An overall response rate(ORR) of 18% and a clinical benefıt rate (CBR) of 44.5%werereported; however, the accrual to this study was prematurelyterminated because of cardiac toxicities. Six patients devel-oped an LVEF decrease, with two exceeding a reduction of15%, and one of them developed symptomatic congestiveheart failure. Another phase II study assessed the combina-tion of pertuzumab plus continued trastuzumab therapy in66 patients with HER2-positive MBC progressing ontrastuzumab-based therapy, showing promising antitumoractivity: ORR reached 24.2% and CBRwas 50%, with fıve pa-tients achieving a complete response (CR; 7.6%) and 11 pa-tients achieving a partial response (PR; 16.7%).14 The cardiacsafety profıle was favorable in this trial in comparison to thesmaller one, with no patient withdrawing from the study as aresult of cardiac toxicity. The cardiac safety of pertuzumabwas proved by a pooled analysis study of 598 patients whoreceived pertuzumab either as monotherapy or in combina-tion with other compounds, showing that pertuzumab didnot result in a notable increase in cardiac-related side ef-fects.15Clinical Evaluation of Pertuzumab and Trastuzumab

(CLEOPATRA; NCT00567190) is a randomized, phase IIIclinical trial that assigned 808 previously untreated patientswith HER2-positive MBC to receive either docetaxel plustrastuzumab or the same regimen combined with pertu-zumab.16 The dual HER2 blockade resulted in a signifıcantprolongation of independently assessed progression-freesurvival (PFS), the trial’s primary endpoint, as comparedwith the standard of care (mPFS 18.5 vs. 12.4 months, HR0.65; 95% CI, 0.51 to 0.75, p � 0.001). An interim analysis ofOS showed a trend favoring the pertuzumab-containing arm,which did not cross the stringent predefıned p value (HR0.64; 95% CI, 0.47 to 0.88, p � 0.005), with few reported OSevents. A secondOS analysis performedwith 69% of planned

events for the fınal analysis showed a statistical signifıcancefavoring the pertuzumab-containing arm (mOS not reachedcompared with 37.6 months in the placebo arm; HR 0.66;95% CI, 0.52–0.84, p � 0.0008).17 ORR reached 69.3% in thecontrol group and 80.2% in the pertuzumab group. Of note,only 10% of the patients enrolled in CLEOPATRA had re-ceived trastuzumab as part of a (neo)adjuvant regimen.How-ever, exploratory analysis revealed no difference in effıcacy ofthe pertuzumab-containing arm according to prior trastu-zumab exposure. In terms of toxicity, diarrhea, rash,mucosalinflammation, febrile neutropenia, and dry skin of any gradewere reported more frequently in the pertuzumab-containing arm. The cardiac safety profıles between the twoarms were comparable, with left ventricular systolic dysfunc-tion grade 3 and greater reported in 2.8%of the patients in thecontrol group and in 1.2% of the patients in the pertuzumab-group. On the basis of these results, pertuzumab is approvedfor use in the fırst-line setting of HER2-positive MBC, incombination with trastuzumab and docetaxel.Pertuzumab has been also assessed in combination with

trastuzumab in the neoadjuvant setting. The phase II studyNeoadjuvant Study of Pertuzumab and Herceptin in EarlyRegimen Evaluation (NeoSphere; NCT00545688) randomlyselected 417 patients to receive four cycles (12 weeks) of oneof the following four preoperative regimens: trastuzumabplus pertuzumab plus docetaxel; trastuzumab plus docetaxel;pertuzumab plus docetaxel; or trastuzumab plus pertu-zumab.18 The pCR rates, the primary endpoint of this trial,were 45.8% for the triplet, 29% for the trastuzumab plus do-cetaxel, 24% for pertuzumab plus docetaxel, and 16.8% forthe chemotherapy-free pertuzumab plus trastuzumab com-bination. In all arms of the study, the rates of pCR werehigher in the subgroup of women with ER-negative disease,with a pCR rate of 27.3% for women receiving thechemotherapy-free combination of pertuzumab plus trastu-zumab. Importantly, this suggests that a proportion ofwomen with HER2-positive disease could be cured withoutcytotoxic chemotherapy. TRYPHAENA (NCT00976989) isanother phase II study that has been reported in the neoad-juvant setting. It randomly selected 225 patients to receiveone of the three following arms: pertuzumab plus trastu-zumab given concurrently with an anthracycline-taxane reg-imen (5-fluorouracil, epirubicin, cyclophosphamide [FEC]followed by docetaxel) or the dual HER2-combination givensequentially to FEC, but concurrently with docetaxel, or thisanti-HER2 doublet given concurrently with carboplatin anddocetaxel.19 The primary endpoint of this trial was cardiacsafety, which was comparable across all three tested arms.pCRwas 61.6% for arm 1, 57% for arm 2, and 66.2% in arm 3,with ER negativity associated with higher pCR rates (79.4%,65.0%, and 83.8% for arms 1, 2, and 3, respectively).The aforementioned clinical effıcacy data led to the devel-

opment of a phase III prospective multicenter clinical trialassessing the effıcacy of pertuzumab in the adjuvant setting:Adjuvant Pertuzumab and Herceptin in Initial Therapy ofBreast Cancer, BIG 4–11 (APHINITY; NCT01358877) is anongoing international collaborative study, aiming to ran-

ZARDAVAS ET AL


domly selected approximately 4,800 patients with early-stageHER2-positive BC to receive either trastuzumab or trastu-zumab with pertuzumab in conjunction with cytotoxic che-motherapy.

HER FAMILY SMALL MOLECULE INHIBITORSNeratinib is an oral, irreversible pan-HER TKI of EGFR/HER1, HER2, and HER4 under clinical development. Aphase II study assessed neratinib in patients with HER2-positive MBC who had either previously received trastu-zumab (66 patients) or not (70 patients).20 Neratinib showedsubstantial antitumor activity with the 16-week PFS rates be-ing 59% for patients who has been pretreated with trastu-zumab and 78% for patients who were trastuzumab-naiveand median PFS reaching 22.3 and 39.6 weeks, respectively.The most common AEs were diarrhea, which was the mostfrequent grade 3 to 4 AE, nausea, vomiting, and fatigue.No neratinib-related cardiotoxicity of grade 3 or 4 was re-ported. Neratinib has also been evaluated in combinationwith trastuzumab in a phase I/II study of 45 patients withtrastuzumab-pretreated HER2-positive MBC.21 Promisingantitumor activity was noted, with a 16-week PFS rate of 47%and a mPFS of 19 weeks. No new toxicity signals were ob-served. Preliminary results of a phase I study assessing a trip-let of weekly paclitaxel with neratinib and trastuzumab inwomen with heavily pretreated HER2-positive MBC havebeen presented; out of 10 patients for whom effıcacy datawere available, there was one CR and three PR, with diarrheabeing the most frequent grade 3 and 4 AE.22 Another phase Istudy assessed the feasibility of neratinib combined witheverolimus, with signs of antitumor activity: out of eight pa-tients with heavily pretreated HER2-positive MBC, fourshowed a PR and one had SD for an RR of 67%.23 A random-ized phase II study compared neratinib with lapatinib pluscapecitabine in the second line of metastatic disease. Nera-tinib demonstrated high antitumor activity (ORR 29%);however, it was inferior to the lapatinib/capecitabine arm interms of mPFS, the primary endpoint of the trial (mPFS 4.5vs. 6.8months,HR1.3; 95%CI, 1.0 to 1.8).24 A phase III studyis currently underway evaluating neratinib in patients withearly-stage HER2-postive BC after completion of standardtrastuzumab based adjuvant chemotherapy (NCT00878709).Afatinib is another promising orally available irrever-

sible pan-HER TKI under clinical development. A phase IIsingle-arm study assessed its activity in 41 patients withtrastuzumab-pretreated HER2-positiveMBC, with amedianof three prior chemotherapy lines (range, 0–15).25 Promisingclinical activity was shown, with four patients (10%) havingPR, 15 patients (37%) having stable disease as best responseand mPFS reaching 15.1 week (95% CI, 8.1 to 16.7 weeks).Themost common grade 3 AEs were diarrhea and rash. Cur-rently, there are ongoing trials assessing afatinib in HER2-positive BC (Table 3), with a phase III trial (LUX-Breast 1,NCT01125566) randomly selecting patients with HER2-positive MBC to receive afatinib plus vinorelbine or trastu-zumab plus vinorelbine as fırst- or second-line treatment.

T-DM1T-DM1 is an ADC consisting of DM1, a maytansinoid anti-microtubule agent, bound to trastuzumab through nonre-ducible thioether bonds. T-DM1 delivers this highly potentcytotoxic agent specifıcally to HER2-expressing cells. OnceT-DM1 binds toHER2 on the cell surface, the T-DM1-HER2complex is internalized and the antibody component is pro-teolytically degraded, releasing the DM1 into the cyto-plasm.26 Importantly, T-DM1 retains the biologic activity oftrastuzumab (i.e., HER2 signaling blockade and induction ofADCC).27 These properties provided the biologic rationalefor testing T-DM1 in the setting of HER2-positive BC.A phase I study conducted in 24 patients with advanced,

heavily pretreated (median, four prior chemotherapeuticagents for metastatic disease) HER2-positive BC assessed thesafety and tolerability of T-DM1 in three-weekly adminis-tered ascending doses.28 The maximum tolerated dose(MTD) of T-DM1was 3.6 mg/kg, with transient thrombocy-topenia being a dose-limiting toxicity. The toxicity profılewas favorable with common drug-related AEs being grade 2and lower thrombocytopenia, elevated transaminases, fa-tigue, nausea, and anemia. Preliminary signs of antitumor ac-tivity were generated, with a CBR (defıned as the objectiveresponse plus stable disease at six months) among the 15 pa-tients treated at the MTD reaching 73%. Two consecutivephase II single-arm studies of T-DM1 in HER2-positiveMBC have been reported. The fırst one reported an ORR of25.9% (95% CI, 18.4% to 34.4%) by independent assessmentamong 112 heavily pretreated patients.29 There were not suf-fıcient events to reach median duration of response, andmPFS was 4.6 months (95% CI, 3.9 to 8.6 months). Anotherphase II study assessed T-DM1monotherapy in 110 patientswith HER2-positive MBC, who were previously treated withtrastuzumab, lapatinib, an anthracycline, a taxane, and cape-citabine (median, seven prior agents formetastatic disease).30The ORR was 34.5% (95% CI, 26.1% to 43.9%), the CBR was48.2% (95%CI, 38.8% to 57.9%), mPFS was 6.9months (95%CI, 4.2 to 8.4 months), and median duration of response was7.2months (95%CI, 4.6months to not estimable). Both stud-ies showed a favorable toxicity profıle, with most AEs beinggrade 1 to 2; the most frequent grade 3 and lower AEs werethrombocytopenia and fatigue.The results of a randomized phase II trial have been also

reported; that study randomly selected 137 previously un-treated patients to receive either T-DM1or trastuzumabwithdocetaxel.31 T-DM1 resulted in a signifıcant prolongation ofmPFS as compared to the comparator arm (14.2 vs. 9.2months, p � 0.035), whereas no signifıcant differences inORR (64% for T-DM1 vs. 58% in the control arm) or in CBR(75% for T-DM1 vs. 81% in the control arm)were found. Thetoxicity profıle of T-DM1 compared favorably with standardtreatment, with the rates of grade 3 or 4 neutropenia being5.8% and 61% and the rates of alopecia being 4.3% and 67%,respectively.Positive results confırming the antitumor activity of

T-DM1 in the setting of HER2-positive MBC were recently



reported from the phase III randomized trial EMILIA(NCT00829166).32 This study randomly assigned 991 pa-tients, previously treated with trastuzumab and a taxane, toreceive T-DM1 or lapatinib plus capecitabine. The studymetits primary endpoint, with T-DM1 resulting in a signifıcantprolongation of mPFS as compared with the lapatinib pluscapecitabine arm (9.6 vs. 6.4 months, HR 0.65; 95% CI, 0.55to 0.77, p � 0.001), assessed by independent review. TheORR also favored the T-DM1 arm (43.6% vs. 30.8%, p �0.001), as did the median duration of response (12.6 monthsvs. 6.5months). ThemedianOS (mOS) at the second interimanalysis crossed the stopping boundary for effıcacy (30.9 vs.25.1 month, HR 0.68; 95% CI, 0.55 to 0.85, p � 0.001) favor-ing the T-DM1 arm. In terms of toxicity, rates of grade 3 or 4AEs were lower for T-DM1 overall (41% vs. 57%). T-DM1was associated with higher incidences of thrombocytopeniaand increased liver enzyme levels and lower incidences ofdiarrhea, nausea, vomiting, and palmar-plantar erythrodys-thesia compared with that for lapatinib plus capecitabine.Although the results of the EMILIA study provide solid

evidence for T-DM1 as the treatment of choice in the second-line setting of HER2-positive MBC, another phase III ran-domized trial termed MARIANNE (NCT01120184) isseeking evidence to support the use of T-DM1 in the fırst-linesetting.With 1,092 enrolled patients, MARIANNEwill com-pare the effıcacy of T-DM1 plus pertuzumab, T-DM1 plusplacebo, and the combination of trastuzumab plus a taxane.A third phase III study, TH3RESA (NCT01419197), is eval-uating T-DM1 in the third or later line of therapy. In thisstudy, patients with HER2-positive MBC who previouslyhave received trastuzumab, lapatinib, and chemotherapyare randomly selected to receive T-DM1 or physician’schoice of therapy. Other ongoing phase II trials are currentlyexploring the feasibility and/or antitumor potency of T-DM1in the adjuvant (NCT01196052) and neoadjuvant setting(NCT01745965) of HER2-positive BC.

Pi3k/Akt/mTOR BLOCKING AGENTSPI3K/Akt/mTOR represents an intracellular signal transduc-tion pathway commonly deregulated in the setting of BC andpotentially mediating resistance to anti-HER2-blockingagents.33 To date, rapalogs represent the most clinically ad-vanced agents targeting this signaling pathway. A phase Istudy assessed the triple combination of everolimuswith pac-litaxel and trastuzumab administered weekly in 33 patientswith heavily pretreated (median of two lines of chemother-apy in metastatic setting, range, 0 to 17 lines) HER2-positiveMBC.34 Encouraging antitumor activity was reported, withan overall disease control rate 6 months or fewer of 74% andmPFS reaching 34 weeks (95%CI, 29.1 to 40.7 weeks). Grade3 to 4 neutropenia was the most common toxicity observed(17 patients; 52%). Another phase I study evaluated a regi-men of everolimus combined with vinorelbine and trastu-zumab in a similar population of 47 patients, with the ORRreaching 19.1%, disease control rate 6 months or fewer of83% and mPFS reaching 30.7 weeks (95% CI, 28 to 44.9

weeks).35 Neutropenia (92%) and stomatitis (70%) of anygrade were the most common AEs. Results of the dual com-bination of everolimus with trastuzumab in the metastaticsetting have been reported from a pooled analysis (47 pa-tients) of two phase I/II studies conducted concurrently: aCBR of 34% and a mPFS of 4.1 month were reported withfatigue, infection, and mucositis being the most frequentAEs.36 Currently, there are two ongoing phase III trials ofeverolimus in the setting ofHER2-positiveMBC:BOLERO-1(NCT00876395) assesses the triplet of everolimus, trastu-zumab, and paclitaxel as fırst-line regimen, and BOLERO-3(NCT01007942) assesses the addition of vinorelbine to thetwo targeted agents in patients previously treated with pa-tients. Importantly, the clinical activity of rapalogs is damp-ened by theirmode ofmolecular action, which is restricted tomTORC1 inhibition, whereas mTORC2 remains unaf-fected.37 To overcome this innate limitation, there are cur-rently many targeted agents blocking different molecularcomponents of the PI3K/Akt/mTOR signaling pathway un-der development in HER2-positive BC, namely pan-PI3K,PI3K isoform-selective, mTORC1/2 selective, dual PI3K/mTOR, and Akt inhibitors.38

OTHER EXPLORATORY ANTI-HER2-BLOCKINGSTRATEGIESUnderstanding the molecular biology of HER2-positive BCcoupled with the elucidation of the molecular mechanismsmediating resistance toHER2 blockade can lead to the devel-opment of effective targeted agents. Ongoing trials combin-ing HER2-targeted agents with compounds blocking othersignaling pathways hold the promise to further improve theclinical outcome of patients with HER2-positive BC. A po-tentially promising approach appears to be the combinationof HER2 blockade with insulin growth factor receptor 1(IGF-1R)–blocking agents. A bidirectional crosstalk ofHER2 and IGF-1R is well documented, with IGF-1R inhibi-tion restoring sensitivity to trastuzumab in preclinical mod-els.39,40 Clinical trials evaluating this hypothesis are currentlyongoing. Another rational combination is the blockade ofHER2 and SRC. SRC is a nonreceptor tyrosine kinase thatwas recently shown to function as a common node down-stream of multiple trastuzumab-resistance mechanisms,both acquired and de novo.41 Blocking SRC in combinationwith trastuzumab showed a heightened synergistic antitu-mor effıcacy in trastuzumab-resistant tumors in vivo, under-lining the potential clinical usefulness of this approach.HER3, as already highlighted preclinically, is another ratio-nal molecular target in the setting of HER2-posititve BC.HER3 is a potent activator of PI3K/Akt signaling pathwaythat has been demonstrated to be upregulated on HER2blockade.42 Importantly, a recent report suggested that dualHER2-blockade in HER2-positive BC cells does not com-pletely abrogate HER3 function, so that the addition of aHER3-targeting monoclonal antibody, with the dual combi-nations of either trastuzumab/pertuzumab or trastuzumab/lapatinib, led to increased antitumor activity.43

ZARDAVAS ET AL


CONCLUSIONThe aforementioned data underline the antitumor potencyof newly developed HER2-blocking agents against HER2-positive BC. Dual HER2 blockade has resulted in clinicalsuccesses in both the neoadjuvant (Fig. 1) andmetastatic set-ting. The development of anti-HER2 ADCs promises furthertreatment options for these patients, achieving selected de-livery of potent chemotherapy coupled with HER2 inhibi-tion. Further treatment options could rise from thedissection of the molecular mechanisms mediating resis-tance to HER2 blockade. Nevertheless, many questions needto be answered: how can we better select the optimal HER2blockade strategy for the individual patient? Clearly, to an-swer this question, robust predictive biomarkers are urgentlyneeded. To the present day, only HER2 remains a suitablebiomarker for therapeutic decisions about HER2 blockade,as was exemplifıed by the failure of the recent biomarkeranalysis of theCLEOPATRA study to identify additional pre-dictive biomarkers.44Howcanwe identify those patientswhocan be effectively treated with HER2 blockade without theaddition of cytotoxic chemotherapy? Recent fındings indi-cate that FDG-PET/CT can serve as a valuable tool to thisend. The Neo-ALTTO PET study assessed whether meta-bolic response with anti-HER2 therapies could predict pCRat the timeof surgery; pCR rateswere twice as high in patients

who were FDG-PET/CT responders (using European Or-ganisation for Research and Treatment of Cancer criteria45)compared with nonresponders (41% vs. 21% with FDG-PET/CT after 2 weeks of chemotherapy-free HER2 block-ade).46 Building on these promising results, the BreastInternational Group (BIG) is planning to launch the I- Inte-grating Imaging to De-escalate Therapy in HER2-positivebreast CancerManagement (DREAM) trial, aiming to defınethe most effective HER2-targeting strategy and identify earlyon the group of patients who could potentially be spared sys-temic chemotherapy. What is the optimal sequencing oftreatment among the variety of treatment options that arebeing constantly added in our therapeutic armamentarium(Table 5)? Is it possible to further improve the antitumor ef-fıcacy of dual HER2 blockade regimens by either HER2blockade triplets or through the addition of other com-pounds targeting anti-HER2-resistance mechanisms? Lastly,will we identify more effective anti-HER2 adjuvant regimenswith acceptable toxicity? Moreover, in a fınancially con-strained environment, issues of increasing costs for the na-tional health care systems are of concern. These questionscan be answered through properly conducted collaborativeclinical trials. The near future is promising for major leapsforward toward more effective treatment of HER2-positiveBC.



Employment or Leadership Position: None. Consultant or Advisory Role: David A. Cameron, GlaxoSmithKline; Roche. Ian E. Krop, Genentech; PumaBiotechnology (U). Martine J. Piccart-Gebhart, Amgen; AstraZeneca; Bayer Schering Pharma; PharmaMar; Roche/Genentech; Sanofi. Stock Ownership: None.Honoraria: David A. Cameron, AstraZeneca; GlaxoSmithKline; Pfizer. Martine J. Piccart-Gebhart, Amgen; AstraZeneca; Bayer Schering Pharma; Novartis;PharmaMar; Roche/Genentech; Sanofi. Research Funding: David A. Cameron, Amgen; Pfizer; Roche. Ian E. Krop, Genentech. Expert Testimony: None. OtherRemuneration: None.

References

1. YardenY, SliwkowskiMX.Untangling theErbB signalling network.NatRev Mol Cell Biol. 2001;2:127-137.

2. Slamon DJ, Clark GM, Wong SG, et al. Human breast cancer: correla-

tion of relapse and survival with amplifıcation of the HER-2/neu onco-gene. Science. 1987l235:177-182.

3. Arteaga CL, Sliwkowski MX, Osborne CK, et al. Treatment of HER2-

TABLE 5. Developmental Status of Targeted Agents in HER2-Positive Breast Cancer, Across the DifferentDisease Settings

Agent Adjuvant Neoadjuvant Metastatic, First Line Metastatic � Second Line

Trastuzumab � lapatinib Phase III (ongoing) Phase III (completed) Phase I (ongoing) Application for approval withdrawn

Pertuzumab Phase III (ongoing) Phase III (completed) Approved Phase II (ongoing)

Trastuzumab-DM1 Phase III (ongoing) Phase II (ongoing) Phase III (ongoing) Approval under evaluation

Afatinib - Phase II (ongoing) - Phase III (ongoing)

Neratinib - Phase II (ongoing) Phase II (ongoing) Phase II (ongoing)

Eeverolimus - - Phase III (ongoing) Phase III (ongoing)



positive breast cancer: current status and future perspectives. NatureReviews Clinical Oncology. 2011;9:16-32.

4. Xia W, Gerard CM, Liu L, et al. Combining lapatinib (GW572016),a small molecule inhibitor of ErbB1 and ErbB2 tyrosine kinases,with therapeutic anti-ErbB2 antibodies enhances apoptosis ofErbB2-overexpressing breast cancer cells. Oncogene. 2005;24:6213-6221.

5. Scaltriti M, Verma C, Guzman M, et al. Lapatinib, a HER2 tyrosine ki-nase inhibitor, induces stabilization and accumulation ofHER2 and po-tentiates trastuzumab-dependent cell cytotoxicity. Oncogene. 2009;28:803-814.

6. Blackwell KL, Burstein HJ, Storniolo AM, et al. Overall survival benefıtwith lapatinib in combination with trastuzumab for patients with hu-man epidermal growth factor receptor 2-positive metastatic breast can-cer: fınal results from the EGF104900 Study. J Clin Oncol. 2012;30:2585-2592.

7. Baselga J, Bradbury I, EidtmannH, et al. Lapatinib with trastuzumab forHER2-positive early breast cancer (NeoALTTO): a randomised, open-label, multicentre, phase 3 trial. Lancet. 2012;379:633-640.

8. Guarneri V, Frassoldati A, Bottini A, et al. Preoperative chemotherapyplus trastuzumab, lapatinib, or both in human epidermal growth factorreceptor 2-positive operable breast cancer: results of the randomizedphase II CHER-LOB study. J Clin Oncol. 2012;30:1989-1995.

9. Chang J, Mayer I, Forero-Torres A, et al. TBCRC 006: a multicenterphase II study of neoadjuvant lapatinib and trastuzumab in patientswith HER2-overexpressing breast cancer. J Clin Oncol. 2011;29 (suppl;abstr 505).

10. Bozovic I, Capelan M, De Azambuja E. Accrual target reached andchanges after IDMC recommendation. Breast International GroupNewsletter 2011. December 2011;13(2):21.

11. Franklin MC, Carey KD, Vajdos FF, et al. Insights into ErbB signalingfrom the structure of the ErbB2-pertuzumab complex. Cancer Cell.2004;5:317-328.

12. Scheuer W, Friess T, Burtscher H, et al. Strongly enhanced antitumoractivity of trastuzumab and pertuzumab combination treatment onHER2-positive human xenograft tumor models. Cancer Res. 2009;69:9330-9336.

13. PorteraCC,Walshe JM,RosingDR, et al. Cardiac toxicity and effıcacy oftrastuzumab combined with pertuzumab in patients with [corrected]human epidermal growth factor receptor 2-positive metastatic breastcancer. Clin Cancer Res. 2008;14:2710-2716.

14. Baselga J, Gelmon KA, Verma S, et al. Phase II trial of pertuzumab andtrastuzumab in patients with human epidermal growth factor receptor2-positive metastatic breast cancer that progressed during prior trastu-zumab therapy. J Clin Oncol. 2010;28:1138-1144.

15. Lenihan D, Suter T, Brammer M, et al. Pooled analysis of cardiac safetyin patients with cancer treated with pertuzumab. Ann Oncol. 2012;23:791-800.

16. Baselga J, Cortes J, Kim S-B, et al. Pertuzumab plus trastuzumab plusdocetaxel for metastatic breast cancer.N Engl J Med. 2012;366:109-119.

17. Swain SM, Kim S, Cortes J, et al. Confırmatory overall survival (OS)analysis of CLEOPATRA: a randomized, double-blind, placebo-controlled phase III study with pertuzumab (P), trastuzumab (T), anddocetaxel (D) in patients (pts) with HER2-positive fırst-line (1L) meta-static breast cancer (MBC). San Antonio Breast Cancer Symposium.2012;abstract P5-18-26.

18. Gianni L, Pienkowski T, Im Y-H, et al. Effıcacy and safety of neoadju-vant pertuzumab and trastuzumab in women with locally advanced, in-flammatory, or early HER2-positive breast cancer (NeoSphere): arandomised multicentre, open-label, phase 2 trial. Lancet Oncol. 2012;13:25-32.

19. Schneeweiss A, Chia S, Hickish T, et al. Neoadjuvant pertuzumab and

trastuzumab concurrent or sequential with an anthracycline-containingor concurrent with an anthracycline-free standard regimen: a random-ized phase II study (TRYPHAENA). Cancer Res. 2011;71(suppl 24):112s.

20. Burstein HJ, Sun Y, Dirix LY, et al. Neratinib, an irreversible ErbB re-ceptor tyrosine kinase inhibitor, in patients with advanced ErbB2-positive breast cancer. J Clin Oncol. 2010;28:1301-1307.

21. Swaby R, Blackwell KL. Neratinib in combination with trastuzumab forthe treatment of advanced breast cancer: Aphase I/II study. J ClinOncol.2009;27:15s (suppl; abstr 1004).

22. Jankowitz R, Abraham J, Tan AR, et al. A phase I dose-escalation studyevaluating weekly paclitaxel with neratinib and trastuzumab in womenwithmetastaticHER2-positive breast cancer,NSABPFB-8. J ClinOncol.2012;30 (suppl; abstr 611).

23. GajriaD,KingT, PannuH, et al. Combined inhibition ofmTORC1withTemsirolimus and HER2 with Neratinib: A Phase I/II Study in Patientswith Metastatic HER2-Amplifıed or Triple-Negative Breast Cancer.Cancer Res. 2011;71(suppl):PD09-08.

24. Martin M, Bonneterre J, Geyer CE Jr., et al. A phase 2, randomized,open-label, study of neratinib (HKI- 272) vs lapatinib plus capecitabinefor 2nd/3rd-line treatment of HER2� locally advanced or metastaticbreast cancer. San Antonio Breast Cancer Symposium. 2011;abstractS5-7.

25. Hickish T,Wheatley D, LinN, et al. Use of BIBW2992, a novel irrevers-ible EGFR/HER1 and HER2 tyrosine kinase inhibitor to treat patientswith HER2-positive metastatic breast cancer after failure of treatmentwith trastuzumab. Cancer Res. 2010;69:5060-5060.

26. Erickson HK, Park PU, Widdison WC, et al. Antibody-maytansinoidconjugates are activated in targeted cancer cells by lysosomal degrada-tion and linker-dependent intracellular processing.Cancer Res. 2006;66:4426-33.

27. Lewis Phillips GD, Li G, Dugger DL, et al. Targeting HER2-positivebreast cancer with trastuzumab-DM1, an antibody-cytotoxic drug con-jugate. Cancer Res. 2008;68:9280-9290.

28. Krop IE, Beeram M, Modi S, et al. Phase I study of trastuzumab-DM1,an HER2 antibody-drug conjugate, given every 3 weeks to patientswith HER2-positive metastatic breast cancer. J Clin Oncol. 2010;28:2698-704.

29. Burris HA 3rd, RugoHS, Vukelja SJ, et al. Phase II study of the antibodydrug conjugate trastuzumab-DM1 for the treatment of human epider-mal growth factor receptor 2 (HER2)-positive breast cancer after priorHER2-directed therapy. J Clin Oncol. 2011;29:398-405.

30. Krop IE, LoRusso P, Miller KD, et al. A phase II study of trastuzumabemtansine in patients with human epidermal growth factor receptor2-positive metastatic breast cancer who were previously treated withtrastuzumab, lapatinib, an anthracycline, a taxane, and capecitabine.J Clin Oncol. 2012;30:3234-3241.

31. EA Perez, L Dirix, J Kocsis, et al. Effıcacy and safety of trastuzumab-DM1 versus trastuzumab plus docetaxel in HER2-positive metastaticbreast cancer patients with no prior chemotherapy for metastatic dis-ease: preliminary results of a randomized,multicenter, open-label phase2 study. 35th ESMO Congress. 2010; abstract LBA3.

32. Verma S, Miles D, Gianni L, et al. Trastuzumab emtansine forHER2-positive advanced breast cancer. N Engl J Med. 2012;367:1783-1791.

33. Berns K, Horlings HM, Hennessy BT, et al. A functional genetic ap-proach identifıes the PI3K pathway as a major determinant of trastu-zumab resistance in breast cancer. Cancer Cell. 2007;12:395-402.

34. Andre F, Campone M, O’Regan R, et al. Phase I study of everolimusplus weekly paclitaxel and trastuzumab in patients with metastaticbreast cancer pretreated with trastuzumab. J Clin Oncol. 2010;28:5110-5115.

ZARDAVAS ET AL


35. JerusalemG, Fasolo A, Dieras V, et al. Phase I trial of oral mTOR inhib-itor everolimus in combination with trastuzumab and vinorelbine inpre-treated patients with HER2-overexpressing metastatic breast can-cer. Breast Cancer Res Treat. 2011;125:447-455.

36. Morrow PK, Wulf GM, Ensor J, et al. Phase I/II study of trastuzumabin combination with everolimus (RAD001) in patients with HER2-overexpressing metastatic breast cancer who progressed ontrastuzumab-based therapy. J Clin Oncol. 2011;29:3126-3132.

37. O’Reilly KE, Rojo F, She Q-B, et al. mTOR inhibition induces upstreamreceptor tyrosine kinase signaling and activates Akt. Cancer Res. 2006;66:1500-1508.

38. Sledge GW, Cardoso F, Winer EP, et al. A Dickens tale of the treatmentof advanced breast cancer: the past, the present, and the future. ASCOEducational Book. 2012;28-38.

39. Nahta R, Yuan LXH, Zhang B, et al. Insulin-like growth factor-I recep-tor/human epidermal growth factor receptor 2 heterodimerization con-tributes to trastuzumab resistance of breast cancer cells. Cancer Res.2005;65:11118-11128.

40. Huang X, Gao L, Wang S, et al. Heterotrimerization of the growth factorreceptors erbB2, erbB3, and insulin-like growth factor-i receptor in breastcancer cells resistant to herceptin.Cancer Res. 2010;70:1204-1214.

41. Zhang S, HuangW-C, Li P, et al. Combating trastuzumab resistance bytargeting SRC, a common node downstream of multiple resistancepathways. Nature Medicine. 2011;17:461-469.

42. Garrett JT, Olivares MG, Rinehart C, et al. Transcriptional and post-translational up-regulation of HER3 (ErbB3) compensates for inhibi-tion of the HER2 tyrosine kinase. Proc Natl Acad Sci U S A. 2011;108:5021-5026.

43. Garrett JT, SuttonCR,KubaMG,et al. Dual blockade ofHER2 inHER2-overexpressing tumor cells does not completely eliminate HER3 func-tion. Clin Can Res. 2013;19:610-619.

44. Baselga J, Cortes J, Im SA, et al. Biomarker analyses in CLEOPATRA: Aphase III, placebo-controlled study of pertuzumab in HER2-positive,fırst-line metastatic breast cancer (MBC). In: Abstracts: Thirty-FifthAnnual CTRC-AACR SanAntonio Breast Cancer Symposium. SanAn-tonio, TX: American Association for Cancer Research; 2012. AbstractS5-1.

45. Young H, Baum R, Cremerius U, et al. Measurement of clinical andsubclinical tumour response using [18F]-fluorodeoxyglucose andpositron emission tomography: review and 1999 EORTC recom-mendations. European Organization for Research and Treatment ofCancer (EORTC) PET Study Group. Eur J Cancer. 1999;35:1773-1782.

46. Gamez C, Flamen P, Holmes E, et al. FDG-PET/CT for early predictionof response to neoadjuvant lapatinib, trastuzumab, and their combina-tion in HER2 positive breast cancer patients: the Neo-ALTTO PETstudy results. Paper presented at: The 2011 EuropeanMultidisciplinaryCancer Congress; September 2011; Stockholm, Sweden.



BREAST CANCER

Finding the Balance in theManagement of Low-Risk BreastCancer

CHAIRJulia R. White, MDThe Ohio State UniversityColumbus, OH

SPEAKERSJavier Cortes, MD, PhDVall d’Hebron University HospitalBarcelona, Spain

Paul Goss, MD, PhDMassachusetts General Hospital Cancer CenterBoston, MA

The Effect of Biology in the Treatment of Small Breast TumorsJose Perez-Garcia, MD, Eva Munoz-Couselo, MD, PhD, and Javier Cortes, MD, PhD

OVERVIEW

Although the outcome of small (T1a/b) node-negative breast tumors is generally excellent, in the absence of prospective clinical trials,we are limited to data derived from retrospective analyses. Overall, the 10-year overall mortality rate is approximately 20%, while the10-year breast cancer-specific mortality is in the range of 4% to 8% among this population in the absence of systemic therapy. Thisclearly reflects that many patients die of causes not related to breast cancer. Due in large part to breast cancer screening programs,the incidence of small tumors is increasing. There is consequently a growing interest in identifying factors that negatively affect theprognosis of these patients. Several studies have shown that patients with triple-negative and HER2� tumors have a worse prognosiscompared with hormone-receptor–positive, HER2- small breast cancers. However, the recent explosion of knowledge of the molecularcharacteristics of tumors is opening a new way to address cancer. Different genomic assays are currently available to help betterpredict the outcome of breast cancer patients. However, none of these techniques have been specifically evaluated in patients withsmall (T1a/b) node-negative tumors, and only a small number of patients with these tumors were included in those studies. In addition,very limited data are available about the role of these assays in patients with triple-negative or HER2-positive cancers. Although achemotherapy-based strategy might be useful for triple-negative or HER2-positive T1b tumors, more information is urgently needed inorder to optimize the treatment of our patients.

The incidence of small (T1a/b) node-negative breast tu-mors is steadily increasing due in part to the implemen-

tation of breast cancer screening programs.1 Unfortunately,the role of adjuvant systemic therapy in this population re-mains unclear.Classically, tumor size and axillary lymph node status have

been considered themost important prognostic factors in pa-tients with breast cancer.2,3 For many years, small (T1a/b)node-negative breast tumors were believed to have such agood prognosis that systemic adjuvant therapy was notviewed to be necessary. However, more recent research onthis issue has shown that certain subgroups of these patientshave a signifıcant risk of systemic recurrence. The challengeis therefore to identify patients most likely to benefıt fromsystemic adjuvant therapy based on validated prognostic andpredictive factors. Although clinical and pathologic featuresare currently themainstay of clinical decisionmaking for thispopulation, a better knowledge of the molecular biology ofbreast cancer and the introduction of new prognostic toolsinto daily clinical practice, such as MammaPrint and On-cotype Dx (and most likely the PAM50 test in the upcomingmonths), are establishing a new and intriguing scenario inthe adjuvant treatment of these patients.4-8The best systemic strategies for small tumors vary across

different guidelines and are a matter of current debate. For

example, according to the National Comprehensive CancerNetwork (NCCN) guidelines, the selection of the adjuvantsystemic treatment for patients with breast cancer is based onhormone receptor and HER2 status, tumor size, and axillarylymph node status.9 In brief, these guidelines do not recom-mend administration of adjuvant chemotherapy in breasttumors that are 0.5 centimeters (cm) or smaller without ax-illary lymph node involvement, irrespective of hormone re-ceptor, and HER2 status. In addition, prognostic tools suchas Oncotype Dx are recommended for HER2- and hormone-receptor–positive disease that is 0.5 cm or larger with nega-tive axillary lymphnodes, whereas adjuvant systemic therapyis recommended for triple-negative and HER2� breast can-cers between 0.6 and 1.0 cm without axillary lymph node in-volvement. Therefore, considering these guidelines, a patientage 40 with a stage T1aN0 triple-negative breast cancer (0.5cm of diameter, high tumor grade, and Ki-67 index of 80%)should not receive adjuvant chemotherapy. But, are we head-ing in the right direction under this or other guidelines?Thus, we discuss the outcome of small (T1a/b) node-

negative breast tumors; the factors that adversely affect theprognosis among this population; and how new prognostictools and the molecular classifıcation of breast cancer canhelp us to identify those patients that might benefıt from ad-juvant systemic therapy.

From the Vall d’Hebron Institute of Oncology, Vall d’Hebron University Hospital, Universitat Autonoma de Barcelona, Barcelona, Spain; Vall d’Hebron Institute of Oncology, Vall d’HebronUniversity Hospital, Medica Scientia Innovation Research (MedSIR), Passeig Vall d’Hebron, Barcelona, Spain.


Corresponding author: Jose Perez-Garcia, MD, Vall d’Hebron Institute of Oncology, Vall d’Hebron University Hospital, Medica Scientia Innovation Research (MedSIR), Passeig Vall d’Hebron 119-129,08035, Barcelona, Spain; email: [email protected].


THE EFFECT OF BIOLOGY IN SMALL BREAST TUMORS


OUTCOME OF SMALL (T1A/B) NODE-NEGATIVEBREAST CANCERSSeveral authors have retrospectively analyzed the outcome ofsmall (T1a/b) node-negative breast tumors. Overall, the 10-year breast cancer-specifıc mortality is in the range of 4% to8%, while the 10-year overall mortality rate is approximately20% among this population in the absence of systemic ther-apy. Most of these studies had a short follow-up and/or in-cluded a small number of patients. In addition, it is importantto take into account that some of them recruited patientstreated with adjuvant systemic therapy. This procedure canlead to an overestimation of the outcome that would befound in an entirely untreated population. For this reason,the studies summarized below have been selected either be-cause they included a signifıcant number of patients and/orbecause they had a longer follow-up. Moreover, only datafrom surgery-treated patients are reported.Fisher et al. retrospectively evaluated the outcome and

treatment of patients with breast tumors smaller than 1 cmwithout axillary lymph node involvement from fıve NationalSurgical Adjuvant Breast and Bowel Project (NSABP) ran-domized clinical trials.10 Two hundred and thirty-fıve pa-tientswith estrogen receptor (ER)-negative tumors and 1,024patients with ER� tumors were identifıed in these trials. Me-

dian follow-up was 8 years. Patients with ER� tumors re-ceived surgery alone (26%); surgery and tamoxifen (53%); orsurgery, tamoxifen, and chemotherapy (21%). Patients withER- tumors received surgery alone (26%) or surgery and che-motherapy (74%). The 8-year recurrence-free survival (RFS)rates for women with ER- and ER� tumors who receivedsurgery alone were 81% and 86%, respectively. The 8-yearoverall survival (OS) rates for surgery-treated patients withER- and ER� tumors were 93% and 90%, respectively, andthe 8-yearOS rate for all patientswas 92%, regardless of treat-ment and ER status. However, only half of these patients diedfrom breast cancer.Rosen et al. retrospectively assessed the prognosis of 767

patients with stage T1/2 node-negative breast cancer treatedwith surgery alone.11 A total of 171 patients with small(T1a/b) node-negative breast tumors were included in theanalysis. Median follow-up was 18 years. This subgroup hadRFS rates of 91% and 88% at 10 and 20 years, respectively.Additionally, there were 178 (23.2%) deaths caused by breastcancer among all patients included in the study. In this re-gard, it is important to note that the probability of death frombreast cancer did not exceed the probability of death causedby other causes among women with small (T1a/b) node-negative breast tumors.Finally, Chia et al. retrospectively evaluated the 10-year

outcomes in a population-based cohort of node-negative,lymphatic, and vascular invasion–negative early breast can-cers without adjuvant systemic therapy.12 The 430 tumors in-cluded in this analysis were smaller than 1 cm. Medianfollow-up of the entire series was 10.4 years. The 10-year OSrate was 79%, and the 10-year breast cancer-specifıc survivalwas 92%.It is necessary to emphasize that one of the major weak-

nesses of these studies, apart from their retrospective nature,is the fact that the effect of HER2 status in the outcome ofthese patients was not analyzed.In summary, the outcome of small (T1a/b) node-negative

breast cancers is generally excellent. However, the breastcancer-specifıc mortality, probably themost appropriate endpoint for this particular subgroup, varies between 4% and 8%among these patients, warranting an optimization of the ad-juvant systemic therapy for this population.

COMMONLY USED FACTORS THAT ADVERSELYAFFECT THE PROGNOSIS OF SMALL (T1A/B)NODE-NEGATIVE BREAST CANCERSAlthough the risk of relapse of small (T1a/b) node-negativebreast tumors is low, there is a growing interest in identifyingfactors that negatively affect the prognosis of these patients,to recognize which patients might benefıt from adjuvant sys-temic therapy as well as to avoid unnecessary side effects.In the study of Fisher et al. the risk of recurrence was

greater in women who had tumors of 1 cm in size than forthose women who had tumors of less than 1 cm; in womenage 49 or younger compared with women age 50 or older;and in women with ductal or lobular carcinoma compared

KEY POINTS

� Due to the increase in breast cancer screening programs,the incidence of small, node-negative tumors is increasing,which leads to a growing interest in identifying factorsthat negatively impact the prognosis of patients withbreast cancer.

� The recent explosion of knowledge about the molecularcharacteristics of tumors is opening a new way to tacklecancer. In breast cancer, different genomic assays areavailable to better predict the outcome of patients withbreast cancer. However, none of them has been specificallyevaluated in patients with T1a/b node-negative tumors andthe number of patients with these tumors included in thesestudies was very small.

� Although clinical and pathologic features are currently themainstay of clinical decision making for this population, abetter knowledge of the molecular biology of breast cancerand the introduction of new prognostic tools into dailyclinical practice, such as MammaPrint and Oncotype Dx(and most likely the PAM50 test in the upcoming months),are establishing a new and intriguing scenario in theadjuvant treatment of these patients.

� The breast cancer intrinsic subtypes may represent arelevant prognostic factor regardless of tumor size.Overall, triple-negative and HER2� breast cancers byimmunohistochemistry are associated with a higher risk ofrecurrence among small (T1a/b) node-negative tumors.

� Although the outcome of small (T1a/b) node-negativebreast tumors is generally excellent, in the absence ofprospective clinical trials, we are limited to clinical dataderived from retrospective analyses.

PEREZ-GARCIA, MUNOZ-COUSELO, AND CORTES


with women who had other histologic tumor types, irrespec-tive of ER status.10 In the analysis of Rosen et al., patientswith infıltrating ductal carcinoma had a worse RFS rate thanpatients with special tumor types (medullary, mucinous,tubular, adenocystic, and papillary) and infıltrating lobularcarcinoma.11In the study of Chia et al. tumor grade appeared to be an

important predictor of risk of breast cancer death, regardlessof tumor size.12 Finally, Hanrahan et al. also retrospectivelyanalyzed the OS and cause-specifıc mortality of patients withstage T1a/bN0 breast cancer registered in the Surveillance,Epidemiology, and End Results (SEER) Program from 1988to 2001.13 The main limitation of this analysis was that theproportion of patients who received adjuvant chemotherapyor hormonal therapywas unknown.A total of 51,246 patientswere identifıed. Median follow-up was 64 months. In thisstudy, a trend toward higher probability of death was foundin patients younger than age 50 at diagnosis; in patients withadverse pathologic features such as high tumor grade, ER-status, and progesterone-receptor–negative status; and in pa-tients with an inadequate axillary lymph node assessment(fewer than six nodes removed at axillary dissection).13Although none of these studies evaluated the influence of

HER2 status in the prognosis of these small (T1a/b) node-negative tumors, other series have appropriately assessed thisissue.Gonzalez-Angulo et al. reviewed the risk of recurrence in

women diagnosed with stage T1a/b, node-negative breastcancer taking HER2 status into account.14 A total of 965 pa-tients were identifıed at the MD Anderson Cancer Centerbetween 1990 and 2002. Patients who received adjuvant che-motherapy or trastuzumabwere excluded.Median follow-upwas 74 months. Ten percent of patients had HER2� tumors,and the 5-year RFS rates were 77.1% and 93.7% in patientswith HER2� and HER2- tumors, respectively (p � 0.001).The 5-year distant RFS rates were 86.4% and 97.2% in pa-tients with HER2� and HER2- tumors, respectively (p �0.001). Among HER2� tumors, the risk of recurrence wassimilar to independent of hormone-receptor status, althoughthe fact that up to 55% of hormone-receptor–positive pa-tients received adjuvant hormone therapy should be takeninto consideration to avoid misinterpretations. Unfortu-nately, the authors did not analyze the outcome of HER2�patients according to tumor stage (T1a vs. T1b).Curigliano et al. also identifıed 150 consecutive patients

with HER2� tumors among a population of 2,130 patientswith stage T1a/b, node-negative breast cancers at the Euro-pean Institute of Oncology between 1999 and 2006.15 Amatched cohort was selected by using variables of hormone-receptor status, age, and year of surgery. No patient receivedadjuvant trastuzumab, but up to 50% of patients received ad-juvant chemotherapy in the hormone-receptor–negativegroup, andmore than 90% of patients received adjuvant hor-mone therapy, alone or in combination with chemotherapy,in the hormone-receptor–positive group. Median follow-upwas 4.6 years. In the hormone-receptor–positive group, the

5-year disease-free survival (DFS) rates were 99% (95% CI,96% to 100%) for HER2- disease and 92% (95% CI, 86% to99%) for HER2� disease. In the hormone receptor–negativegroup, the 5-year DFS rates were 92% (95%CI, 84% to 100%)forHER2- disease and 91% (95%CI, 84% to 99%) forHER2�disease. It is important to emphasize that in the hormone-receptor–negative group, the authors compared HER2�with patients with triple-negative breast cancer. Moreover,the DFS rates according to tumor stage (T1a vs. T1b) andhormone receptor status were adequately assessed. AmongHER2� and hormone-receptor–positive tumors, the 5-yearDFS rates were 88% (95% CI, 67% to 96%) in patients withstage T1a and 95% (95% CI, 82% to 99%) in patients withstage T1b. Among HER2� and hormone-receptor–negativetumors, the 5-year DFS rates were 93% (95%CI, 72% to 98%)in patients with stage T1a and 85% (95% CI, 60% to 95%) inpatients with stage T1b. However, considering the limitednumber of patients and the treatment heterogeneity, defıni-tive conclusions cannot be drawn from this subgroup analy-sis.In sum, a younger age at diagnosis, ER- status, HER2� sta-

tus, and high tumor grade adversely affect the prognosis ofnode-negative breast cancers smaller than 1 cm. These fac-tors should be taken into account to defıne the best adjuvantsystemic therapy for these patients.

HOW CAN NEW PROGNOSTIC TOOLS HELP TOBETTER IDENTIFY WHICH PATIENTS WITH SMALL(T1A/B) NODE-NEGATIVE TUMORS MIGHT BENEFITFROM ADJUVANT SYSTEMIC THERAPY?MammaPrint and Oncotype Dx are two prognostic testsbased on retrospective analyses that are currently used inclinical practice. These tools can help physicians to deter-mine whether or not a patient with breast cancer will benefıtfrom adjuvant chemotherapy.MammaPrint was initially evaluated in a series of 295 con-

secutive women diagnosed with stage T1/2, ER- or ER�breast cancer, with or without axillary lymph node in-volvement.6 Up to 40% of patients had received adjuvantchemotherapy. At 10 years, the probability of remaining freeof distant metastases was lower in the group with a poor-prognosis signature than in the group with a good-prognosissignature (HR for distant metastases � 5.1; p � 0.001). Theprognosis profıle was signifıcantly associated with tumorgrade, ER status, tumor diameter, and age at diagnosis,but not with the number of axillary positive nodes. The au-thors did not specifıcally analyze these fındings amongsmall (T1a/b) node-negative breast cancers. Subsequently,MammaPrint was validated with the same results in a seriesof 307 patients with stage T1/2, node-negative, ER- or ER�breast cancer.8 No patient received adjuvant systemic ther-apy. Only 11 patients with small (T1a/b) tumors were in-cluded in the analysis. Of these tumors, six had a poor-prognosis signature. The authors did not evaluate in detailthe characteristics of these patients, although ER- statusand high tumor grade were associated with a poor-prognosis



signature again. Finally, MammaPrint was assessed in 965patients with stage T1 breast cancer irrespective of node sta-tus, age, and ER, PR, andHER2 status.16 Only 10%of patientshad received adjuvant chemotherapy. The results were con-sistent with those reported in previous studies. A total of 140small (T1a/b) tumors were included in the analysis. How-ever, given the heterogeneity of these patients, solid conclu-sions should not be drawn.MammaPrint is not only an independent prognostic factor

for patients with early stage breast cancer, but it may also bepredictive for the benefıt of chemotherapy. Knauer et al. an-alyzed a pooled database from six prior studies.17 Five hun-dred and forty-one women diagnosed with stage T1–3, ER-or ER� breast cancer, with or without axillary lymph nodeinvolvement were included. Only patients with a poor-prognosis signature had a signifıcant benefıt from chemo-therapy (distant disease-free survival, 76% vs. 88%; p� 0.01),whereas patients with a good-prognosis signature did notbenefıt from chemotherapy treatment (distant disease-freesurvival, 93% vs. 99%; p � 0.2). The chemotherapy benefıtamong small (T1a/b) node-negative breast cancers was notassessed in this study.In contrast, Oncotype DX was retrospectively evaluated in

the NSABP trial B14, a clinical trial that analyzed the treat-ment with tamoxifen in node-negative, ER� patients.7 Thelevels of expression of 21 genes were used to calculate a re-currence score (RS). The rates of distant recurrence at 10years in the low-risk (� 18), intermediate-risk, and high-riskgroups (� 31) were 6.8%, 14.3%, and 30.5%, respectively.The rate in the low-risk group was signifıcantly lower thanthat in the high-risk group (p� 0.001). A total of 109 patientswith small (T1a/b) tumors were included in the analysis. Ofthese patients, 59.5% had a low RS, 25% had an intermediateRS, and 15.5% had a high RS. A high RS was also associatedwith a higher rate of distant recurrence at 10 years amongthese patients, although the differences with respect to thepatients with intermediate and low RS tumors were not assignifıcant as observed in larger tumors. Later, Oncotype DXwas assessed in the NSABP trial B20 that tested the effect ofadding cyclophosphamide,methotrexate, and fluorouracil ormethotrexate and fluorouracil chemotherapy to 5 years of ta-moxifen in the treatment of patients with node-negative,ER� breast cancer.5 Patients with high RS tumors had a largebenefıt from chemotherapy (relative risk for distant recur-rence, 0.26; 95% CI, 0.13 to 0.53), although patients with lowRS tumors did not benefıt from chemotherapy treatment(relative risk for distant recurrence, 1.31; 95% CI, 0.46 to3.78). The chemotherapy benefıt in patients with intermedi-ate RS tumors remains controversial. A total of 110 patientswith small (T1a/b) tumors were included in the analysis. Ofthese patients, 64% had a low RS, 20% had an intermediateRS, and 16%had a highRS.Unfortunately, the chemotherapybenefıt according to tumor stage (T1a vs. T1b vs. others) wasnot evaluated in this study.Finally, neither MammaPrint nor Oncotype Dx have been

assessed among small (T1a/b) node-negative HER2� and

triple-negative breast cancers, and their use in clinical prac-tice is not recommended in these patients.In summary, some small (T1a/b) node-negative breast

cancers are considered high-risk tumors by MammaPrint orOncotypeDx, in particular high-grade ER� andER- tumors,and might benefıt from adjuvant systemic therapy. How-ever, the effect of tumor stage (T1a vs. T1b vs. others) in thisbenefıt, its influence in the selection of the adjuvant treat-ment, and the optimal chemotherapy regimen merit furtherinvestigation.

IS THE MOLECULAR CLASSIFICATION OF BREASTCANCER RELEVANT FOR DEFINING THE BESTADJUVANT SYSTEMIC TREATMENT FOR PATIENTSWITH SMALL (T1A/B) NODE-NEGATIVE BREASTCANCERS?Breast cancer is an extremely heterogeneous disease withmultiple clinical presentations and tumor characteristics.Gene-expression profıling studies have classifıed breast tu-mors into a number of distinct biological and intrinsic sub-types with prognostic and therapeutic implications, thusproviding a newmolecular classifıcation of breast cancer. Ac-cording to this classifıcation, at least four different molecularsubtypes have been identifıed: luminal A, luminal B, HER2-enriched, and basal-like.18Molecular profıling is not currently ready for use in clinical

decisionmaking. Therefore, a combination of immunohisto-chemical surrogate markers (using ER and PR status, HER2status, tumor grade, and KI-67 index) have been validatedfor molecular subtyping. However, as previously men-tioned, neither HER2 status nor Ki-67 index were analyzedin some of the studies discussed above. Surrogate defınitionsof intrinsic subtypes of breast cancer are summarized inTable 1.19,20The PAM50 gene-expression assay is one of the tools in

current development to classify breast cancers into intrinsicsubtypes.4 Recently, Bastien et al. evaluated the concordancebetween PAM50 breast cancer subtyping and immunohisto-chemistry, concluding that a standard immunohistochemi-cal panel for breast cancer does not adequately identify thePAM50 gene-expression subtypes.21Theriault et al. analyzed the outcome of patients with stage

T1a/b, node-negative breast tumors according to breast can-

TABLE 1. Surrogate Definitions of Intrinsic Subtypesof Breast Cancer

Subtype ER and/or PR HER2 Ki67 Index

Luminal A Positive (PR � 20%) Negative � 14%

Luminal B Positive Negative � 14%

Luminal B/HER2 Positive Positive

HER2-overexpression Negative Positive

Basal-like Negative Negative

Abbreviations: ER, estrogen receptor; PR, progesterone receptor.



cer subtype determined by immunohistochemistry.22 Onethousand and twelve patients diagnosed between 1990 and2002 at theMDAndersonCancer Center who did not receivechemotherapy or trastuzumab were included. Medianfollow-up was around 60 months. There were 771 hormone-receptor–positive, 98 HER2�, and 143 triple-negative breastcancers. Compared with patients with hormone-receptor–positive disease, patients withHER2� breast cancer had 4.98-times (95% CI, 2.91–8.53) the risk of worse RFS, and pa-tients with triple-negative breast cancer had 2.71-times (95%CI, 1.59–4.59) the risk of worse RFS. Amar et al. also evalu-ated the outcome of small (T1a/b) node-negative breast can-cers according to hormone receptor and HER2 status.23 Ofthe 421 tumors identifıed, 364 (86.5%) were HER2-, 28(6.7%) were HER2�, and 29 (6.9%) were triple-negativebreast cancers. The median follow-up time was only 1,015days. Unlike the previous study, 3%, 25%, and 27.6% ofHER2-, HER2�, and triple-negative patients with breastcancer received adjuvant chemotherapy, respectively, and17.8% of HER2� patients received trastuzumab-based adju-vant therapy. During the follow-up, the tumor recurred innine patients: four were HER2- tumors (1.1%), two wereHER2� tumors (7.1%), and three were triple-negative tu-mors (10.7%). Unfortunately, no correlation between immu-nohistochemistry and PAM50 test was performed in thisboth studies.Among ER� tumors, there is evident tumor heterogeneity

by gene-expression assays. In this way, luminal B tumors areassociated with poor breast cancer-specifıc survival than lu-minal A tumors.24 Both are usually ER� tumors. However,luminal B tumors are associated with increased expression ofproliferative genes resulting in higher tumor grade andKi-67index. In addition, up to 20% of luminal B tumors areHER2�.25 As stated above, ER- status, HER2� status, andhigh tumor grade adversely affect the prognosis of small(T1a/b) node-negative breast cancers. These characteristicsare very infrequent in luminal A tumors. Therefore, it couldindirectly be concluded that luminal A small (T1a/b) node-negative tumors are associated with the best outcome. Nev-ertheless, it would have been interesting to analyze in thepreviously mentioned studies the outcome of ER� patientstaking into account Ki-67 index and tumor grade. In thismanner, we would have been able to identify a well-defınedsubgroup of patients with ER�, small (T1a/b) node-negativebreast cancers with an excellent prognosis, essentially com-posed of luminal A tumors, who might not benefıt from ad-juvant chemotherapy without the need to use MammaPrint,Oncotype Dx, or even PAM50 recurrence test. Additionally,the risk of relapse of luminal B small (T1a/b) node-negativetumors would have also been reported.Unfortunately, there is no data about the prognostic effect

of PAM50 intrinsic subtyping among small (T1a/b) node-negative breast cancers. For this reason, it would be interest-ing to evaluate the outcome of these patients according tomolecular subtyping. Despite this, the breast cancer intrinsicsubtypes have shown prognostic signifıcance, remainingmore signifıcant inmultivariate analyses than other standard

parameters (ER status, histologic grade, tumor size, and nodestatus), and have also been used to generate a risk of recur-rence score.4 Therefore, the breast cancer intrinsic subtypesmay represent an important prognostic factor despite the tu-mor size. Considering that only the luminal A subtype con-tains low-risk patients, whereas the luminal B, HER2-enriched, and basal-like subtypes include intermediate- andhigh-risk tumors, the PAM50-based, low-risk luminal A tu-mors among small (T1a/b) node-negative breast cancerscould constitute a subgroup with a favorable prognosis thatwould not benefıt fromadjuvant chemotherapy. In regards tothe management of PAM50-based high-risk small (T1a/b)node-negative tumors, including all the intrinsic subtypes,would tumor stage (T1a vs. T1b vs. others) have an effect inthe benefıt derived from adjuvant systemic therapy? Shouldcancer staging guide the choice of the adjuvant systemic ther-apy, or should we rely only on the biologic behavior?What isthe breast cancer-specifıc mortality among the different mo-lecular subtypes without adjuvant systemic therapy? Thesequestions remain open andmost of themwill probably neverbe fully answered.In sum, the breast cancer intrinsic subtypes may represent

a relevant prognostic factor regardless of tumor size. Overall,triple-negative and HER2� breast cancers by immunohisto-chemistry are associated with a higher risk of recurrenceamong small (T1a/b) node-negative tumors.

CONCLUSIONAlthough the outcome of small (T1a/b) node-negative breasttumors is generally excellent, in the absence of prospectiveclinical trials, we are limited to clinical data derived from ret-rospective analyses. Different studies have shown that ayounger age at diagnosis, ER- status,HER2� status, and hightumor grade adversely affect the prognosis of these patientsand should be taken into consideration in order to defıne thebest adjuvant systemic therapy for this population.Additionally, several genomic assays are currently avail-

able to help better predict the outcome of breast cancer pa-tients. In this way, some small (T1a/b) node-negative breastcancers are classifıed as high-risk tumors byMammaPrint orOncotypeDx, in particular high-grade ER� and ER- tumors,andmight benefıt from adjuvant systemic therapy. However,the effect of tumor stage (T1a vs. T1b vs. others) in this ben-efıt, its influence in the selection of the adjuvant treatment,and the optimal chemotherapy regimen merit further evalu-ation.Finally, triple-negative and HER2� breast cancers by im-

munohistochemistry have been associated with a higher riskof recurrence among small (T1a/b) node-negative tumors.Unfortunately, there is no data about the prognostic impactof PAM50 intrinsic subtyping. For this reason, it would beinteresting to evaluate the outcome of these patients accord-ing tomolecular subtyping in the absence of systemic therapyin order to optimize the management of these tumors.



APPENDIX 1: TREATMENT ALGORITHM FOR SMALL(T1A/B) NODE-NEGATIVE BREAST TUMORSER-Positive/HER2-Negative1. T1a3adjuvant endocrine therapy2. T1b

• Adjuvant endocrine therapy in postmenopausalpatients with low-grade/intermediate-grade tumors(PR positive� 20% and Ki67� 14% are typically as-sociated with luminal A subtype and can help to bet-ter identify those patients that are unlikely to benefıtfrom adjuvant chemotherapy).

• Consider MammaPrint or Oncotype Dx mainly inpremenopausal women, and/or in patients withhigh-grade tumors:• High risk/high recurrence score3adjuvant che-motherapy � adjuvant endocrine therapy

• Low risk/low recurrence score3adjuvant endo-crine therapy

• Intermediate recurrence score3adjuvant endo-crine therapy

Triple-Negative Breast Cancer1. T1a3consider adjuvant chemotherapy in very young

women (age � 40), and/or in patients with high-gradetumors

2. T1b3consider adjuvant chemotherapy

ER-Positive/HER2-Positive1. T1a3adjuvant endocrine therapy. Consider adjuvant

chemotherapy plus trastuzumab in very young pa-

tients (age � 40), and/or in patients with high-gradetumors.

2. T1b3adjuvant endocrine therapy. Consider adjuvantchemotherapy plus trastuzumab.

ER-Negative/HER2-Positive1. T1a3consider adjuvant chemotherapy plus trastu-

zumab in very young patients (age � 40), and/or in pa-tients with high-grade tumors.

2. T1b3consider adjuvant chemotherapy plus trastu-zumab

Following these proposed guidelines, our recommendationfor a 40-year-old patient with a stage T1aN0 triple-negativebreast cancer (0.5 cm of diameter, high tumor grade, andKi67 index of 80%) would be chemotherapy with taxanes-based therapy.This recommendation is based on the following:1. ER-negative status, younger age at diagnosis, and high tu-

mor grade are independent adverse prognostic factorsamong small (T1a/b) node-negative breast cancers.10,12,13

2. Triple-negative breast cancers by immunohistochemis-try are associated with a higher risk of recurrenceamong small (T1a/b) node-negative tumors.22,23

3. Most basal-like tumors are triple-negative by immuno-histochemistry.25

4. Basal-like tumors show a high chemosensitivity.265. Basal-like tumors have a signifıcantly worse clinical

outcome.26



Employment or Leadership Position: None. Consultant or Advisory Role: Javier Cortes, Celgene; Novartis; Roche. Stock Ownership: None. Honoraria:Javier Cortes, Eisai. Research Funding: None. Expert Testimony: None. Other Remuneration: None.

References

1. Kennedy T, Stewart AK, Bilimoria KY, et al. Treatment trends and fac-tors associated with survival in T1aN0 and T1bN0 breast cancer pa-tients. Ann Surg Oncol. 2007;14:2918-2927.

2. Hilsenbeck SG, Ravdin PM, de Moor CA, et al. Time-dependence ofhazard ratios for prognostic factors in primary breast cancer. BreastCancer Res Treat. 1998;52:227-237.

3. Carter CL, Allen C, Henson DE. Relation of tumor size, lymph nodestatus, and survival in 24,740 breast cancer cases.Cancer. 1989;63:181-187.



6. van’t Veer LJ, Dai H, van de Vijver MJ, et al. Gene expression profıl-

ing predicts clinical outcome of breast cancer. Nature. 2002;415:530-536.


8. Buyse M, Loi S, van’t Veer L, et al. Validation and clinical utility of a70-gene prognostic signature forwomenwith node-negative breast can-cer. J Natl Cancer Inst. 2006;98:1183-1192.

9. NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines).Breast Cancer. Http://www.nccn.org/professionals/physician_gls/pdf/breast.pdf. Accessed March 10, 2013.

10. Fisher B, Dignam J, Tan-Chiu E, et al. Prognosis and treatment of pa-tients with breast tumors of one centimeter or less and negative axillarylymph nodes. J Natl Cancer Inst. 2001;93:112-120.



11. Rosen PP, Groshen S, Kinne DW, et al. Factors influencing prognosisin node-negative breast carcinoma: analysis of 767 T1N0M0/T2N0M0patients with long-term follow-up. J Clin Oncol. 1993;11:2090-2100.

12. Chia SK, Speers CH, Bryce CJ, et al. Ten-year outcomes in a population-based cohort of node-negative, lymphatic, and vascular invasion-negative early breast cancers without adjuvant systemic therapies. J ClinOncol. 2004;22:1630-1637.

13. Hanrahan EO, Gonzalez-Angulo AM, Giordano SH, et al. Overall sur-vival and cause-specifıc mortality of patients with stage T1a,bN0M0breast carcinoma. J Clin Oncol. 2007;25:4952-4960.

14. Gonzalez-Angulo AM, Litton JK, Broglio KR, et al. High risk of recur-rence for patients with breast cancer who have human epidermalgrowth factor receptor 2-positive, node-negative tumors 1 cm orsmaller. J Clin Oncol. 2009;27:5700-5706.

15. Curigliano G, Viale G, Bagnardi V, et al. Clinical relevance ofHER2 overexpression/amplifıcation in patients with small tumorsize and node-negative breast cancer. J Clin Oncol. 2009;27:5693-5699.

16. Mook S, Knauer M, Bueno-de-Mesquita JM, et al. Metastatic potentialof T1 breast cancer can be predicted by the 70-gene MammaPrint sig-nature. Ann Surg Oncol. 2010;17:1406-1413.

17. KnauerM,Mook S, Rutgers EJ, et al. The predictive value of the 70-genesignature for adjuvant chemotherapy in early breast cancer. Breast Can-cer Res Treat. 2010;120:655-661.

18. Perou CM, T, Eisen MB, et al. Molecular portraits of human breast tu-mours. Nature. 2000;406:747-752.

19. Goldhirsch A, Wood WC, Coates AS, et al. Strategies for subtypes—dealing with the diversity of breast cancer: highlights of the St. GallenInternational Expert Consensus on the Primary Therapy of Early BreastCancer 2011. Ann Oncol. 2011;22:1736-1747.

20. Prat A, Cheang MC, Martin M, et al. Prognostic signifıcance ofprogesterone receptor-positive tumor cells within immunohistochemically defıned luminal a breast cancer. J Clin Oncol. 2013;31:203-209.

21. Bastien RR, Rodríguez-Lescure A, EbbertMT, et al. PAM50Breast Can-cer Subtyping by RT-qPCR and concordancewith standard clinicalmo-lecular markers. BMCMed Genomics. 2012;5:44.

22. Theriault RL, Litton JK,Mittendorf EA, et al. Age and survival estimatesin patients who have node-negative T1ab breast cancer by breast cancersubtype. Clin Breast Cancer. 2011;11:325-331.

23. Amar S, McCullough AE, TanW, et al. Prognosis and outcome of small(��1 cm), node-negative breast cancer on the basis of hormonal andHER-2 status. Oncologist. 2010;15:1043-1109.

24. Cheang MC, Chia SK, Voduc D, et al. Ki-67 index, HER2 status, andprognosis of patients with luminal B breast cancer. J Natl Cancer Inst.2009;101:736-750.

25. Sotiriou C, Pusztai L. Gene-expression signatures in breast cancer.N Engl J Med. 2009;360:790-800.

26. Prat A, Parker JS, Karginova O, et al. Phenotypic and molecular char-acterization of the claudin-low intrinsic subtype of breast cancer. BreastCancer Res. 2010;12:R68.



Competing Risks in Low-Risk Breast CancerKathrin Strasser-Weippl, MD, MBA, and Paul E. Goss, MD, PhD

OVERVIEW

In recent years a growing amount of data on prognostic features of breast cancer has allowed for identification of tumors with a very lowrisk of recurrence. Markers used to predict the risk of distant spread include classic clinicopathologic features as well as newer tumor genesignatures, which have been validated and are being used in cohorts of patients with breast cancer patients who have low-risk disease.However, the definition of “low-risk” breast cancer requires consideration of patient-related factors such as comorbidities and age in additionto tumor characteristics, as high competing risks for mortality might be more important than cancer recurrence from a patient’s point of view.In addition, identification of low-risk breast cancer cohorts is only valuable if treatment decisions are based on this information. Therefore,the magnitude of any treatment benefit in low-risk disease needs to be quantified in a comprehensible way. However, treatment benefit inlow-risk situations is hard to quantify, may vary over time, and needs to be compared to individual risks for side effects. Decision modelsconsidering tumor and patient characteristics as well as predictive markers for treatment efficacy and toxicity will be needed. Tools such asAdjuvant! Online ultimately need to include information from gene signatures in order to reliably recommend specific treatment options forpatients with breast cancer patients who have low-risk disease.

Historically, the decision of whether, and how, to treatpatients with limited-stage breast cancer has been

based on clinicopathologic parameters such as tumor size orextent of nodal involvement.1 In recent years however, “atrend to using tumor biology rather than numerical diseaseindicators such as tumor size or extent of nodal involve-ment”1 has become increasingly common. This concept wasfırst implemented for deciding about the use of tamoxifen forestrogen receptor (ER)-positive breast cancer and developedfurther for decision making around the use of trastuzumabfor HER2� disease. In 2005, the St. Gallen InternationalBreast Cancer Conference divided endocrine responsivenessof breast cancers into three categories “highly endocrine re-sponsive,” “endocrine response uncertain,” and “endocrineunresponsive.“2 Increasingly, relying on predictive biomark-ers for treatment decisions has become the basis of decidingon cancer therapies and is leading to approval of targetedtherapies for many cancers.Despite the value of both predictive and prognostic bio-

markers in dividing high-risk patients from lower-risk pa-tients, inevitably it remains debatable whether those withvery low risk of recurrence should still be treated relativelyaggressively if biomarkers point to even a small therapeuticbenefıt. To make these decisions numerous factors needconsideration including tumor and patient characteristics,treatment side effects, economic deliberations, and patient

preference. In this article we discuss the nuances of some ofthese factors.

DEFINING “LOW-RISK” BREAST CANCERBefore considering treatment options for low-risk patientswith breast cancer, the term “low-risk disease” needs to bedefıned. Commonly, the term “low risk” has been used forsubtypes of breast cancer with very favorable prognosticmarkers, predicting “objectively and independently” favor-able clinical outcome “independent of systemic treatmentgiven.”3 However, as today it is usual practice to treat mostpatients with breast cancer with endocrine therapy or che-motherapy, it is increasingly diffıcult to estimate the prog-nostic value of a new marker independent of treatmenteffects. In addition, amarker can be both prognostic and pre-dictive, and the possibility of certain disease subtypes becom-ing “low risk” only in the presence of a specifıc treatmentneeds to be taken into account. A typical example would be ifcertain HER2� ER-positive cancers become “low risk” iftrastuzumab and endocrine therapy are administered.In addition, the label “low risk” should not be used merely

for breast cancers with a lower risk of recurrence or deaththan other breast cancers, because this ignores the patient’sperspective. Rather, “low risk” from a patient’s point of viewis a disease that has little or no risk of death and/ormorbiditywithout (systemic) treatment. A patient’s competing risks for

From the Harvard Medical School and Avon Breast Cancer Center of Excellence, Massachusetts General Hospital Cancer Center, Boston, MA; Center for Oncology, Hematology and Palliative Care,Vienna, Austria.


Corresponding author: Paul E. Goss, MD, PhD, Avon Breast Cancer Center of Excellence, Massachusetts General Hospital Cancer Center, 55 Fruit St., Lawrence House, LRH-302, Boston, MA 02114;email: [email protected].


STRASSER-WEIPPL AND GOSS


survival and quality of life, accounting for her expectation ofremaining quality-adjusted life years (QALY) without breastcancer need to be considered in addition to disease charac-teristics. This is a diffıcult task, as expected QALYs can onlybe roughly estimated andmay vary considerably between pa-tients on the basis of comorbidities, lifestyle, genetic, andother unknown factors.In summary, the identifıcation of patients with low-risk

breast cancer needs a two-tiered approach: (1) identifıcationof favorable tumor subtypes by prognostic and predictivebiomarkers; and (2) identifıcation of patients by markers ofquality of life and life expectancy who, with a specifıc tumorsubtype, will not be able to benefıt suffıciently from (moreaggressive) treatment during their remaining QALYs.

DOES “INDOLENT” BREAST CANCER EXIST AT ALL?When reviewing the prognostic spectrum of breast cancers,an important question is whether breast cancers can ever beconsidered “indolent,” that is whether there are patientswhose natural life expectancy will not be altered by the pres-ence of the disease. This question has recently been answeredin an indirect but very convincingmanner. By examining theincidence of early- and late-stage breast cancers in the Sur-veillance, Epidemiology, and End Results Program databasefrom 1976 through 2008, Bleyer et al. showed that in 2008invasive breast cancer was over-diagnosed in more than30,000 women or 20% of all invasive breast cancers diag-nosed that year in the United States.4 These and other datademonstrate that there are a number of patients with breastcancer who are treated for breast cancers that would not havecaused clinical illness had they remained undetected.5,6 Ifthere are patients whowill live to their normal life expectancywith no treatment at all, one may conclude that—across theprognostic spectrum of breast cancers— there are certainly anumber of patients with breast cancerwhowill remain free ofrecurrence after surgical (and possibly radiation) treatmentalone or for whom endocrine therapy is suffıcient. A major

goal of our efforts to tailor treatment to individual patients isto identify these groups.

TUMOR CHARACTERISTICSThe fırst step toward identifying “low-risk” patients withbreast cancer is to defıne favorable tumor characteristics thatmay be purely prognostic markers, or predictive markers fornontoxic therapies that allow for the omission of more ag-gressive treatment modalities.

TUMOR SIZE AND NODAL INVOLVEMENTIt has long been acknowledged that the prognosis of patientswith breast cancer varies with the size of the primary tumorand the extent of nodal involvement, because both indicatethe capacity of the tumor to cause distant spread.7-10 It wasalso shown in the 1970s—when tumor size and nodal in-volvement were the only markers available—that adjuvantchemotherapy improves the survival of patients with breastcancer with node-positive disease.11,12However, although all these studies consistently reported a

higher risk of recurrence and death with higher values ofthese numerical markers,13 they also showed that not all pa-tients with breast cancer with large tumors and/or extensivenodal involvement ultimately suffer disease recurrence.

ESTROGEN RECEPTOR STATUSThe ER is predominantly used for treatment decision mak-ing, because ER expression is highly correlated with responseto endocrine treatment.14 However, expression of the ER isalso a prognostic marker as the survival curves and the cor-responding hazard rates vary signifıcantly by ER expressionof the primary tumor15: the hazard rate for recurrence of ER-positive breast cancer is much lower than for ER-negativebreast cancer in the fırst years after diagnosis, but it remainsalmost unchanged over many years so that starting fromabout year 8 to 10, the yearly rate of recurrence of ER-positivebreast cancer is higher than for ER-negative disease.15,16

GRADING AND PROLIFERATIVE INDEXThe biomarkers tumor grade and proliferative index (Ki-67)correlate with each other17 and do so inversely with expres-sion of ER.18 The prognostic value of Ki-67 staining fordisease-free and overall survival independent of node statushas also been established in numerous studies and severalmeta-analyses.19-21 In addition, pretreatment Ki-67 scoresare predictive for response to chemotherapy.22,23

HER2 POSITIVITYOverexpression of theHER2 oncogene (HER2 positivity) hasbeen known to be of prognostic value in breast cancer formany years.24 When the monoclonal antibody trastuzumabbecame available, HER2 moved from being a merely prog-nostic to being amainly predictive biomarker for response toanti-HER2 agents.25 In addition, HER2 also seems to be pre-

KEY POINTS

� Breast cancer with a low risk of recurrence can be definedby clinicopathologic and molecular markers.

� The definition of “low-risk” breast cancer requiresconsideration of patient-related factors such as age andcomorbidities in addition to tumor characteristics.

� Identification of low-risk breast cancer cohorts is onlyvaluable if treatment decisions are based on thisinformation.

� Treatment benefit in low-risk situations is hard to quantifyand needs to be compared to the individual risk of sideeffects.

� Ultimately, decision models considering tumor and patientcharacteristics as well as predictive markers for treatmentefficacy and toxicity will be needed.

LOW-RISK BREAST CANCER


dictive for response to certain chemotherapies, including an-thracyclines and taxanes.26,27Like ER expression, HER2 positivity moved from being a

prognostic to a predictive marker, which in turn yields a fa-vorable prognostic value in the presence of anti-HER2 treat-ment.

“LOW-RISK” BY STANDARD CLINICOPATHOLOGICMARKERSWhen clinical and pathologic biomarkers are combined,breast cancers can be classifıed into subgroups with distinctpredictive and prognostic features. The “low-risk” categoryresulting from combinations of these markers (which is anapproximation of the breast cancer subtypes identifıed bygene-expression profıling, see below) comprises tumors thatare ER-positive, HER2-, with a low proliferative index, andlow or no involvement of the axillary lymph nodes.16,28-33 Inthe presence of systemic treatment, the risk of distant recur-rence for patients belonging to this category is below 5% at 5to 10 years.34 On the basis of clinical information and estimatesof ER, progesterone receptor, HER2, and Ki-67 (IHC4), sev-eral prognostic scores have been developed. In the neoadju-vant setting, the preoperative endocrine prognostic indexscore (PEPI) was able to identify three prognostically distinctgroups among a cohort of 158 patients treated with preoper-ative letrozole or tamoxifen for stage II or III breast cancer.35The IHC4prognostic index proved to be a strong predictor ofsurvival in cohorts of women with ER-positive breast cancertreated with adjuvant endocrine therapy.36Considering that responsiveness to chemotherapy of ER-

positive, HER2- tumors with a low proliferative index islow,37 it suggests that most of these low-risk patients wouldderive little or no benefıt from adjuvant chemotherapy. Foryoung patients without comorbidities and, thus, a highremaining-life expectancy, however, even small absolutebenefıts in distant recurrence rates might be worth pursuing.On the other hand, in the presence of a high competing riskof death from other causes, even tumors with a greater risk ofrecurrence might mean a comparably low(er) risk of deathfrom breast cancer.

GENE-EXPRESSION SIGNATURESIn the past two decades the development of new technologieshas allowed for the broad analysis of tumor tissue on a mo-lecular level. By using gene-expression–based assays, it hasbeen possible to identify breast cancers with highly favorableprognostic features. Several groups have used the results ofmultigene expression analyses in cohorts with a known out-come (i.e., study populations) to retrospectively developmultigene tests (called “signatures”) with a prognostic valuethat goes beyond that gained from standard clinico-pathologicmarkers (Table 2). For example it has been shownthat even tumors with a high grade or proliferation index,large size, and/or positive lymph nodesmay fall into the low-risk category as defıned by the 21-gene recurrence score (On-cotype DX).39-41

Although development of these signatures was initially “sur-vival driven,”42 they also allowed for a more precise pathologicdefınition of “intrinsic” breast cancer subtypes into luminal A,luminal B, HER2-enriched, and basal-like tumors.42,43 Thesesubtypes are mainly distinguished on the basis of proliferationand ER-regulated genes,42 and they therefore largely overlap,but are not identical with, the risk groups identifıed by standardclinicopathologic markers (Table 1).In addition to their prognostic value, gene-expression sig-

natures can be used to obtain predictive information regard-ing response to treatment. It has been shown that tumorswith favorable prognostic features (luminal A) may at thesame time have unfavorable predictive characteristics. Forexample in the neoadjuvant setting pathologic complete re-sponse rates of only 7% for luminal A comparedwith 36% forHER2-enriched and 43% for basal-like tumors to chemother-apy can be achieved.44

“LOW-RISK” BY GENE-EXPRESSION PROFILINGLow-risk groups defıned by gene-expression profıling consistmostly of a subset of luminal A tumors45 so that the value ofthe signatures is mainly in the ER-positive, HER2- cohort.46In this group, the independent prognostic information ofgene signatures beyond standard clinicopathologic markershas been established in numerous studies.39,47-52 Patientswith node-negative disease whose tumors fall into a low-riskmolecular category have an excellent prognosis, so that in twoongoing prospective studies (MINDACT and TAILORx) the

TABLE 1. “Surrogate Definition” of Intrinsic BreastCancer Subtypes Based on the St. Gallen Consensus201138

Intrinsic Subtype Clinicopathologic Definition

Luminal A Luminal A

ER� and/or PR�

HER2-

Ki-67 low (�14%)

Luminal B Luminal B (HER2-)

ER� and/or PR�

HER2-

Ki-67 high

Luminal B (HER2�)

ER� and/or PR�

Any Ki-67

HER2 overexpressed or amplified

ERBB2 overexpression HER2 � (nonluminal)

HER2 overexpressed or amplified

ER and PR absent

Basal-like Triple-negative (ductal)

ER and PR absent

HER2-

Abbreviations: ER, estrogen receptor; PR, progesterone receptor.



effect of omitting chemotherapy on survival is being evalu-ated (Table 2).Survival can also be predicted by gene-expression signa-

tures within cohorts with HER2� or triple-negative tumors,but even the prognostically favorable subgroups within thiscategory have a relatively high risk of relapse,53,54 confıningthe clinical utility of gene signatures to patients with in-creased competing risks of death in these cohorts.Importantly, most gene expression signatures were de-

veloped and validated in cohorts with node-negative breastcancer, making their benefıt in node-positive breast cancerquestionable.42 In different cohorts of patients with ER-positive, node-positive tumors, the recurrence rates inthe identifıed “low risk” groups were too high to consideromitting chemotherapy solely on the basis of prognosisand in the absence of competing risks for survival.41,49,55This, however, does not preclude the possibility thatintermediate-risk tumors constitute a much lower risk formortality compared with other competing risks in certainpatients.

PERSPECTIVES RELATED TO PATIENTS’ AGE ANDCOMORBIDITIESAsmentioned above, apart from tumor characteristics manypatient-specifıc parameters need to be considered whentreating patients with breast cancer. This is particularly truein the presence of “low-risk” disease, in which the absolutebenefıt from any treatment is bound to be small and compet-ing causes of deathmay considerably reduce the possibility oflong-term benefıt from adjuvant therapies.Among the patient-related factors, the patient’s age and

comorbidities act as the major competing causes of mor-tality. In a study of 900 women with early breast cancer,those with more than two comorbid conditions were 20-times more likely to die from non-breast cancer causes.60Any benefıt from systemic adjuvant treatment thereforeneeds to be weighed against these competing causes ofdeath.61A diagnosis of breast cancer with intermediate-risk fea-

turesmight turn into a “low-risk” situation in the presence ofa higher likelihood of death from other causes. On the otherhand, up to 40% of patients with breast cancer older than age80will die from their cancer,62 and age itself is known as a riskfactor for under treatment, particularly when life expectancyis underestimated.63 Even in a very low-risk situation, such as

extended adjuvant endocrine therapy (in women who havealready survived 5 years without recurrence) it has beenshown that breast cancer accounts for 30% of the observedmortality in patients age 70 or older.64 The dilemma in thetreatment of elderly women with breast cancer is aggravatedby the fact that frail and older individuals are often not in-cluded or not represented adequately in clinical trials becauseof inclusion criteria precluding their study participation or aselection bias.On the other hand, in young and fıt patients with high

life expectancy, treatment decisions need to be based on stud-ies with extended follow-up. As mentioned, the recurrencerate of ER-positive tumors remains high during years 5 to 10and beyond, exceeding the hazard rate of the assumedhigher-risk ER-negative tumors. However, most molecularbiomarkers used today have been developed and validatedusing outcome in the fırst years after breast cancer diagnosisonly.41,48 When the Breast Cancer Index (BCI) signature wascompared with Oncotype DX and the IHC4 in patientstreated with endocrine therapy in the ATAC trial, only theBCI yielded signifıcantly more prognostic information thanthe “clinical tumor score” (involved nodes, tumor grade andsize, age, and treatment) for years 5 to 0 and beyond (p �0.0005).65As yet, there are few clinical tools that consider tumor

and patient characteristics for breast cancer at the sametime. Adjuvant! Online (AO) is an online tool attemptingto estimate the natural mortality of a patient with breastcancer as well as the baseline prognosis of the disease andtreatment effects considering both patient-related factorssuch as age and comorbidities as well clinicopathologictumor characteristics. On the basis of these data, the ab-solute benefıt of systemic adjuvant therapies is estimatedup to 10 years after diagnosis. Despite its many limita-tions AO is an important tool taking into account the pa-tient’s remaining life expectancy as well as tumor character-istics.In the future, the information derived from gene-expression

signatures of breast tumors needs to be combined with (notmerely compared to) patient-related factors to be able to cor-rectly estimate the benefıt fromadjuvant treatment options. Forexample, in a retrospective study in 265 tamoxifen-treated pa-tients with breast cancer, the addition of AO to BCI improvedpredictive accuracy of the gene signature.66

TABLE 2. Multigene Signatures with Prognostic Value

Gene Signature No. of Genes Assessed Tissue Needed Ongoing Prospective Trials Ref.

MammaPrint 70 FF MINDACT (ER�, N�4) 56

Oncotype DX 21 FFPE TAILORx (ER�, HER2-, N0) 57

PAM50 risk of relapse (ROR) 55 FFPE 58

Breast cancer index 7 (combines HOXB13/IL17BR and5-gene molecular grade index)

FFPE 51,59

FF: fresh-frozen; FFPE: formalin-fixed paraffin-embedded.



SPEAKING OF “TREATMENT BENEFIT”As discussed, patients with “low-risk” breast cancer can beidentifıed when both tumor and patient characteristics arecombined. However, the identifıcation of these patients isonly valuable if treatment decisions are based on both ofthese. Therefore other parameters, for example, expectedtreatment benefıt and toxicity, need to be taken into consid-eration.Onemight argue that if the risk of recurrence is smallenough, any possible absolute treatment benefıt is so smallthat (certain) systemic therapies can be safely omitted. How-ever, in healthy patients with a long life expectancy, even asmall absolute treatment benefıt might be worth pursuing iftoxicity is low.The “low-risk” breast cancer subgroup identifıed by gene-

expression analysis is characterized by high ER expressionand low markers of proliferation (Ki-67), thus suggesting alow expected benefıt from adjuvant chemotherapy (seeabove) and a high sensitivity to endocrine treatment options.Considering, in addition, the relatively constant risk of recur-rence (hazard rate) in ER-positive breast cancer over a longperiod of time after diagnosis, this argues for using endocrinetreatment over a long period of time in these patients. Whenlooking at treatment benefıt, however, one needs to keep inmind that there are numerous ways to describe such benefıt,including the hazard ratio of two survival curves, the ab-solute benefıt of treatments at certain time points, measuresof time such as median survival, time-to-event end points(like overall or progression-free survival), and more “eco-nomical” outcome measures (like the number-needed-to-treat or “cost per QALY”). In current oncologic literature,the hazard ratio in combination with a measure of time,commonly median survival, is generally used; however, it isnot clear whether this is the best way to discuss treatmentbenefıt for patients with low-risk breast cancer. It was shownby the EBCTCG that the hazard ratios achieved by variouschemotherapies are more or less the same across most risksubgroups,67 but the absolute benefıt and the number-needed-to-treat (being the inverse of the absolute benefıt),are not. The magnitude of absolute benefıt depends on thetime when it is measured (assuming constant hazard rates)and tends to increase with longer follow-up, so that it maybe underestimated in studies with short follow-up. It is alsoknown from psychological studies that outcomes that are al-most certain (e.g., a 96% chance of remaining recurrence-free) are underweighted and that people are willing to paydisproportionally high prices to achieve (near-)certainties(e.g., a 100% chance of remaining recurrence-free).68 An ab-solute increase in the chance of survival from 93% to 99%might therefore mean much more to a patient consideringsystemic treatment than an absolute increase from 60% to66% and if absolute benefıts are measured at different timepoints that adds to this complexity. Ultimately, the choice ofadjuvant systemic treatment might be very subjective andstrongly related to the risk/benefıt ratio of a particular treat-ment.

BENEFIT COMPARED WITH SIDE EFFECTSIn treatment situations with low expected benefıt and a longfollow-up, the risk/benefıt ratio of any therapy becomes anessential parameter. Side effects are usually diligently re-corded in clinical trials, and thewillingness to accept a certainlevel of toxicity is higher in oncology than in other areas ofmedicine. In low-risk cohorts with near-normal life expec-tancies, however, toxicitiesmight be the deciding factor. Thishas been extensively shown in the breast cancer preventionsituation, in which numerous clinical trials have reported ameasurable benefıt for selective ER modulators and an aro-matase inhibitor at little toxicity,69,70 but still only a minorityof women uses these agents for prevention.71 It has been ar-gued that the diffıculty of weighing multiple risks comparedwith multiple benefıts, the general risk aversion of hu-mans,71,72 and the fact that treatment side effects are oftenexperienced early whereas benefıts come later, are responsi-ble for this phenomenon. In addition, the risk for certain sideeffects may vary considerably between individuals and overtime, so that general recommendations are hard to make insituations in which the threshold for unacceptable toxicity isvery low.

CONCLUSIONOver the past decade, understanding of breast cancer hasevolved from being considered a single tumor type to now amultitude of diseases with different prognostic and predic-tive features. Understanding the disease on a molecular levelhas pointed to the existence of tumors that harbor an excep-tionally low risk of recurrence, which is undoubtedly the fırststep toward identifying “low-risk” breast cancer. However,there are many other patient- and treatment-related factorsthat need to be considered to allow reliable classifıcation ofpatients into a “low-risk” group. If, by “low risk,” we meanbreast cancer that will not alter a patient’s expected QALYs,then her competing risks for morbidity and mortality mustbe taken into consideration. In addition, when talking abouttreatment options in this setting, the complex issue of treat-ment benefıts that can bemeasured— and framed—in differ-ent ways, needs to be discussed. Those small absolutetreatment benefıts must then be contrasted with expectedside effects, which are a fundamental basis of treatment de-cisions in low-risk situations with long treatment durations.Tools that take into consideration molecular information

from the tumor as well as markers of patients’ life expectancyandquality of lifewill ultimately be needed to provide reliabletreatment recommendations for patients with breast cancerwho have low-risk disease.In ongoing research, somatic mutations in breast cancer

are being analyzed further, including identifıcation of“driver” and “passenger” mutations and development ofmodels that address the time dependence of breast cancerrecurrence.73,74





Employment or Leadership Position: None. Consultant or Advisory Role: None. Stock Ownership: None. Honoraria: Paul E. Goss, GlaxoSmithKline;Pfizer. Research Funding: None. Expert Testimony: None. Other Remuneration: None.

References

1. Gnant M, Harbeck N, Thomssen C. St. Gallen 2011: Summary of theConsensus Discussion. Breast Care. 2011;6:136-141.

2. Goldhirsch A, Glick JH, Gelber RD, et al. Meeting highlights: Interna-tional expert consensus on the primary therapy of early breast cancer2005. Ann Oncol. 2005;16:1569-1583.

3. Weigel MT, Dowsett M. Current and emerging biomarkers in breastcancer: prognosis and prediction. Endocr Relat Cancer. 2010 Sep 23;17:R245-62.

4. Bleyer A, Welch HG. Effect of three decades of screening mammogra-phy on breast-cancer incidence. N Engl J Med. 2012;367:1998-2005.

5. Zahl PH, Maehlen J, Welch HG. The natural history of invasive breastcancers detected by screening mammography. Arch Intern Med. 2008;168:2311-2316.

6. Welch HG, Schwartz LM, Woloshin S. Ramifıcations of screening forbreast cancer: 1 in 4 cancers detected bymammography are pseudocan-cers. BMJ. 2006;332727.

7. Koscielny S, Le MG, Tubiana M. The natural history of human breastcancer. The relationship between involvement of axillary lymph nodesand the initiation of distant metastases. Br J Cancer. 1989;59:775-782.

8. Valagussa P, Brambilla C, Zambetti M, et al. G. Salvage treatments inrelapsing resectable breast cancer. Recent Results Cancer Res. 1989;115:69-76.

9. Fisher B, SlackN, KatrychD, et al. Ten year follow-up results of patientswith carcinoma of the breast in a co-operative clinical trial evaluatingsurgical adjuvant chemotherapy. Surg Gynecol Obstet. 1975;140:528-534.

10. Ferguson DJ, Meier P, Karrison T, et al. Staging of breast cancer andsurvival rates. An assessment based on 50 years of experience with rad-ical mastectomy. JAMA. 1982;248:1337-1341.

11. Fisher B, Carbone P, Economou SG, et al. 1-Phenylalanine mustard (L-PAM) in the management of primary breast cancer. A report of earlyfındings. N Engl J Med. 1975;292:117-122.

12. Bonadonna G, Valagussa P. Current status of adjuvant chemotherapyfor breast cancer. Semin Oncol. 1987;14:8-22.

13. Quiet CA, Ferguson DJ, Weichselbaum RR, et al. Natural history ofnode-positive breast cancer: The curability of small cancers with a lim-ited number of positive nodes. J Clin Oncol. 1996;14:3105-3111.

14. Early Breast Cancer Trialists’ Collaborative Group. Effects of chemo-therapy and hormonal therapy for early breast cancer on recurrence and15-year survival: An overview of the randomised trials. Lancet. 2005;365:1687-1717.

15. Anderson WF, Chen BE, Jatoi I, et al. Effects of estrogen receptor ex-pression and histopathology on annual hazard rates of death frombreast cancer. Breast Cancer Res Treat. 2006;100:121-126.

16. Dignam JJ, Dukic V, Anderson SJ, et al. Hazard of recurrence and ad-juvant treatment effects over time in lymph node-negative breast can-cer. Breast Cancer Res Treat. 2009;116:595-602.

17. Trihia H,Murray S, Price K, et al. Ki-67 expression in breast carcinoma:Its association with grading systems, clinical parameters, and otherprognostic factors—a surrogate marker? Cancer. 2003;97:1321-1331.

18. Haerslev T, Jacobsen GK, Zedeler K. Correlation of growth fraction byKi-67 and proliferating cell nuclear antigen (PCNA) immunohisto-chemistry with histopathological parameters and prognosis in primarybreast carcinomas. Breast Cancer Res Treat. 1996;37:101-113.

19. Urruticoechea A, Smith IE, Dowsett M. Proliferation marker Ki-67 inearly breast cancer. J Clin Oncol. 2005 Oct 1;23(28):7212-20.

20. Azambuja E, Durbecq V, Rosa DD, et al. HER-2 overexpression/ampli-fıcation and its interaction with taxane-based therapy in breast cancer.Ann Oncol. 2008;19:223-232.

21. Stuart-Harris R, Caldas C, Pinder SE, et al. Proliferation markers andsurvival in early breast cancer: A systematic review andmeta-analysis of85 studies in 32,825 patients. Breast. 2008;17:323-334.

22. Chang J, Powles TJ, AllredDC, et al. Prediction of clinical outcome fromprimary tamoxifen by expression of biologic markers in breast cancerpatients. Clin Cancer Res. 2000 Feb;6:616-621.

23. Faneyte IF, Schrama JG, Peterse JL, et al. Breast cancer response to neo-adjuvant chemotherapy: predictive markers and relation with outcome.Br J Cancer. 2003;88:406-412.

24. Slamon DJ, Clark GM, Wong SG, et al. Human breast cancer: Correla-tion of relapse and survival with amplifıcation of the HER-2/neu onco-gene. Science. 1987;235:177-182.

25. Mass RD, Press MF, Anderson S, et al. Evaluation of clinical outcomesaccording to HER2 detection by fluorescence in situ hybridization inwomen with metastatic breast cancer treated with trastuzumab. ClinBreast Cancer. 2005;6:240-246.

26. Dressler LG, BerryDA, BroadwaterG, et al. Comparison ofHER2 statusby fluorescence in situ hybridization and immunohistochemistry topredict benefıt from dose escalation of adjuvant doxorubicin-basedtherapy in node-positive breast cancer patients. J Clin Oncol. 2005;23:4287-4297.

27. HayesDF, Thor AD,Dressler LG, et al. HER2 and response to paclitaxelin node-positive breast cancer. N Engl J Med. 2007;357:1496-1506.

28. Nielsen TO,Hsu FD, Jensen K, et al. Immunohistochemical and clinicalcharacterization of the basal-like subtype of invasive breast carcinoma.Clin Cancer Res. 2004;10:5367-5374.

29. Blows FM, Driver KE, Schmidt MK, et al. Subtyping of breast cancer byimmunohistochemistry to investigate a relationship between subtypeand short and long term survival: A collaborative analysis of data for10,159 cases from 12 studies. PLoS medicine. 2010;7:e1000279.

30. Hugh J, Hanson J, Cheang MC, et al. Breast cancer subtypes and re-sponse to docetaxel in node-positive breast cancer: Use of an immuno-histochemical defınition in the BCIRG 001 trial. J Clin Oncol. 2009;27:1168-1176.

31. Cheang MC, Chia SK, Voduc D, et al. Ki67 index, HER2 status, andprognosis of patients with luminal B breast cancer. J Natl Cancer Inst.2009;10:736-750.

32. Phipps AI, Buist DS, Malone KE, et al. Reproductive history and risk ofthree breast cancer subtypes defıned by three biomarkers. CancerCauses Control. 2011;22:399-405.



33. Liedtke C, Mazouni C, Hess KR, et al. Response to neoadjuvant therapyand long-term survival in patients with triple-negative breast cancer.J Clin Oncol. 2008;26:1275-1281.

34. Nguyen PL, Taghian AG, Katz MS, et al. Breast cancer subtype approx-imated by estrogen receptor, progesterone receptor, andHER-2 is asso-ciatedwith local and distant recurrence after breast-conserving therapy.J Clin Oncol. 2008;26:2373-2378

35. Ellis MJ, Tao Y, Luo J, et al. Outcome prediction for estrogen receptor-positive breast cancer based on postneoadjuvant endocrine therapy tu-mor characteristics. J Natl Cancer Inst . 2008;100:1380-1388.

36. Cuzick J, Dowsett M, Pineda S, et al. Prognostic value of a combinedestrogen receptor, progesterone receptor, Ki-67, and human epidermalgrowth factor receptor 2 immunohistochemical score and comparisonwith the Genomic Health recurrence score in early breast cancer. J ClinOncol. 2011;29:4273-1278.

37. Aebi S, Sun Z, BraunD, et al. Differential effıcacy of three cycles of CMFfollowed by tamoxifen in patients with ER-positive and ER-negative tu-mors: Long-term follow up on IBCSG Trial IX. Ann Oncol. 2011;22:1981-1987.

38. Goldhirsch A, Wood WC, Coates AS, et al. Strategies for subtypes-dealing with the diversity of breast cancer: Highlights of the St. GallenInternational Expert Consensus on the Primary Therapy of Early BreastCancer 2011. Ann Oncol Aug. 2011;22:1736-1747.


40. O Gluz, H Kreipe, T Degenhardt. Prospective Comparison of Risk As-sessment Tools in Early Breast Cancer (Recurrence Score, uPA/PAI-1,Central Grade, and Luminal Subtypes): Final CorrelationAnalysis fromthe Phase III WSG-Plan B Trial. Cancer Res. 2011;71(suppl):107s.

41. Albain KS, BarlowWE, Shak S, et al. Prognostic and predictive value ofthe 21-gene recurrence score assay in postmenopausal women withnode-positive, oestrogen-receptor-positive breast cancer on chemo-therapy: A retrospective analysis of a randomised trial. Lancet Oncol.2010;11:55-65.

42. Prat A, Ellis MJ, Perou CM. Practical implications of gene-expression-based assays for breast oncologists. Nat Rev Clin Oncol. 2012;9:48-57.


44. Parker JS. Breast cancer molecular subtypes predict response toanthracycline/taxane-based chemotherapy.Cancer Res. 2009;69(suppl):a2019.

45. Prat A, Parker JS, Fan C. Concordance among gene-expression-basedpredictors for ER-positive breast cancer treated with adjuvant tamox-ifen. Ann Oncol. 2012;23:2866-2873.

46. Fan C, Prat A, Parker JS, et al. Building prognostic models for breastcancer patients using clinical variables and hundreds of gene expressionsignatures. BMC medical genomics. 2011;4:3.

47. Habel LA, Shak S, Jacobs MK. A population-based study of tumor geneexpression and risk of breast cancer death among lymph node-negativepatients. Breast Cancer Res. 2006;8:R25.

48. Buyse M, Loi S, van’t Veer L, et al. Validation and clinical utility of a70-gene prognostic signature forwomenwith node-negative breast can-cer. J Natl Cancer Inst. 2006;98:1183-1192.

49. van de Vijver MJ, He YD, van’t Veer LJ, et al. A gene-expression signa-ture as a predictor of survival in breast cancer. N Engl J Med. 2002;347:1999-2009.

50. Cardoso F, Van’t Veer L, Rutgers E, Loi S, Mook S, Piccart-GebhartMJ. Clinical application of the 70-gene profıle: The MINDACT trial.J Clin Oncol. 2008;26:729-735.

51. MaXJ, Salunga R,Dahiya S, et al. A fıve-genemolecular grade index and

HOXB13:IL17BR are complementary prognostic factors in early stagebreast cancer. Clin Cancer Res. 2008;14:2601-2608.

52. Jerevall PL,MaXJ, LiH, et al. Prognostic utility ofHOXB13:IL17BR andmolecular grade index in early-stage breast cancer patients from theStockholm trial. Br J Cancer. 2011;104:1762-1769.

53. Alexe G, Dalgin GS, Scanfeld D, et al. High expression of lymphocyte-associated genes in node-negative HER2� breast cancers correlateswith lower recurrence rates. Cancer Res. 2007;67:10669-10676.

54. Teschendorff AE, Miremadi A, Pinder SE, et al. An immune responsegene expressionmodule identifıes a good prognosis subtype in estrogenreceptor negative breast cancer. Genome Biol. 2007;8:R157.

55. Dowsett M, Cuzick J, Wale C, et al. Prediction of risk of distant recur-rence using the 21-gene recurrence score in node-negative and node-positive postmenopausal patients with breast cancer treated withanastrozole or tamoxifen: A TransATAC study. J Clin Oncol. 2010;28:1829-1834.

56. Bueno-de-Mesquita JM, Linn SC, Keijzer R, et al. Validation of 70-geneprognosis signature in node-negative breast cancer. Breast Cancer ResTreat. 2009; 117:483-495.

57. Sparano JA, Paik S. Development of the 21-gene assay and its applica-tion in clinical practice and clinical trials. J Clin Oncol. 2008;26:721-708.

58. Nielsen TO, Parker JS, Leung S, et al. A comparison of PAM50 intrinsicsubtyping with immunohistochemistry and clinical prognostic factorsin tamoxifen-treated estrogen receptor-positive breast cancer. CancerRes. 2010;16:5222-5232.

59. Goss PE, Ingle JN,Martino S, et al. Randomized trial of letrozole follow-ing tamoxifen as extended adjuvant therapy in receptor-positive breastcancer: Updated fındings from NCIC CTG MA. 17. J Natl Cancer Inst.2005;97:1262-1271.

60. Satariano WA, Ragland DR. The effect of comorbidity on 3-year sur-vival of women with primary breast cancer. Ann Intern Med. 1994;120:104-110.

61. Biganzoli L, Wildiers H, Oakman C, et al. Management of elderlypatients with breast cancer: Updated recommendations of the Inter-national Society of Geriatric Oncology (SIOG) and European Societyof Breast Cancer Specialists (EUSOMA). Lancet Oncol. 2012;13:e148-160.

62. Rosso S, Gondos A, Zanetti R, et al. Up-to-date estimates of breast can-cer survival for the years 2000-2004 in 11 European countries: The roleof screening and a comparison with data from the United States. Eur JCancer. 2010;46:3351-3357.

63. Bouchardy C, Rapiti E, Fioretta G, et al. Undertreatment strongly de-creases prognosis of breast cancer in elderly women. J Clin Oncol. 2003;21:3580-3587.

64. Chapman JA, Meng D, Shepherd L, et al. Competing causes of deathfrom a randomized trial of extended adjuvant endocrine therapy forbreast cancer. J Natl Cancer Inst. 2008;100:252-260.

65. Sgroi DS, Cuzick J, Zhang Y, et al. Comparative Performance of BreastCancer Index (BCI) vs. OncotypeDx and IHC4 in the Prediction of LateRecurrence in Hormonal Receptor-Positive Lymph Node-NegativeBreast Cancer Patients: A TransATAC Study. Cancer Res. 2012;72(suppl):S1-9.

66. Jankowitz RC, Cooper K, Erlander MG, et al. Prognostic utility of thebreast cancer index and comparison to Adjuvant! Online in a clinicalcase series of early breast cancer. Breast Cancer Res. 2011;13:R98.

67. Early Breast Cancer Trialists’ Collaborative Group, Peto R, Davies C, etal. Comparisons between different polychemotherapy regimens forearly breast cancer: Meta-analyses of long-term outcome among100,000 women in 123 randomised trials. Lancet. 2012;379:432-444.

68. Tversky A, Kahneman D. Rational Choice and the Framing of Deci-sions. The Journal of Business. 1986;59:S251-S278.

69. Fisher B,Costantino JP,WickerhamDL, et al. Tamoxifen for prevention



of breast cancer: Report of the National Surgical Adjuvant Breast andBowel Project P-1 Study. J Natl Cancer Inst. 1998;90:1371-1388.

70. Goss PE, Ingle JN, Ales-Martínez JE, et al. Exemestane for breast-cancerprevention in postmenopausal women. N Engl J Med. 2011;364:2381-2391.

71. Waters EA, Cronin KA, Graubard BI, et al. Prevalence of tamoxifen usefor breast cancer chemoprevention among U.S. women. Cancer Epide-miol Biomarkers Prev. 2010;19:443-446.

72. Kahnemann D, Tversky A. Choices, Values and Frames.American Psy-chologist. 1984;39: 341-350.

73. Stephens PJ, Tarpey PS, Davies H, et al. The landscape of cancergenes and mutational processes in breast cancer. Nature. 2012;486:400-404.

74. Esserman LJ, Moore DH, Tsing PJ, et al. Biologic markers determineboth the risk and the timing of recurrence in breast cancer. Breast Can-cer Res Treat. 2011;129:607-616.



Do We Need to Irradiate All Small Invasive Breast Cancersand DCIS?Julia White, MD

OVERVIEW

Breast radiotherapy after lumpectomy is considered standard for nearly all patients with invasive breast cancer and is recommendedfor many patients after lumpectomy for ductal carcinoma in situ (DCIS). However, there is recognition that lumpectomy alone canachieve optimal cancer control for some patients with invasive breast cancer and DCIS. Patients with breast cancers with lower riskof recurrence are less likely to derive benefit from breast radiotherapy. This review will focus on defining populations of patients withinvasive breast cancer and DCIS with a low risk of recurrence post-lumpectomy and the evidence supporting omission of breastradiotherapy post-lumpectomy.

Breast radiotherapy after lumpectomy is considered stan-dard for nearly all patients with invasive breast cancer

and is recommended formany patients after lumpectomy forDCIS. On the surface, the application of post-lumpectomyradiotherapy for these two disease entities seems to be a sim-ilar issue. In reality, they represent two separate clinicalman-agement questions because the goals of post-lumpectomybreast radiotherapy are different for DCIS than for invasivebreast cancer. Regardless, in each case there is recognitionthat some patients with DCIS and invasive breast cancer willhave achieved optimal cancer control in the breast withlumpectomy alone. It is these cases that will not have a clin-ically meaningful benefıt from post-lumpectomy breast ra-diotherapy. For populations of patients with breast cancerwith lower risk of in-breast recurrences after lumpectomy,the debate about the relative benefıt of breast radiotherapycontinues. Ideally, ongoing genomic and molecular researchwill lead to ameans to accurately characterize an individual’spersonalized prognosis for local cancer recurrence such thata patient with breast cancer could safely decide to omit breastradiotherapy post-lumpectomy without worrying about thetrade-off of inferior outcomes.

INVASIVE BREAST CANCERFor patients with invasive breast cancer, the primary goal ofbreast radiotherapy after lumpectomy is to maximize in-breast cancer control so it is equivalent to that achieved bymastectomy and minimizes risk for breast cancer mortality.Secondary but important goals are to preserve a sensate, cos-metically appealing breast and to avoid mastectomy. The

benefıt of breast radiotherapy after lumpectomy has beenstudied in numerous clinical trials that have randomizedwomen post-lumpectomy to breast radiotherapy comparedwith observation.1-5 Each of these trials demonstrated ahighly signifıcant reduction in cancer recurrence (typically70% to 75%) in the treated breast with addition of breastradiotherapy.1-5 Despite fairly large reductions in local recur-rence, none of these trials individually documented a signif-icant improvement in breast cancer survival fromradiotherapy after lumpectomy. However, on meta-analysesof these trials, reduction in local regional recurrence (LRR)after lumpectomy by radiotherapy resulted in a reduction inbreast cancer mortality. This was demonstrated in the EarlyBreast Cancer Trialists Group (EBCTCG)meta-analsyis thatstudied the relationship of local regional recurrence at 5 yearsand breast cancer mortality at 15 years in 7,311 patient leveldata from 10 randomized clinical trials.8 The 5-year local re-currence was 7% among those allocated radiotherapy and26% among those observed, corresponding to an absolute re-duction of 19%. The 15-year risk of death as a result of breastcancer was 30.5% among those allocated post-lumpectomyradiotherapy and 35.9% among those observed (correspond-ing to an absolute breast cancer mortality reduction of 5.4%,SE 1�7). Themagnitude in reduction of breast cancer mortal-ity is much smaller than that for local recurrence but still ahighly signifıcant breast cancer death rate ratio of 0.83, SE0.05, 95% CI 0.75–0.91, 2, p � 0.0002. Subgroup analysesdemonstrated that the absolute benefıt produced by radio-therapy was determined principally by the magnitude of thelocal recurrence risk in unirradiatedwomen. In particular, ananalyses that divided absolute local recurrence risk reduction

From the Department of Radiation Oncology, The James, Ohio State University Comprehensive Cancer Center.


Corresponding author: Julia White, MD, Department of Radiation Oncology, The James, Ohio State University Comprehensive Cancer Center, 300 W. 10th Ave., Ste. 094D, Columbus, OH 43210;email: [email protected].


JULIA WHITE


after lumpectomy or mastectomy by 5 years into three cate-gories of less than 10%, 10% to 20%, or greater than 20%demonstrated that for those with less than 10% absolute re-duction in local recurrence by 5 years no improvement inbreast cancer mortality by 15 years is seen. A more recentmeta-analysis from the EBCTCG of 10,801 women in 17 tri-als further explored the effect of post-lumpectomy breast ra-diotherapy on 10-year overall breast cancer recurrence rate(locoregional or distant) and 15-year breast cancer mortalityeffects.9 The 10-year risk of any (locoregional or distant) fırstrecurrence was 19.3% in women allocated to radiotherapyand 35.0% in women allocated to breast-conserving surgeryonly, corresponding to an absolute risk reduction of 15.7%(95%CI 13.7–17.7, 2p�0.00001). Radiotherapy also reducedbreast cancer death at 15 years by a moderate amount: 25.2%without and 21.4% with radiotherapy, respectively, for 15-year absolute risk reduction of 3.8% (95% CI 1.6–6.0, 2p �0.00005). Analysis of 7,287 pathologically node-negative pa-tients by patient and tumor characteristics revealed that theabsolute recurrence reduction varied strongly by patient age,tumor grade, tumor size, and ER status. Older women andthose with low-grade tumors had a smaller risk of recurrencewithout radiotherapy than younger women and those withhigh-grade tumors. Each woman with pN0 disease was as-signed an individually predicted absolute reduction in 10-year recurrence risk from radiotherapy that was categorizedinto three risk groups: lower (� 10%), intermediate (10%to 19%), and large (� 20%). The 10-year recurrence riskswithout radiotherapy was 50.3%, 24.8%, 18.9%, comparedwith 26%, 12.4%, and 12% (absolute reductions of 24.3%,12.4%, and 6.9%) with radiotherapy for patients in the large,

intermediate, and lower risk groups, respectively. The corre-sponding absolute reductions in 15-year risk of breast cancerdeath in the three groups were 7.8% (95% CI 3.1–12.5), 1.1%(�2.0 to 4.2), and 0.1% (�7.5 to 7.7), respectively. Theseanalyses suggest that patients with breast cancer post-lumpectomy with an anticipated � 10% absolute reductionin locoregional risk by 5 years and overall recurrence (loco-regional and distant) by 10 years can omit breast radio-therapy without risk of excess breast cancer mortality.Furthermore, older patients with node-negative, less than2 cm, ER positive, low grade breast cancer who have takentamoxifen characterize this lower risk group.Table 1 lists several randomized controlled trials designed to

evaluate the benefıt of breast radiotherapy in patients consid-ered to be at low risk of recurrence post-lumpectomy.3,10-15The breast cancer populations in these trials were all node-negative, typically older than age 50, had received antiendo-crine therapy for cancers less than 2 cm in size that wereER/PR�, and grade 1–2. In the NSABP B21 and PMH trials,high grade lesions were associated with higher rates of in-breast recurrence.3,10 In contrast, small, grade 1 breast can-cers, such as those enrolled on the BASO II study that has hadlimited reporting so far, have been associated with low recur-rence rates after lumpectomy alone.11,14,15Two trials focused specifıcally on populations of older pa-

tients with breast cancer and provide additional informationabout the secondary goals of breast conservation: namely,cosmetic appearance and avoiding mastectomy. CALGB9343 randomized post-lumpectomy 626 women older thanage 70 with hormone receptor positive breast cancer that in-tended to take 5 years of tamoxifen, to breast radiation com-pared with observation.12 The most recent reporting of thistrial at 10.5 years follow-up demonstrated a low rate of in-breast recurrence of 8% without compared with 2% with ra-diotherapy. Notably, 43% of women participating in thisstudy had died, but only 7% as a result of breast cancer, dem-onstrating that in this older population comorbidity wasmore frequently the primary health risk than the breast can-cer. There was statistically similar freedom frommastectomyafter lumpectomy alone; 96% compared with 98% with theaddition of breast radiotherapy.12 A SEER-Medicare data-base that identifıed 8,724 women older than age 70 who weretreated with lumpectomy for small, lymph node-negative,hormone receptor positive breast cancer demonstrated goodagreement with CALGB 9242.16 In this study, by 8 years offollow-up, women who received breast radiotherapy post-lumpectomy had a 5.7% absolute benefıt; 8% ipsilateralbreast cancer event without compared with 2.3% with treat-ment. Healthy women ages 70 to 79 were more likely to ex-perience the benefıt associated with radiation therapy.However, a separate study of the SEER-Medicare observa-tional cohort showed that omission of radiotherapy resultedin higher rates of mastectomy in 7,074 women aged 70 to 79with receptor positive, node-negative breast cancer, less than2 cm in size.17 Overall, the 10-year risk of mastectomy with-out radiotherapy was 6.3% and 3.2% with radiotherapy (p �0.001). In particular, women ages 70 to 74 and those with

KEY POINTS

� Breast radiotherapy after lumpectomy is consideredstandard for nearly all patients with invasive breast cancerand is recommended for many patients after lumpectomyfor DCIS.

� The goals of breast radiotherapy post-lumpectomy aredifferent for patients with invasive breast cancer than forpatients with DCIS.

� For patients with invasive disease, the risk of in-breastrecurrence is lowest post-lumpecomty in node-negative,small, grade 1–2, hormone sensitive breast cancer in olderpostmenopausal women who will take antiendocrinetherapy.

� For patients with DCIS, the risk of in-breast recurrence islower postlumpectomy in mammographically detected,small, low-grade hormone sensitive DCIS with negativemargins in postmenopausal women that will taketamoxifen.

� These patients with breast cancer can still have variablerisks of in-breast recurrence post-lumpectomy thatunderlies the need for better tests to predict anindividual’s risk so optimal treatment decisions can bemade.

IS POST-LUMPECTOMY RADIATION MANDATORY?


high grade breast cancer had 10-year absolute reduction inmastectomy rates of 3.8% and 6.7%, respectively.The PRIME trial of 255 women older than age 65 with neg-

ative nodes and less than 3 cm size breast cancer treated withlumpectomy and antiendocrine therapy to evaluate quality oflife in those randomized to observation compared with ra-diotherapy. More breast symptoms were noted at 15 monthsfollow-up in the irradiated group but no longer signifıcant by5 years follow-up.In summary, for most patients with invasive breast cancer

that have undergone lumpectomy, radiotherapy is needed tomaximize local control and minimize breast cancer mortal-ity. Older postmenopausal women who have node-negative,hormone receptor positive breast cancer, less than 2 cm insize and grade 1–2 have a lower risk of subsequent breast can-cer recurrence and will likely not have worse breast cancermortality with omission of radiotherapy. For woman age 70or older, competing health risks may influence outcomesmore than intrinsic risk features of the breast cancer. Addi-tional studies are warranted in younger postmenopausalwoman to further identify patients for whom the radiationtherapy might be safely omitted.

DCISFor DCIS, the primary goal of breast radiotherapy afterlumpectomy is to maximize in-breast cancer control to pre-vent or minimize occurrence of the fırst invasive breast can-cer and to avoid mastectomy. In the treatment of DCIS,breast cancer mortality has not been associated with localtreatment. Five randomized control trials have demonstratedthat breast radiotherapy signifıcantly reduces cancer recur-rence following complete excision for DCIS.18-21,23 AnEBCTCG meta-analysis provides a concise overview of ra-diotherapy effect following lumpectomy for DCIS.22 A totalof 3,729 women were eligible for analysis from the NSABPB17,EOTRC 10953, SweDCIS, and United Kingdom ANZ trials

with a median follow-up of 8.9 years. Radiotherapy morethan halved the 10-year rate of ipsilateral breast events (rateratio 0.46, standard error [SE] 0.05, 2 p � 0.00001) from28.1% with lumpectomy alone to 12.9% with the addition ofbreast radiotherapy, p � 0.00001. At 5 years after random-ization the absolute reduction in risk was 10.5% (SE 1.2%,7.6% vs 18.1%), and at 10 years it was 15.2% (SE 1.6%, 12.9%vs 28.1%). Radiotherapy resulted in a larger proportional re-duction in the rate of ipsilateral breast recurrence for womenaged more than 50 than for younger women. The propor-tional reduction did not differ signifıcantly according to anyother factor. There was no signifıcant difference in the meta-analysis for breast cancer or overall mortality between treat-ment arms. There were 50 out of 1,878 (2.7%) deaths as aresult of breast cancer for the radiotherapy groups and 44 outof 1,851 (2.3%) for observation post-lumpectomy.Within the EBCTCGmeta-analysis a “low-risk” group was

sought in which the absolute risk of ipsilateral breast eventswas so low that the addition of radiotherapy would providelittle absolute gain. There were 291 such cases of DCIS withlow-grade, less than 20mm in size and with negative surgicalmargins identifıed. Among them, the 10-year risk of an ipsi-lateral event in those allocated to lumpectomy alonewas sub-stantial at 30.1%, and the addition of radiotherapy resulted ina 10-year absolute gain of 18.0% (SE 5.5%).20Recently, the Radiation TherapyOncologyGroup (RTOG)

reported the results from its RTOG 9804 clinical trial for“Good Risk” DCIS post-lumpectomy, randomizing patientsto observation compared with breast radiotherapy23; seekingto determine radiotherapy benefıt after lumpectomy forDCIS patients who are expected to have a low risk of ipsilat-eral breast recurrence. In comparison to the other random-ized trials, RTOG 9804 enrolled all mammography screendetected lesions, 72.4% were less than 10mm in size, all werelow or intermediate nuclear grade, and adjuvant tamoxifenwas taken by 62%.After amedian follow-up of 7.2 years therehavebeen19 in-breast recurrences (42% invasive: 58%DCIS) inthe observation arm (7-year rate 6.7%) and two (50% invasive:

TABLE 1. “Low Risk” Invasive Breast Cancer Post-Lumpectomy: Randomized Controlled Trials EvaluatingRadiotherapy Benefit

Clinical Trial N F/U yrs Age � 50 yrs (%) ER/PR�(%) Tam/AI (%) Grade 1–2 (%)

In-BreastRecurrence (%)

RT No RT

NSABP B21 1009 8 80 56.5 67 67 9.31 2.82 16.5

PMH 769 5.6 100 80.5 100 68.3 0.6 7.7

ABCSG Study 8a 869 4.48 99 100 100 95 0.4 5.1

CALGB 9343 626 10.5 1003 97 100 - 2 8

GBSG-V 347 9.9 91.4 88 50 97.2 6 20

BASO II 1172 4.5 - - 25 1005 1.3 3.6

PRIME 255 5 1004 - - 94.5 0 6

Abbreviations: NSABP, National Surgical Breast and Bowel Program; PMH, Princess Margaret Hospital; ABCSG, Austrian Breast and Colorectal Cancer Study Group; CALGB, Cancer and LeukemiaGroup B; GBSG, German Breast Cancer Study Group; BASO, British Association of Surgical Oncology; PRIME, Postoperative Radiotherapy in a Minimum Risk Population.1RT � placebo, 2RT � tamoxifen. 3All were �70. 4All were � age 65. 5All were grade 1.

JULIA WHITE


50% DCIS) in the radiotherapy group (7-year rate 0.9%) for ahazard ratio of 0.11 (95 CI 0.03–0.47, p � 0.0003).23There have been numerous retrospective studies from sin-

gle institutions reporting in-breast recurrence rates of 15% to18% at 5–6 year median follow-up from excision alone forDCIS.24-26 Two prospective single arm trials sought to estab-lish a combination of size of lesion, grade, and surgical mar-gin width that might defıne a subset of patients at low risk forlocal failure without irradiation, Table 2. In the Harvardstudy of 158 women with predominant grade 1–2 DCIS andmargin of 10 mm, 13 patients developed local recurrence asthe fırst site of treatment failure 7 to 63 months after studyentry, corresponding to a 5-year rate of 12%. In the ECOGstudy, ipsilateral breast recurrence of 6.1% at 5 year and10.5% at 7 years was seen in the low-intermediate grade stra-tum, which the authors conclude may be acceptable to manypatients and physicians.There are numerous factors associated with risk of local

recurrence for a woman with DCIS who has undergonelumpectomy. In many cases it is diffıcult to weight the con-tribution of multiple risk factors to arrive at an individualestimate for in-breast recurrence. As a result, clinical toolshave been developed by using these factors to estimate of riskof local recurrence for an individual and to guide treatmentdecision making. The Van Nuys Prognostic Index was fırstdescribed in 1995 using combinations of nuclear grade andnecrosis. It now incorporates four statistically signifıcant in-dependent prognostic factors for local tumor recurrence (tu-mor sizes, margin width, pathology class that includes gradeand necrosis, and patient age) to which a grade of 1–3 is as-signed for a score of 4–12.25 DCIS cases with a score of 4–6are recommended to receive lumpectomy alone without ra-diotherapy based on their institutional experience in 320 pa-tients for which a 6% in-breast recurrence risk was seen.25Investigators from Memorial Sloan-Kettering have devel-oped a nomogram based on 1,868 consecutive patientstreated for DCIS with lumpectomy, approximately 50% ofwhom received breast radiotherapy. A combination of 10clinical, pathologic, and treatment factors associated with in-breast recurrence were studied with multivariate analysis to

devise themonogram that has been validated and is availableon the institution website.24 Gene expression has been corre-lated with DCIS prognosis post-lumpectomy and representsan important area of future investigation. ECOG investiga-tors presented a “DCIS Score” subset of the 21 Gene On-cotype DX Recurrence Score for hormone sensitive invasivecancer. The DCIS Score was able to stratify cases by risk ofipsilateral invasive recurrence on a subset of 327 cases en-rolled on the ECOG 5194.29 Similarly, loss of retinoblastoma(RB) and phosphate and tensin homolog (PTEN) suppressorgenes was strongly associated with ipsilateral invasive breastcancer recurrence in 236 patients with DCIS treated withlumpectomy alone.30In summary, breast radiotherapy following lumpectomy

for DCIS reduces the risk of in-breast recurrence by morethan half. There has not been a subset of DCIS reproduciblyidentifıed that does not receive some benefıt in reduction oflocal recurrences with radiotherapy. A lower risk of in-breastrecurrence is seen after lumpectomy alone for mammo-graphically detected, small less than 10 mm, grade 1–2 with-out necrosis, hormone receptor positive DCS, with at least 3mm surgical margins. Patients should be included in thetreatment decision-making to learn what magnitude of riskreduction is meaningful to them.

CONCLUSIONIt is becoming increasingly evident that not all invasive breastcancer or DCIS needs breast radiotherapy post-lumpectomy.The challenge remains to accurately predict an individual pa-tient’s risk of recurrence post-lumpectomy to make treat-ment recommendations. Future development of genomicand molecular discriminators will be an important step toaccurately characterize an individual woman’s personalizedprognosis for local cancer recurrence. Identifıcation of pa-tients unlikely to benefıt from radiation therapy post-lumpectomy may also serve to reduce use of mastectomy. Theindications for mastectomy should be no longer be relevant iflocal recurrence risk is suffıciently low to omit radiotherapypost-lumpectomy.



TABLE 2. Prospective Studies of Lumpectomy Alone for DCIS

Study N F/U (yrs.) Median Age Median Size (cm) Grade Margin Width

In-BreastRecurrence (%)

5-yr 7 yr

Harvard 158 3.6 51 0.9 Low-intermediate grade 10 mm 12 -

ECOG 5194 565 6.2 60 0.6 Low-intermediate grade 3 mm 6.1 10.5

105 6.7 59 0.5 High grade 15.3 18

Abbreviation: ECOG, Eastern Cooperative Oncology Group.

IS POST-LUMPECTOMY RADIATION MANDATORY?


References

1. Fisher B, Bryant J, Dignam JJ, et al. Tamoxifen, radiation therapy, orboth for prevention of ipsilateral breast tumor recurrence after lumpec-tomy in women with invasive breast cancer in of one centimeter or less.J Clin Oncol. 2002;20:4141-4149.

2. Liljegren G, Holmberg l., Bergh J, et al. 10-year results after sector re-section with or without postoperative radiotherapy for stage I breastcancer: A randomized trial. J Clin Oncol. 1999;17:2326-2333.

3. Clark RM, Whelan T, Levine M., et al. Randomized clinical trial ofbreast irradiation following lumpectomy and axillary dissection fornode-negative breast cancer: An update. J Natl Cancer Inst. 1996;88:1659-1664.

4. Forrest AP, StewartHJ, EveringtonD, et al. Randomised controlled trialof conservation therapy for breast cancer: 6-years analysis of the Scot-tish trial. Lancet. 1996;348:708-713.

5. Whelan T, Clark R, Roberts R, Levine M, Foster G, Investigators of theOntario Clinical Oncology Group. Ipsilateral breast tumor recurrencepost lumpectomy is predictive of subsequent mortality: Results from arandomized trial. Investigators of the Ontario Clinical OncologyGroup. Int J Radiat Oncol Biol Phys. 1994;30:11-16.

6. Early Breast Cancer Trialists Collaborative Group. Effects of radiother-apy and of differences in the extent of surgery for early breast cancer onlocal recurrence and 15-year survival: An overview of the randomizedtrials. Lancet. 2005;366:2087-2106.

7. Early Breast Cancer Trialists Collaborative Group. Effect of radiother-apy after breast-conserving surgery on 10-year recurrence and 15-yearbreast cancer death: Meta-analysis of individual patient data for 10 801women in 17 randomised trials. Lancet. 2011;378:1707-1716.

8. Fyles AW,McCreadyDR,Manchul LA, et al. Tamoxifenwith orwithoutbreast irradiation in women 50 years of age or older withearly breastcancer. N Engl J Med. 2004;351:963-970.

9. Ptter R, Gnant M, Kwasny W, et al. Lumpectomy plus tamoxifen ora-nastrozole with or without whole breast irradiation in women with fa-vorable early breast cancer. Int J Radiat Oncol Biol Phys. 2007;68:334-340.

10. Hughes KS, Schnaper LA, Berry D, et al. Lumpectomy plustamoxifenwith or without irradiation in women 70 years of age or older with earlybreast cancer. N Engl J Med. 2004;351:971-977.

11. Winzer KJ, Sauerbrei W, Braun M, et al. Radiation therapy and tamox-ifen after breast-conserving surgery; updated results of a 2x2 ran-domised clinical trial in patients with low risk of recurrence. Eur JCancer. 2010;46:95-101.

12. Blamey RW, Chetty U, Bates T, et al. Radiotherapy and/or tamoxifenafter conserving surgery for breast cancers of excellent prognosis: BASOII TRIAL. Eur J Cancer Suppl. 2007;5:O-11

13. Williams LJ, Kunkler IH, Prescott RJ, et al. A randomised controlledtrial of postoperative radiotherapy following breast-conserving surgeryin a minimum-risk older population. The PRIME trial. Health TechnolAssess. 2011;15(12):1-59.

14. Smith BD, Gross CP, Smith GL, et al. Effectiveness of radiation therapyfor older women with early breast cancer. J Natl Cancer Inst. 2006;98:681-690.

15. Albert JM. Pan IW, Shih YCT, et al. Effectivenss of radiation for preven-tion of mastectomy in older breast cancer patients treated with conser-vative surgery. Cancer. 2012;118:4642-4651.

16. Wapnir IL, Dignam JJ, Fisher B, et al. Long-term outcomes of invasiveipsilateral breast tumor recurrences after lumpectomy in NSABP B-17and B-24 randomized clinical trials for DCIS. J Natl Cancer Inst. 2011;103:478-488.

17. Bijker N, Meijnen P, Peterse JL, et al. Breast-conserving treatment withorwithout radiotherapy in ductal carcinoma-in-situ: Ten-year results ofEuropeanOrganisation for Research andTreatment of Cancer random-ized phase III trial 10853-a study by the EORTC Breast Cancer Cooper-ative Group and EORTC Radiotherapy Group. J Clin Oncol. 2006;24:3381-3387.

18. Cuzick J, Sestak I, Pinder SE, et al. Effect of tamoxifen and radiotherapyin women with locally excised ductal carcinoma in situ: Long-term re-sults from the UK/ANZ DCIS trial. Lancet Oncol. 2011;12:21-29.

19. Holmberg L, GarmoH,Granstrand B, et al. Absolute risk reductions forlocal recurrence after postoperative radiotherapy after sector resectionfor ductal carcinoma in situ of the breast. J Clin Oncol. 2008;26:1247-1252.

20. Correa C, McGale P, Taylor C, et al. Overview of the randomized trialsof radiotherapy in ductal carcinoma in situ of the breast. J Natl CancerInst Monographs. 2010;2010:162-177.

21. McCormick BW,K;Hudis, C; Kuerer, et al. Low-risk Breast Ductal Car-cinoma In Situ (DCIS): Results From the Radiation Therapy OncologyGroup 9804 Phase 3 Trial. In:American Society for Radiation Oncology.2012;Boston, MA; 2012. p. S5.

22. Rudloff U, Jacks LM, Goldberg JI, et al. Nomogram for predicting therisk of local recurrence after breast-conserving surgery for ductal carci-noma in situ. J Clin Oncol. 2010;28:3762-3769.

23. SilversteinMJ, Lagios MD. Choosing treatment for patients with ductalcarcinoma in situ: Fine tuning the University of Southern California/Van Nuys Prognostic Index. J Natl Cancer Inst Monographs. 2010;2010:193-196.

24. Fong J, Kurniawan ED, Rose AK, et al. Outcomes of screening-detectedductal carcinoma in situ treated with wide excision alone.Ann Surg On-col. 2011;18:3778-3778.

25. Wong JS, Kaelin CM, Troyan SL, et al. Prospective study of wide exci-sion alone for ductal carcinoma in situ of the breast. J Clin Oncol. 2006;24:1031-1036.

26. Hughes LL, Wang M, Page DL, et al. Local excision alone without irra-diation for ductal carcinoma in situ of the breast: A trial of the EasternCooperative Oncology Group. J Clin Oncol. 2009;27:5319-5324.

27. Solin LJ, Gray R, Baehner FL, et al. A quantitative multigene RT-PCRassay for predicting recurrence risk after surgical excision alonewithoutirradiation for ductal carcinoma in situ (DCIS): a prospective validationstudy of the DCIS Score from ECOG E5194. In: 34th Annual San Anto-nio Breast Cancer Symposium. San Antonio, TX; 2011.

28. Knudsen ES, Pajak TF, QeenanM, et al. Retinoblastoma and phosphateand tensin homolog tumor suppressors: Impact on ductal carcinoma insitu progression. J Natl Cancer Inst. 2012;104:1825-1836.

JULIA WHITE


BREAST CANCER

Obesity and Inflammation: TheDangerous Duo in Breast Cancer

CHAIRJennifer Ligibel, MDDana-Farber Cancer InstituteBoston, MA

SPEAKERSHoward Strickler, MD, MPHAlbert Einstein College of MedicineBronx, NY

Andrew Dannenberg, MDWeill Cornell Medical CollegeNew York, NY

Obesity and Inflammation: New Insights into Breast CancerDevelopment and ProgressionNeil M. Iyengar, MD, Clifford A. Hudis, MD, and Andrew J. Dannenberg, MD

OVERVIEW

The importance of inflammation in promoting carcinogenesis and tumor progression is well recognized. Chronic inflammation causedby a variety of infectious agents can lead to the development of several common malignancies. Similarly, inflammatory bowel diseaseis a well-known risk factor for colorectal cancer. Much less is known about the link between inflammation and the development of breastcancer. Recent data suggest that obesity causes both in-breast and systemic inflammation that contribute to the development andprogression of breast cancer. This observation has potentially important implications in terms of prevention and treatment of breastcancer, especially given the rising worldwide overweight and obesity rates. Inflamed white adipose tissue (WAT) within the breast isassociated with elevated levels of proinflammatory mediators, enhanced expression of aromatase (the rate-limiting enzyme forestrogen biosynthesis), and increased estrogen receptor-� (ER-�)-dependent gene expression. Systemic consequences of obesityincluding altered adipokine levels, elevated circulating estrogen levels, and insulin resistance are also believed to play a role in thepathogenesis of breast cancer. Collectively, these findings suggest a significant role for inflammation in the pathogenesis of breastcancer in obese and overweight patients.

Among its wide range of clinical consequences, obesity isa well-known risk factor for the development of several

common epithelial malignancies.1 In postmenopausalwomen, the risk of developing breast cancers that express theestrogen and progesterone receptors (ER and PR), is signifı-cantly elevated for those who are obese or overweight.2 Fur-thermore, obese and overweight patients, once diagnosed,suffer fromworse disease-related outcomes than their leanercounterparts, regardless of breast cancer subtype.3 Aftermenopause, estrogen is mostly derived peripherally from thenoncyclical conversion of androgen precursors within adi-pose tissue. The rate-limiting step in this conversion is cata-lyzed by the cytochrome P450 enzyme, aromatase, which isencoded by the CYP19 gene. Circulating estrogens, such asestradiol, are known to stimulate the proliferation of breastepithelial cells and potentially exert a mutagenic effect.Higher levels of circulating estradiol as a result of increasedadiposity and aromatase expression are thought to contrib-ute, in part, to the greater risk of ER/PR-positive breast can-cer in obese postmenopausal women.2 However, comparedto the premenopausal state, circulating estradiol levels aresignifıcantly lower as the ovaries no longer produce substan-tial amounts of estrogen. Nevertheless, the incidence of ER-positive disease rises with age, approaching nearly 85% ofbreast cancers diagnosed in women during their ninth de-cade of life.4 This seemingly paradoxical association between

the increasing incidence of hormone-sensitive tumors in ag-ing women whose circulating estradiol levels have declinedafter menopause highlights an important scientifıc question:Why are hormone-sensitive breast cancers more prevalentafter circulating levels of tumor-stimulating estrogens havesignifıcantly and naturally declined? One possibility is thatthe microenvironment in which the neoplasm arises andgrows plays an important role and that it compensates for thedecrease in circulating estrogens. Indeed, locally producedestrogens and, possibly, ligand-independent activation ofER-�have been associatedwith carcinogenesis.5 Local effectsmay also be important for understanding the link betweenobesity and increased risk of hormone receptor-positivebreast cancer in obese postmenopausal women.Additional evidence suggests that there are likely to be sev-

eral estrogen-independent mechanisms involved in the linkbetween obesity and breast carcinogenesis (Fig. 1). In addi-tion to the elevated risk of hormonally sensitive breast tu-mors, obesity has also been associated with an increased riskof ER-negative breast cancers in some studies.6 This observa-tion, in particular, points to the involvement of estrogen-independent pathways (although thesemay also be operativein hormone receptor-positive cancers). Obesity is associatedwith chronic, systemic inflammation characterized by ele-vated levels of circulating proinflammatory mediatorsknown to promote tumorigenesis and growth.1,7,8 Both the

From the Memorial Sloan-Kettering Cancer Center, New York, NY; Weill Cornell Medical College, New York, NY.


Corresponding author: Andrew J. Dannenberg, MD, Department of Medicine and Weill Cornell Cancer Center, 525 East 68th St, Room F-206, New York, NY 10065; email: [email protected].


IYENGAR, HUDIS, AND DANNENBERG


systemic and local consequences of chronic adipose inflam-mation thus provide key potential mechanistic links betweenobesity and breast cancer. In addition to inflammation, otherobesity-related effects promote cell proliferation and sur-vival. These include hyperinsulinemia and increased insulin-like growth factor-1 (IGF-1) signaling, aswell as altered levelsof the adipokines, adiponectin, and leptin. This review ad-dresses the complex biologic interactions between obesityand breast cancer, with a particular focus on inflammation asa key mediator.

CONNECTING OBESITY AND BREAST CANCER VIAINFLAMMATIONTumor MicroenvironmentThe links between inflammation and cancer have long beendescribed. In the mid-nineteenth century, Virchow fırst hy-pothesized that tissue injury and the ensuing inflammationpromote enhanced cellular proliferation, thereby predis-posing cells to neoplastic transformation.9 Subsequent obser-vations that many cancers arise at sites of chronic inflamma-tion, coupled with epidemiologic data demonstratingincreased cancer rates in patients with underlying inflamma-tory conditions, have led to signifıcant efforts to better un-derstand the complex role of inflammation. Adding urgencyto this ongoing investigation are reports linking infectionand the inflammatory response to 15% to 20% of all cancer-related deaths worldwide.9 A critical observation is that thetumor microenvironment closely resembles that of a healingwound, including the influx of immune cells with resultantproduction of proinflammatory mediators as well as tissueremodeling and angiogenesis.10 Furthermore, malignantcells may co-opt the inflammatorymechanism, leading to in-creased growth, invasion, and metastasis.11 Together, the

epidemiologic and histologic observations strongly suggestthat inflammation plays a key role in tumor biology, and isnot simply a result of the development of cancer.Although many cancers including colon, lung, bladder,

esophagus, liver, and others, are known to have infectiousand/or inflammation-related etiologies, this link has been farless clear for breast cancer. Recently, inflammation, charac-terized by the hallmarks of wound healing, has been demon-strated to occur in breast tumors.12 Emerging evidence pointsto a causal relationship between obesity, chronic in-breast in-flammation, and tumorigenesis. Increased body mass indexhas been associated with adipocyte hypertrophy in thebreast.13 Moreover, a positive correlation has been observedbetween increased adipocyte size and cell death in the humanbreast. Adipocytes can release a variety of proinflammatorycytokines and adipokines, which are likely to contribute tothe recruitment and activation of immune cells includingmac-rophages.14 Dying adipocytes also release free fatty acids (FFA),which can stimulate the production of proinflammatory cy-tokines via toll-like receptor-4 (TLR-4) and NF�B signaling.

Adipocyte InteractionsA growing understanding of the complex interactions be-tween adipocytes and immune cells within theWAT stromalvascular fraction is shedding new light on the role of adiposeinflammation in tumor development and progression. Theactivatedmacrophage is a keymediator of adipose inflamma-tion, and the adipocyte-macrophage interaction is emergingas a central theme. Adipocyte death leads to myeloid cell re-cruitment in a characteristic pattern whereby macrophagesform a crown surrounding the dead adipocyte. This forma-tion is histologically apparent as crown-like structures (CLS),which have been observed in subcutaneous and visceral fat inassociation with the metabolic syndrome.15 More recently,these inflammatory lesions were discovered to occur in themammary gland of obese mice and the WAT of the humanbreast, termed CLS-B.13,16 Notably, CLS-B are detected moreeffıciently using immunohistochemical staining for CD68, amacrophage marker, rather than with standard hematoxylinand eosin (H&E) staining (Fig. 2).13 For this reason as well asthe usual focus on tumor, rather than benign tissue, CLS-Band their associated biology may not have been previouslyappreciated.As in the preclinicalmousemodels, the presenceof CLS-B in women was associated with activation of NF�Band increased levels of TNF�, IL-1�, IL-6, and COX-2- de-rived PGE2. Relating to the epidemiological association ofobesity and postmenopausal hormone receptor–positivebreast cancer, a critical consequence of CLS-B and the asso-ciated elevation in tissue levels of proinflammatory cytokinesis increased transcription of the CYP19 gene encoding aro-matase.19 Indeed, several of the proinflammatory mediatorsassociated with CLS-B, including TNF�, IL-1�, IL-6, andPGE2, are known to upregulate aromatase expression via spe-cifıc promoters that give rise to unique mRNA species foundin breast tissue.19-21 Increased levels and activity of aromataselead to enhanced estrogen biosynthesis and upregulation ofPR, an ER target gene. Additionally, increased circulating

KEY POINTS

� In obese women, low-grade or “smoldering” inflammationoccurs in multiple white adipose tissue depots including thebreast.

� Overweight and obesity are characterized by adipocytehypertrophy and adipocyte death in association withrecruitment of immune cells into white adipose tissue.

� Infiltrating macrophages surround dead adipocytes inbreast white adipose tissue to form inflammatory crown-like structures (CLS-B) that are associated with NF�Bactivation and increased levels of several proinflammatorymediators that are known inducers of aromatase andthereby estrogen biosynthesis.

� Breast inflammation, quantified by the CLS-B index,correlates with increased aromatase activity, thusestablishing an obesity 3 inflammation 3 aromatasesignaling axis.

� Excess adiposity is associated with a state of metabolicdysregulation and systemic inflammation, which have beenlinked to breast carcinogenesis.

OBESITY, INFLAMMATION, AND BREAST CANCER


levels of proinflammatory mediators in obese patients withbreast cancer correlatewith poor prognosis.17,18 For example,elevated levels of IL-6 in serum have been associated withworse survival in patients with metastatic hormone-refractory breast cancer.17Given the signifıcant biologic consequences of these in-

flammatory pathways and the need to further explore the his-tologic structures associated with their activation, a CLS-Bindex was developed to quantify the severity of breast WAT

inflammation. This index is defıned as the number of slideswith histologic evidence of CLS-B compared to the numberof slides examined, reported on a scale ranging from 0 to1.0. Consistent with the hypothesis that the obesity3inflammation3aromatase axis is operative, the severity ofbreast inflammation as defıned by the CLS-B index corre-lated with aromatase activity.13 Furthermore, the presence ofCLS-B was not merely a surrogate for increasing body massindex. In support of this point, CLS-B was found in approx-imately 75% of obese women. Inflammatory foci can also befound in a minority of lean women. These fındings are con-sistent with the observation that not all obese individuals suf-fer from the metabolic syndrome and that other factors mayalso induce localized inflammation inWAT. Overall, the dis-covery of CLS-B implicates inflammation specifıcally, ratherthan obesity alone, as a key driver of aromatase activity inthe breast. Taken together, these fındings demonstrate anobesity3inflammation3aromatase signaling axis, charac-terized by a complex system of paracrine interactionsbetween macrophages and other cells, for example, preadi-pocytes, in human breast WAT (Fig. 3). This axis placesinflammation at the center of ER-positive breast cancer patho-genesis for many patients and provides an important mech-anistic and targetable link between obesity and breast cancer.

FIG 1. Pathways linking obesity and breast cancer. Reprinted with permission from Sinicrope and Dannenberg.30© 2010 American Society of Clinical Oncology. All rights reserved.Abbreviations: IGF-1, insulin-like growth factor-1; IL-6, interleukin-6; IL-1�, interleukin-1�; SHBG, sex hormone binding globulin; TNF�, tumor necrosis factor alpha; VEGF, vascular endothelialgrowth factor.

FIG 2. Crown-like structures of the breast (CLS-B,arrow). A. Hematoxylin and eosin stain. B. Anti-CD68immunostain identifies macrophages. Reprinted withpermission from Morris et al.13



Metabolic Dysfunction and InflammationIn addition to activation of estrogen signaling pathwaysvia inflammation-mediated upregulation of aromatase,obesity promotes breast cancer through the systemic effectsof dysregulated metabolism. A primary function of adiposetissue is to store energy in the form of triglycerides withinthe WAT adipocytes. Hyperadiposity is characterized by astate of energy imbalance, which plays an important rolein immune activation. In addition to cell wall stretch andsubsequent release of proinflammatory cytokines, adipocytehypertrophy is associated with increased lipolysis, resultingin the release of FFA. FFA are believed to stimulate in-flammatory pathways in adipose tissue, but can also be de-posited in various organs contributing to atherosclerosis,hypertension, insulin resistance, and glucose intolerance.Indeed, excess adiposity is associated with higher circulatingtriglyceride levels, reflecting the state of energy imbalancethat occurs in many obese/overweight patients. As noted,FFA stimulate multiple inflammatory signaling pathways,ultimately leading to activation of the transcription factor,NF�B, which plays a key role in the immune responseand adipose inflammation.22 Activation of NF�B can bemediated by TLR-4, a member of the TLR family of patternrecognition molecules that are centrally involved in theimmune response.23 Lipopolysaccharide (LPS), a compo-nent of the cell wall of Gram-negative bacteria, is a proto-typic TLR-4 agonist and stimulates the innate immune re-sponse. In a similar manner, TLR-4 is thought to orchestrateinflammation in response to excess endogenous lipids.23,24FFA signaling through TLR-4 can lead to macrophage acti-vation, resulting in elevated levels of TNF�, IL-6, and otherproinflammatory mediators as detailed above.24 These cyto-kines stimulate further lipolysis and release of FFA, thusestablishing a reinforced feedback cycle of sustained inflam-mation.

CONNECTING OBESITY AND BREAST CANCER VIAENDOCRINE DYSFUNCTIONEndocrine-mediated consequences of obesity also contributeto breast carcinogenesis and tumor progression. These in-clude elevated levels of insulin and bioavailable insulin-likegrowth factor-1 (IGF-1), partly related to alterations in levelsof the adipokines, adiponectin, and leptin (Fig. 1). Both ofthese adipokines have angiogenic properties. Elevated leptinlevels, as occur in obesity, stimulate breast tumor cell prolif-eration through several signal transduction pathways and byaltering cell-cycle checkpoints via upregulation of cyclin D1and cdk2.25,26 In contrast to leptin, adiponectin levels are di-minished in obesity. Although the underlying mechanismsare not completely understood, decreased adiponectin levelsare associated with hyperinsulinemia and increased breastcancer risk. These observations may be partly explained bythe effects of adiponectin on the PTEN/PI3K/mTOR,MAPK, and AMPK pathways, leading to enhanced cell pro-liferation and inhibition of apoptosis.1,25Insulin resistance, characterized by elevated plasma insulin

levels, occurs in association with obesity-related inflamma-tion and is likely to play an important role in breast cancerprogression.27 Both insulin and IGF-1 promote cell prolifer-ation and inhibition of apoptosis by stimulating the PI3K/Akt and Ras/Raf/MAPK systems.28 Additionally, IGF-1 andinsulin, to a lesser degree, interact with estrogen signalingpathways to promote hormone-sensitive breast cancers.IGF-1 has been shown to stimulate aromatase activity.29 In-sulin, on the other hand, acts systemically by inhibiting he-patic synthesis of sex-hormone-binding globulin (SHBG),which binds and transports the hormones testosterone, dihy-drotestosterone, and estradiol, in a biologically inactivestate.1 Increased free estradiol, as a result of diminishedSHBG levels, is then available to stimulate growth of ER-expressing breast tumors. Together, these fındings suggestthat obesity-associated insulin resistance has multiple local

FIG 3. Paracrine interactions involving activated macrophages establish an obesity3inflammation3aromataseaxis in the breast tissue of obese/overweight women.Abbreviations: FFA, free fatty acids; IL-1�, interleukin-1�; PGE2, prostaglandin E2; TNF�, tumor necrosis factor alpha; TF, transcription factors.



and systemic consequences that help to explain the link be-tween obesity and breast cancer.

CONCLUSIONRates of obesity and overweight are predicted to climb in thecoming years and as such, we are likely to see a proportionalincrease in neoplastic disease and obesity-related mortality.Understanding the biologic underpinnings of this relation-ship is critical to the development of preventative and ad-junctive therapies that could disrupt the obesity-cancer linkeven if obesity itself remains. To this end, it is becoming in-creasingly apparent that activation of multiple inflammatorysignaling pathways provides a critical link between obesityand cancer. Recent observations reporting the interactionsbetween adipocytes and immune cells leading to a proinflam-matory local milieu are providing new insights into thewound-like tumormicroenvironment described by Virchowmore than a century ago. Specifıcally, the discovery of an

obesity3inflammation3aromatase axis highlights inflam-mation as a central mediator of breast carcinogenesis formanypatients. Identifying patients inwhom this axis is activecould uncover a specifıc population that may benefıt fromanti-inflammatory interventions. The potential role of be-havioral interventions, nutritional changes, and pharmaco-logic strategies to attenuate obesity-related inflammationneed to be explored. BecauseWAT inflammation is found inmost but not all overweight and obese women, the develop-ment of noninvasive methods to detect chronic in-breast in-flammation to identify the at-risk population is a keyresearch objective.

ACKNOWLEDGMENTSThis work was supported by NIH R01CA154481, the BreastCancer Research Foundation, and grant UL1TR000457 ofthe Clinical and Translational Science Center at Weill Cor-nell Medical College.


Relationships are considered self-held and compensated unless otherwise noted. Relationships marked “L” indicate leadership positions. Relationships marked “I” are those held by animmediate family member; those marked “B” are held by the author and an immediate family member. Relationships marked “U” are uncompensated. Authors marked with an asterisk (*) areparticipants in ASCO’s Disclosure Management System Pilot; their disclosure is not limited to subject matter under consideration in this article and includes payments to themselves, animmediate family member (I), and/or their institutions (Inst). For information on the pilot program, or to provide feedback, please visit coipilot.asco.org.

Employment or Leadership Position: None. Consultant or Advisory Role: None. Stock Ownership: None. Honoraria: None. Research Funding:*Clifford A. Hudis, Merck. Expert Testimony: None. Other Remuneration: *Clifford A. Hudis, Estee Lauder Company (I).

References

1. van Kruijsdijk RC, van derWall E, Visseren FL. Obesity and cancer: therole of dysfunctional adipose tissue.Cancer Epidemiol Biomarkers Prev.2009;18:2569-2578.

2. ClearyMP, GrossmannME.Minireview: Obesity and breast cancer: theestrogen connection. Endocrinology. 2009;150:2537-2542.

3. Calle EE, Rodriguez C,Walker-ThurmondK, et al. Overweight, obesity,and mortality from cancer in a prospectively studied cohort of U.S.adults. N Engl J Med. 2003;348:1625-1638.

4. Li CI, Daling JR,MaloneKE. Incidence of invasive breast cancer by hor-mone receptor status from 1992 to 1998. J Clin Oncol. 2003;21:28-34.

5. Lorincz AM, Sukumar S. Molecular links between obesity and breastcancer. Endocr Relat Cancer. 2006;13:279-292.

6. Daling JR, Malone KE, Doody DR, et al. Relation of body mass index totumor markers and survival among young women with invasive ductalbreast carcinoma. Cancer. 2001;92:720-729.

7. Pierce BL, Ballard-Barbash R, Bernstein L, et al. Elevated biomarkers ofinflammation are associated with reduced survival among breast cancerpatients. J Clin Oncol. 2009;27:3437-3444.

8. Olefsky JM, Glass CK. Macrophages, inflammation, and insulin resis-tance. Annu Rev Physiol. 2010;72:219-246.

9. Balkwill F, Mantovani A. Inflammation and cancer: Back to Virchow?Lancet. 2001;357:539-545.

10. Mantovani A, Allavena P, Sica A, et al. Cancer-related inflammation.Nature. 2008;454:436-444.

11. Coussens LM, Werb Z. Inflammation and cancer. Nature. 2002;420:860-867.

12. Balkwill F, Charles KA,MantovaniA. Smoldering and polarized inflam-mation in the initiation and promotion of malignant disease. CancerCell. 2005;7:211-217.

13. Morris PG,Hudis CA, Giri D, et al. Inflammation and increased aroma-tase expression occur in the breast tissue of obese women with breastcancer. Cancer Prev Res (Phila). 2011;4:1021-1029.

14. Monteiro R, Azevedo I. Chronic inflammation in obesity and the met-abolic syndrome.Mediators Inflamm. 2010;2010.

15. Cinti S, Mitchell G, Barbatelli G, et al. Adipocyte death defınes macro-phage localization and function in adipose tissue of obese mice and hu-mans. J Lipid Res. 2005;46:2347-2355.

16. Subbaramaiah K, Howe LR, Bhardwaj P, et al. Obesity is associated withinflammation and elevated aromatase expression in the mouse mam-mary gland. Cancer Prev Res (Phila). 2011;4:329-346.

17. Bachelot T, Ray-Coquard I,Menetrier-CauxC, et al. Prognostic value ofserum levels of interleukin 6 and of serum and plasma levels of vascularendothelial growth factor in hormone-refractory metastatic breast can-cer patients. Br J Cancer. 2003;88:1721-1726.

18. Dandona P, Weinstock R, Thusu K, et al. Tumor necrosis factor-alphain sera of obese patients: fall with weight loss. J Clin Endocrinol Metab.1998;83:2907-2910.

19. Subbaramaiah K, Morris PG, Zhou XK, et al. Increased levels of COX-2



and prostaglandin E2 contribute to elevated aromatase expression ininflamed breast tissue of obese women. Cancer Discov. 2012;2:356-365.

20. Zhao Y, Agarwal VR, Mendelson CR, et al. Estrogen biosynthesis prox-imal to a breast tumor is stimulated by PGE2 via cyclic AMP, leading toactivation of promoter II of the CYP19 (aromatase) gene. Endocrinol-ogy. 1996;137:5739-5742.

21. Irahara N, Miyoshi Y, Taguchi T, et al. Quantitative analysis of aro-matase mRNA expression derived from various promoters (I.4, I.3,PII and I.7) and its association with expression of TNF-alpha, IL-6and COX-2 mRNAs in human breast cancer. Int J Cancer. 2006;118:1915-1921.

22. Hotamisligil GS. Endoplasmic reticulum stress and the inflammatorybasis of metabolic disease. Cell. 2010;140:900-917.

23. Shi H, Kokoeva MV, Inouye K, et al. TLR4 links innate immunityand fatty acid-induced insulin resistance. J Clin Invest. 2006;116:3015-3025.

24. Nguyen MT, Favelyukis S, Nguyen AK, et al. A subpopulation of mac-rophages infıltrates hypertrophic adipose tissue and is activated by free

fatty acids via Toll-like receptors 2 and 4 and JNK-dependent pathways.J Biol Chem. 2007;282:35279-35292.

25. JardeT, Perrier S,VassonMP, et al.Molecularmechanisms of leptin andadiponectin in breast cancer. Eur J Cancer. 2011;47:33-43.

26. Okumura M, Yamamoto M, Sakuma H, et al. Leptin and high glucosestimulate cell proliferation in MCF-7 human breast cancer cells: recip-rocal involvement of PKC-alpha and PPAR expression. Biochim Bio-phys Acta. 2002;1592:107-116.

27. Goodwin PJ, Ennis M, Pritchard KI, et al. Fasting insulin and outcomein early-stage breast cancer: results of a prospective cohort study. J ClinOncol. 2002;20:42-51.

28. Pollak MN, Schernhammer ES, Hankinson SE. Insulin-like growth fac-tors and neoplasia. Nat Rev Cancer. 2004;4:505-518.

29. Su B, Wong C, Hong Y, et al. Growth factor signaling enhances aroma-tase activity of breast cancer cells via post-transcriptional mechanisms.J Steroid Biochem Mol Biol. 2011;123:101-108.

30. Sinicrope FA, Dannenberg AJ. Obesity and breast cancer prognosis:weight of the evidence. J Clin Oncol. 2011;29:4-7.



Obesity and Its Impact on Breast Cancer: Tumor Incidence,Recurrence, Survival, and Possible InterventionsJennifer A. Ligibel, MD, and Howard D. Strickler, MD, MPH

OVERVIEW

A positive association between obesity and the risk of incident postmenopausal breast cancer has been consistently observed inepidemiologic studies. Although most studies of premenopausal women have not found a similar relationship between breast cancerand obesity, the prognosis for both pre- and postmenopausal breast cancer is substantially worse among obese than normal-weightindividuals. Increasing evidence suggests that these associations may be mechanistically related to sex hormones, insulin, and certainadipokines. Insulin, for example, has important mitogenic/antiapoptotic activity in addition to its metabolic effects, and many breasttumors express high levels of the insulin receptor (IR)-A isoform. Further, the use of metformin, a diabetes medication that reducesinsulin levels, has been epidemiologically associated with reduced breast cancer risk among patients with diabetes, and a recentobservational study found a higher rate of pathologic complete responses among patients with diabetes and breast cancer who wereusing metformin. Formal clinical trials of metformin as adjuvant breast cancer therapy have been initiated and are ongoing. Similarly,the effect of lifestyle changes on breast cancer outcomes is actively being investigated. Several lifestyle intervention studies havedemonstrated that weight loss, increased physical activity, and dietary changes are feasible in breast cancer populations, and thatindividuals who make lifestyle changes after breast cancer diagnosis experience several physical and psychologic benefits. In thisarticle, the authors review the evidence linking obesity with breast cancer risk and outcomes and provide an overview of lifestyleintervention studies in patients with breast cancer.

The prevalence of obesity among U.S. women has morethan doubled since the late 1970s, with 34% of women

above 20 years of age currently estimated to be obese, andan additional 27% considered overweight (Fig. 1).1,2 Al-though the prevalence of obese and overweight womenhas, in recent years, stabilized at this high level, certain riskgroups have especially high prevalence of excess adiposity,including women of black or Hispanic race/ethnicity, andthose of lower social economic status.1,2 Obesity is a well-established risk factor for several common conditions in-cluding diabetes, cardiovascular disease, stroke, renaldisease, nonalcoholic fatty liver disease/hepatic fıbrosis, anddisability. It is also signifıcantly associated with overall mor-tality.3More recently, obesity has been associated with cancer

risk and mortality.3,4 Although data suggesting this asso-ciation goes back to the 1990s and earlier, it is only in thepast few years that the evidence has reached a point thatthe obesity-cancer relationship can be said to be well-established and broadly accepted. This change in perspectivereflects not only the growing number of large, well-conducted prospective investigations regarding obesityand cancer, but also the growing number of laboratory and

molecular epidemiologic studies that have helped demon-strate biologic plausibility. Table 1 shows a list of the cancersreported to have a signifıcant positive association with obe-sity in women.4 In addition to breast cancer, these tumorsinclude endometrial, gallbladder, esophageal adenocarci-noma, renal, leukemia, thyroid, pancreas, multiplemyeloma,colon, non-Hodgkin lymphoma, rectal, and possibly ovariancancers.3,4The review that follows provides an overview of the

evidence relating obesity with breast cancer by subtype,including the associations of obesity with cancer incidence,recurrence, and survival. The laboratory and clinical/epidemiologic data that implicate specifıc signaling pathwaysin the obesity–breast cancer relationship are also discussed,and the authors review both published and ongoing inter-ventional studies being conducted to target these path-ways. Although there are currently no randomized trialstesting the effect of purposeful weight loss after diagnosison breast cancer prognosis, dozens of smaller-scale lifestyleintervention studies have been conducted in patients withbreast cancer. This review provides an overview of thegrowing literature on lifestyle intervention studies in breastcancer.

From the Department of Adult Oncology, Dana-Farber Cancer Institute, Boston, MA; the Department of Epidemiology & Population Health, Albert Einstein Cancer Center, Albert Einstein College ofMedicine, Bronx, NY.


Corresponding author: Jennifer A. Ligibel, MD, Dana-Farber Cancer Institute, 450 Brookline Ave, Y1234, Boston, MA 02215; email: [email protected].


LIGIBEL AND STRICKLER


ASSOCIATIONS OF OBESITY WITH INCIDENTBREAST CANCER BY SUBTYPEPostmenopausal Compared with PremenopausalThere is a substantial body of research showing that post-menopausal women who are overweight or obese have in-creased risk of breast cancer compared with normal weightindividuals. Although the effect of obesity estimated by re-centmeta-analyses ismodest, a relative risk of approximately1.15,4,5 these types of studies often lack pertinent data to con-trol for relevant cofactors or conduct subset analyses. As an

important example, a prospective study in the Women’sHealth Initiative found that excluding the 53% of womenwho were using hormone therapy (HT) increased the associ-ation of obesity and postmenopausal breast cancer from ahazard ratio [HR] of 1.13 (95% confıdence interval [CI]:0.83–1.55) to a HR of 1.91 (95% CI: 1.11–3.27).6 Similar re-sults were reported in the Nurses’ Health Study7 and in othercohorts.8 The effect of HT use likely reflects the strong fırst-pass effect of oral estrogens on the liver, which alters hepaticprotein production and processing, including proteins re-lated to glucose metabolism such as insulin and other hor-mones6; a frequent consideration in interpreting the obesityand breast cancer literature.In contrast, studies of obesity and premenopausal breast

cancer have had conflicting fındings. Most studies have re-ported either an inverse association4 or no association be-tween obesity and cancer risk among menstruating women,5although a few studies found a positive association.4 Hypoth-eses to explain the possible protective effect of obesity on pre-menopausal breast cancer risk have largely focused on therelation of obesitywith irregular or longmenstrual cycles andanovulatory infertility, as having fewer ovulations is thoughtto correlate with reduced estrogen exposure.9

Estrogen Receptor (ER) and Progesterone Receptor(PR) ExpressionMultiple studies of postmenopausal women have shownstrong associations between obesity and the risk of ER�/PR� breast cancer but not other breast cancer subtypes.10,11For example, a meta-analysis of nine cohorts and 22 case-control studies found that each fıve unit change in BMI wasassociated with a 33% increased risk of ER�/PR� post-menopausal breast cancer and a 10% decreased risk inER�/PR� premenopausal breast cancer, whereas no asso-ciations were observed for ER-/PR- or ER�/PR-.11 Al-though this study could not address use of HT in postmeno-pausal women, a large case-control study found a similarpositive association of obesity with all ER�/PR� postmeno-pausal breast cancer (OR � 1.3; 95% CI: 1.0–1.7) that was

KEY POINTS

� Obesity is associated with the risk of incident breastcancer in postmenopausal but not menstruating women andwith risk of cancer recurrence and mortality in patientswith early-stage breast cancer, regardless of menopausalstatus.

� Sex hormones, insulin, and certain adipokines may have amechanistic role in the development of ER�postmenopausal breast cancer.

� Metformin, a diabetes medicine that reduces insulin levels,is being studied as an adjuvant breast cancer therapy inclinical trials and has been shown in observational studiesto be associated with reduced breast cancer risk andbetter prognosis among patients with diabetes.

� There are no data from randomized trials testing the effectof purposeful weight loss on prognosis in overweight orobese women with early breast cancer.

� Dozens of interventional studies have shown that weightloss, increased physical activity, and improvements indietary quality are feasible in breast cancer survivors, andthat individuals who make these changes feel better,improve cardiorespiratory fitness, and experience favorablechanges in biomarkers linked to breast cancer risk andoutcomes.

FIG 1. Prevalence of obesity among women in theU.S. by (a) calendar time (1971 to 2004), and(b) race/ethnicity (average obesity prevalence during1999 to 2004). Overweight was defined as a bodymass index (BMI) of 25 to < 30 kg/m2, obese> 30 kg/m2, and extreme obesity as > 40 kg/m2.(Adapted from Ogden and colleagues.2)

TABLE 1. Cancers Associated with Obesity Accordingto Meta-analysis Studies

Breast

Colon

Endometrial

Esophageal adenocarcinoma

Gallbladder

Leukemia

Multiple myeloma

Non-Hodgkin lymphoma

Pancreas

Renal

Thyroid*Adapted from Renehan et al, 20084

OBESITY AND BREAST CANCER


nearly two-fold higher for obesity when HT users were ex-cluded (OR � 2.3; 95% CI:1.3–3.8).10 A meta-analysis thatspecifıcally addressed “triple negative tumors” (i.e., ER-/PR-/HER2-) found positive associations with obesity but onlyamong premenopausal women,12 although that study couldnot account for HT use in postmenopausal women.

POSSIBLE MECHANISMS UNDERLYING THEOBESITY–BREAST CANCER ASSOCIATIONMost hypotheses regarding the biologic pathways that mayunderlie the relation of obesity with breast cancer focus onthree areas: sex hormones, including both estrogens and an-drogens; the insulin/insulin-like growth factor (IGF)-axis;and adipocytokines, including both metabolic and inflam-matory factors.

Sex HormonesHigh endogenous serum estrogen levels have been consis-tently associated with breast cancer risk in postmenopausalwomen6,9,13 and are, on average, higher in overweight andobese women than in normal-weight women, likely due theactivity of aromatase in adipose tissue.9 Estrogens are re-ported to be carcinogenic through at least two pathways,namely, the mitogenic/antiapoptotic activity of estrogens inthe breast as well as other tissues, and the possibly mutageniceffects of estrogen metabolites. Postmenopausal breast can-cer risk has additionally been associated with endogenousandrogens, including testosterone, DHEAS, and androstene-dion,14 which may reflect the conversion of androgens to es-trogens by aromotase. Serum androgen levels are, likeestrogen levels, correlated with adiposity. Recent prospectivedata suggest that the associations with estrogens and andro-gens may be specifıc, or at least strongest, for ER�/PR�tumors.13

The Insulin/IGF-AxisObese women have high rates of impaired fasting glucoseand diabetes, often accompanied by high circulating levels ofinsulin. In addition to insulin’s well-known metabolic activ-ity, it is also an important growth factor for a wide range oftissues and plays a substantial role in normal organogenesis.Indeed, insulin shares 40% amino acid sequence homologywith insulin-like growth factor (IGF)-I, the primarymediator ofthe effects of growth hormone, and also shares mitogenic/antiapoptotic activity and downstream signaling pathwayswith IGF-I.15As comparedwith insulin levels, however, levels of IGF-I in

individuals with impaired fasting glucose and early diabetesare variable.16 Thus, it is insulin, and not IGF-I, that is in-creasingly thought to play a major role in the obesity-diabetes-cancer relationship. In animal models, high insulinlevels have an independent mitogenic effect, and the insulinreceptor (IR) is highly expressed in cancers of the breast.15More specifıcally, tumors over-express the IR-A isoform,

which is important in fetal development but does not havea primary role in metabolism; this observation may helpexplain increased cancer risk related to obesity, despite thepresence of metabolic insulin resistance.15Diabetes has been prospectively associatedwith postmeno-

pausal breast cancer risk in many but not all studies.15 Ameta-analysis found an overall relative risk (RR) of 1.20 (95%CI: 1.11–1.30) for the association of diabetes and breast can-cer.17 This modest association could reflect the complicatedrelationship between diabetes and levels of insulin, as pa-tients may use antidiabetic medications, and insulin levelsmay fall in advanced diabetes. Although few large prospec-tive studies of insulin using fasting blood specimens havebeen reported, two of the three studies that addressedHTusefound a strong relation of insulin levels with breast canceramong non-HT users.6,18 The largest of these studies mea-sured fasting insulin, estradiol and IGF-I levels, and foundthat the highest compared with lowest insulin quartile wasassociated with a HR of 2.40 (95% CI: 1.30–4.41) innondiabetic/non-HT users.6 In other recent studies, it wasshown that patients with diabetes who used metformin, amedication that inhibits hepatic glucose production and re-duces circulating insulin levels, had lower risk of postmeno-pausal breast cancer risk than patients with diabetes who didnot use metformin.19

AdipokinesLeptin and adiponectin are the adipokines most extensivelystudied in relation to cancer.20Obese individuals tend to havehigh levels of leptin,whichhasmitogenic/antiapoptotic activity,and low levels of adiponectin, which has antimitogenic/pro-apoptotic activity. However there are only limited, con-flicting epidemiologic data regarding the relation of adi-ponectin levels and breast cancer risk. For example, a cross-sectional study byGaudet and colleagues (2010) failed to fındassociations between circulating adiponectin or other adipo-kines with postmenopausal breast cancer,21 whereas anothercross-sectional study reported that the leptin/adiponectin ra-tio was increased in postmenopausal breast cancer cases rel-ative controls.22 Overall, there is a paucity of relevantprospective cohort data regarding adipokines and breast can-cer, and future studies will need to additionally control forboth insulin and sex hormone levels.

RECURRENCE AND SURVIVALObesity has been repeatedly shown to be a risk factor forbreast cancer recurrence and poor survival. A meta-analysisof 43 studies, for example, found aHR of 1.33 (95%CI: 1.19–1.50) for breast cancer-related mortality and a similar HRfor all-cause mortality, when contrasting obese and non-obese patients with breast cancer.23 These association didnot differ by menopausal status,23 nor have results beenfound to differ by the ER status of tumors.23,24 Further, arecent study that accounted for use of screening, access to



treatment, the type of treatment, use of adjuvant therapy, andtumor characteristics found a more than two-fold increasedrisk of recurrence (HR� 2.43; 95%CI: 1.34–4.41) and breastcancer-related death (HR� 2.41; 95% CI: 1.00–5.81) amongobese patients compared with normal weight patients withbreast cancer.24Similarly, several studies found positive associations be-

tween high pretreatment insulin levels and poor progno-sis.25,26 For example, in a recent study, those with highinsulin levels had a two-fold worse distant disease-free sur-vival (HR � 2.05; 95% CI: 1.16–3.62) and overall survival(HR � 2.57; 95% CI: 1.48–5.50) than other patients withbreast cancer.25 Other studies have suggested that highleptin levels25 and low adiponectin levels were associatedwith poor prognosis.26 However, studies that concurrentlyaddress sex hormone, insulin, leptin, and adiponectin levels,before and after treatment, will be needed to comprehen-sively study these relationships with disease recurrence andsurvival.

METFORMIN AND OTHER PHARMACOLOGICINTERVENTIONSMetformin is an oral medication for treatment of type II di-abetes that acts by suppressing AMPK-mediated gluconeo-genesis in the liver, and results in lower serum insulin levels.It is well tolerated and does not induce hypoglycemia in non-diabetics. Metformin may reduce the risk of breast cancerand its recurrence through severalmechanisms.Most impor-tantly, in addition to its effects on insulin levels, metforminmay inhibit AMP-activated protein kinase activity in themTOR cancer-related pathway.27Meta-analyses have shown that metformin use in patients

with diabetes is associated with reduced risk of postmeno-pausal breast cancer OR � 0.82; 95% CI: 0.71–0.97),19 andreduced breast cancer mortality (summary relative risk �0.63; 95% CI: 0.40–0.99).28 This included improved overallsurvival in patients with diabetes with HER2� tumors, andpossibly distant metastasis free survival in patients with dia-beteswith triple-negative breast cancer.28 There are currentlyseveral clinical trials of metformin use as adjuvant therapy inthe treatment of patientswith breast cancer.27,29 In particular,there is a large phase III randomized clinical trial of 3,582womenwith stage I-III breast cancer (NCICMA.32)27, whichis the only studywith adequate sample size to assess the effectof metformin on breast cancer recurrence and mortality.Other pharmacologic regulators of AMPK are also being de-veloped, as are methods to increase adiponectin levels andagents that directly target the IR-A and IR/IGF-IR hybrid re-ceptors.However, considerable effort is additionally being put to-

ward the development of nonpharmacologic interventions.Given that obesity is a modifıable condition, behavioral in-terventions designed to change dietary and exercise practicesin patients represent a major opportunity to reduce breastcancer risk and improve disease outcomes.

BEHAVIORAL INTERVENTIONS TARGETING OBESITYWeight Loss Interventions in Breast Cancer SurvivorsUntil recently, therewere relatively few studies evaluating theeffıcacy and potential benefıts of weight loss interventions inpatients with breast cancer. Many of the early studies of life-style change in breast cancer survivors focused on elementsof weight maintenance, such as physical activity or dietaryquality (described below), and did not include weight loss asan objective. However, as the evidence linking obesity tobreast cancer recurrence andmortality has grown, alongwiththe proportion of U.S. adults who are overweight and obese,the need to develop and test weight loss interventions inbreast cancer populations has become more pressing.Several small studies have evaluated different weight loss

strategies in breast cancer survivors.30-34 Studies generallydemonstrate that weight loss can be successfully imple-mented in this population through a variety of interventions.For example, Shaw and colleagues randomized 64 obesewomen with early stage breast cancer to a low-fat diet group,a low-calorie diet group, or a control group, and demon-strated that participants randomized to low-fat and low-calorie diets experienced signifıcant weight loss andreductions in body fat as compared with controls (p � 0.006and 0.008, respectively).33 Another small study randomized48 obese breast cancer survivors to one of three dietary inter-vention arms (Weight Watchers, individualized dietarycounseling, or a combination of the two) or to a controlgroup31 and found that the individual counseling and com-bination groups lost signifıcantly more weight than controls(p� 0.05 at 12months for comparison of each group to con-trol), whereas the Weight Watchers alone group did not ex-perience signifıcant weight loss.Only one large-scale weight loss study has been reported to

date in breast cancer survivors. The Lifestyle InterventionStudy for Adjuvant Treatment of Early Breast Cancer(LISA)35 randomized 338 postmenopausal women with hor-mone receptor-positive breast cancer to an educational con-trol group or to a 2-year, telephone-based weight lossintervention, modeled on the Diabetes Prevention Program.The weight loss intervention focused on calorie reduction toattain a 500-1000 kcal per day defıcit; reduction in fat intaketo approximately 20% of calories; increased intake of fruits,vegetables, and grains; and increased physical activity to atleast 150 minutes of moderate-intensity recreational activityper week. Intervention participants lost signifıcantly moreweight than control participants at 6, 12, 18, and 24 months(4.3 vs 0.6 kg, p � 0.001 at 6 months and 3.6 vs 0.3 kg, p �0.001 at 24) and reported a signifıcant improvement in phys-ical functioning scores and higher levels of physical activityas compared with controls.35Several other weight loss intervention studies are currently

ongoing in breast cancer survivors (Table 2), although it isnot clear that any of these will be adequately powered to testthe effect of purposeful weight loss on the risk of breast can-cer recurrence or mortality. These studies will compare theeffects of different dietary interventions on weight change



and biomarkers linked to breast cancer prognosis and willalso evaluate the feasibility of in-person and distance-basedweight loss interventions, thus providing information thatwill be essential in the design of future studies evaluating theimpact of purposeful weight loss on disease outcomes inoverweight and obese women with early breast cancer.

Other Lifestyle Interventions in Breast CancerSurvivorsMany studies have looked at the feasibility and potentialbenefıts of modifying two key components of weight control(dietary quality and physical activity patterns) after breastcancer diagnosis. The two largest lifestyle interventionstudies in breast cancer survivors to date have been theWomen’s Interventional Nutrition Study (WINS)36 and theWomen’s Healthy Eating and Living (WHEL) study,37 whichboth looked at the impact of dietary interventions on dis-ease recurrence inwomenwith early-stage breast cancer. The

WHEL study randomized 3,088 breast cancer survivors to adietary intervention focused on increasing fruits, vegetables,and fıber and lowering fat intake, or to a usual care controlgroup.37 Intervention participants substantially increased in-take of fruits and vegetables and decreased percentage of di-etary calories from fat. The diet was designed to be isocaloricwith the baseline diet of study participants, and no weightloss was observed. There was no difference in recurrencerates in the diet and control groups (16.7 vs. 16.9%, p� 0.63).In contrast, the WINS randomized 2,437 women with early-stage breast cancer to a low-fat dietary intervention or to ausual care control group.36 Intervention participants sub-stantially decreased dietary fat intake, lost approximately 6pounds, and experienced signifıcantly better disease free sur-vival compared with controls (HR 0.76, 95% CI 0.60–0.98,p� 0.034).With further follow-up, these fındings lost statis-tical signifıcance, but an exploratory subgroup analysis dem-onstrated a signifıcant survival benefıt of the intervention in

TABLE 2. Select Ongoing Weight Loss Intervention Studies in Patients with Breast Cancer

N Population Study Design/Intervention(s) Aims

CHOICE50 370 - Postmenopausal breast cancersurvivors

Participants are allowed to choose1 of 3 arms:

-Primary: Change in C-reactive protein

-BMI 25–34.9kg/m2 -Control -Secondary: Changes in other metabolic andinflammatory biomarkers linked to recurrence

-Low-fat diet � exercise

-Low carb diet � exercise

DIANA-551 1208–2000 -Stage I-III breast cancer withinthe past 5 yr

1:1 randomization: -Primary: Breast cancer recurrence/new primary

-Age35–70 -Usual care control group -Secondary:

-One of the following: -“Mediterranean-Macrobiotic” lifestyle � Biomarkers

� ER- tumor � Weight loss or maintenance � Anthropometric measures

� Metabolic syndrome � Physical activity: 210 min/wk

� High serum insulin or testosterone � Modest calorie restriction

� Reduction in intake of foods with highglycemic index

ENERGY 800 -Stage I-III breast cancer 1:1 randomization: -Primary: weight loss

-BMI 25–45kg/m2 -Usual care control group -Secondary:

-Supervised weight loss intervention � Quality of life

� Calorie restriction � Serum biomarkers

� Increased physical activity

� Behavioral strategies & social support

SUCCESS-C52 `1200* -Newly diagnosed breast cancer: Lifestyle randomization: -Primary: Comparison of disease free survival betweengroups

� Node � OR -Usual care control group

High risk node -: ER-, age �35,grade 3 tumor, T2–3

-Telephone-based lifestyle intervention

� BMI 24–40 kg/m2 � Calorie restriction

� Low fat diet

� Increased physical activity to 150–200min/wk

*Weight loss portion of the SUCCESS-C trial will include the subset of participants with BMI 24 –40 kg/m2 in the parent adjuvant chemotherapy trial



ER- patients (HR0.41, p� 0.003).38 These fındings have beeninterpreted as providing further support for the role ofweight loss in breast cancer survivors.

Mixed interventions. A few recent studies have looked at theimpact of increased physical activity and dietary change onquality of life and other outcomes in breast cancer survivors.The Reach Out to Enhance Wellness (RENEW) study ran-domized 641 survivors of breast, prostate, and colon cancersto a 1-year telephone- and print materials-based lifestyle in-tervention, designed to increase physical activity and im-prove dietary quality, or to a wait-list control group.39 Thestudy demonstrated that participants randomized to the in-tervention group experienced signifıcantly less decline infunctional status, improvements in dietary quality, increasedphysical activity, and modest weight loss, as compared withcontrol participants (all p � 0.05).39 The FRESH Start trial40randomized 543 survivors of breast or prostate cancer to a10-month tailored, mail-based intervention, designed to in-crease intake of fruit and vegetables, decrease fat intake, andincrease exercise or to a control group who received nontai-loredmaterials about a healthy diet and exercise. Both groupssignifıcantly improved lifestyle behaviors (p� 0.05), but par-ticipants in the tailored intervention were more likely to im-prove two ormore behaviors (34% vs.18%, p� 0.001), and alsoexperienced modest weight loss and improvements in exer-cise patterns and dietary quality as compared with controlparticipants.

Physical activity interventions.Many studies have focused onthe feasibility and potential benefıts of physical activity in pa-tients with breast cancer both during and after adjuvant ther-apy.41,42 Given that studies have shown that decreasedphysical activity in the adjuvant period is one of the strongestpredictors of weight gain in patients with breast cancer,43 ef-fective means of increasing physical activity may be espe-cially important in breast cancer survivors. The AmericanCollege of Sports Medicine performed a comprehensive re-view of exercise intervention studies of populations of pa-tients with cancer, including data from 54 randomizedcontrolled trials of exercise in breast cancer survivors.44 Stud-ies evaluated a variety of exercise modalities including walk-ing, cycling, yoga, strength training, and rowing. In both theadjuvant and post-treatment settings, the review found con-sistent evidence that physical activity interventions were safeand led to increased aerobic fıtness and strength. Many stud-ies also found that individuals who participated in physicalactivity interventions experienced improvements in qualityof life, anxiety, depression, fatigue, body image, body size,and body composition, although not all studies were consis-tent in these fındings.

Lifestyle Intervention Trials with Biologic EndpointsLifestyle intervention studies also provide the opportunity tolearn more about the complex biologic mechanisms throughwhich obesity and other lifestyle factors could impact breastcancer risk and outcomes. Evaluating the effıcacy of lifestyle

interventions in affecting changes in biomarkers linked tobreast cancer risk and recurrence could help determinewhich interventions are most promising for further testing,as well as identifying the patient populations most likely tobenefıt from lifestyle change after breast cancer diagnosis.For example, the Nutrition and Exercise forWomen (NEW)Trial evaluated the effect of different lifestyle interventionson biomarkers linked to breast cancer in 439 postmeno-pausal, sedentary, overweight women. Participants were ran-domized to one of four groups: dietary weight loss, exercisealone, dietaryweight loss� exercise, or control.Women ran-domized to the weight loss groups, with or without exercise,experienced the most signifıcant reductions in insulin, othermetabolic hormones,45 sex steroids, and inflammatorymedi-ators.46 Women randomized to exercise alone experiencedsmaller but signifıcant changes in estrogen, testosterone andleptin levels (p� 0.001, p� 0.04, and p� 0.01, respectively),but no signifıcant changes in other hormones.Several studies have looked at the impact of exercise alone

on biomarkers linked to cancer risk and outcomes in breastcancer survivors. For example, Ligibel and colleagues dem-onstrated a 28% decrease in insulin levels in a group of sed-entary, overweight breast cancer survivors participating in a16-week mixed strength-training and aerobic exercise inter-vention (p � 0.07)47; similar effects were reported by Irwinand colleagues.48 Individual studies have also demonstrated afavorable impact of exercise on IGF-I and other metabolicand inflammatory hormones,49 but data are too limited todraw fırm conclusions.More work is needed to better under-stand the effect of lifestyle change on biologic mechanismslinked to breast cancer.

CONCLUSIONObesity is associated with breast cancer incidence and prog-nosis, and increasing evidence implicates sex hormones, in-sulin, adipokines, and their inter-related biologic pathwaysasmajor factors underlying these relationships.More impor-tantly, these hormones and biologic pathways represent im-portant potential targets for primary and secondary breastcancer prevention. TheMA-32 study, for example, is the fırstformal randomized clinical trial of metformin use and its ef-fect on prognosis in early-stage breast cancer.However, thereare additional potential biologic targets related to the impli-cated signaling pathways, and appropriate pharmacologicstudies to assess these opportunities for cancer preventionand treatment are needed. Furthermore, the ability of pur-poseful weight loss to improve prognosis has not been testedin the setting of randomized trials. Several lifestyle interven-tion trials have been conducted in breast cancer survivors,demonstrating the feasibility of implementing weight lossand other interventions after cancer diagnosis. Data fromtwo large-scale dietary intervention studies provide furtherevidence that dietary changes that produce weight loss mayimprove outcomes in breast cancer survivors, and otherstudies suggest that weight loss induces signifıcant favorablechanges in biomarkers linked to breast cancer risk and out-comes. Adequately powered adjuvant studies are needed to



defıne the role of weight loss in the management of over-weight and obese breast cancer survivors. Overall, the asso-ciation of obesity with breast cancer incidence and prognosis

represents a very signifıcant, but still under-exploited oppor-tunity to improve prognosis for patients with early-stagebreast cancer.



References

1. Ogden CL, Carroll MD, Kit BK, et al. Prevalence of Obesity in the US,2009-2010. NCHS Data Brief No. 2012;82:1-7.

2. OgdenCL, Yanovski SZ, CarrollMD, et al. The epidemiology of obesity.Gastroenterology. 2007;132:2087-2102.

3. Flegal KM, Graubard BI, Williamson DF, et al. Cause-specifıc excessdeaths associated with underweight, overweight, and obesity. JAMA.2007;298:2028-2037.

4. Renehan AG, Tyson M, Egger M, et al. Body-mass index and incidenceof cancer: a systematic review and meta-analysis of prospective obser-vational studies. Lancet. 2008;371:569-578.

5. Cheraghi Z, Poorolajal J, Hashem T, et al: Effect of body mass index onbreast cancer during premenopausal and postmenopausal periods: Ameta-analysis. PLoS One. 7:e51446.

6. GunterMJ,HooverDR,YuH, et al. Insulin, insulin-like growth factor-I,and risk of breast cancer in postmenopausal women. J Natl Cancer Inst.2009;101:48-60.

7. Huang Z, Hankinson SE, Colditz GA, et al. Dual effects of weight andweight gain on breast cancer risk [see comments]. JAMA. 1997;278:1407-1411.

8. Kerlikowske K,Walker R,Miglioretti DL, et al. Obesity, mammographyuse and accuracy, and advanced breast cancer risk. J Natl Cancer Inst.2008;100:1724-1733.

9. Key TJ, Appleby PN, Reeves GK, et al. Bodymass index, serum sex hor-mones, and breast cancer risk in postmenopausal women. J Natl CancerInst. 2003;95:1218-1226.

10. Li CI,MaloneKE, Daling JR. Interactions between bodymass index andhormone therapy and postmenopausal breast cancer risk (UnitedStates). Cancer Causes Control. 2006;17:695-703.

11. Suzuki R, Iwasaki M, Inoue M, et al: Body weight at age 20 years, sub-sequent weight change and breast cancer risk defıned by estrogen andprogesterone receptor status-the Japan public health center-based pro-spective study. Int J Cancer. 129:1214-1224.

12. PierobonM, Frankenfeld CL: Obesity as a risk factor for triple-negativebreast cancers: A systematic review and meta-analysis. Breast CancerRes Treat. 137:307-314.

13. Zhang X, Tworoger SS, Eliassen AH, et al: Postmenopausal plasma sexhormone levels and breast cancer risk over 20 years of follow-up. BreastCancer Res Treat. 137:883-892.

14. Kaaks R, Berrino F, Key T, et al. Serum sex steroids in premenopausalwomen and breast cancer risk within the European Prospective Inves-tigation into Cancer and Nutrition EPIC. J Natl Cancer Inst. 2005;97:755-765.

15. Strickler HD, Wylie-Rosett J, Rohan T, et al. The relation of type 2 dia-betes and cancer. Diabetes Technol Ther. 2001;3:263-274.

16. Rajpathak SN, Gunter MJ, Wylie-Rosett J, et al. The role of insulin-likegrowth factor-I and its binding proteins in glucose homeostasis and type2 diabetes. Diabetes Metab Res Rev. 2009;25:3-12.

17. Larsson SC,Mantzoros CS,Wolk A. Diabetesmellitus and risk of breastcancer: a meta-analysis. Int J Cancer. 2007;121:856-862.

18. Kabat GC, Kim M, Caan BJ, et al. Repeated measures of serum glucoseand insulin in relation to postmenopausal breast cancer. Int J Cancer.2009;125:2704-2710.

19. Col NF, Ochs L, SpringmannV, et al: Metformin and breast cancer risk:A meta-analysis and critical literature review. Breast Cancer Res Treat.135:639-646.

20. JardeT, Perrier S,VassonMP, et al:Molecularmechanisms of leptin andadiponectin in breast cancer. Eur J Cancer. 47:33-43.

21. Gaudet MM, Falk RT, Gierach GL, et al: Do adipokines underlie theassociation between known risk factors and breast cancer among a co-hort of United States women? Cancer Epidemiol. 34:580-586.

22. ChenDC, Chung YF, Yeh YT, et al. Serum adiponectin and leptin levelsin Taiwanese breast cancer patients. Cancer Lett. 2006;237:109-114.

23. Protani M, CooryM,Martin JH: Effect of obesity on survival of womenwith breast cancer: Systematic review and meta-analysis. Breast CancerRes Treat. 123:627-635.

24. Kamineni A, Anderson ML, White E, et al: Body mass index, tumorcharacteristics, and prognosis following diagnosis of early-stage breastcancer in a mammographically screened population. Cancer CausesControl. 24:305-312.

25. Goodwin PJ, Ennis M, Pritchard KI, et al: Insulin- and obesity-relatedvariables in early-stage breast cancer: Correlations and time course ofprognostic associations. J Clin Onco.l 30:164-171.

26. DugganC, IrwinML,Xiao L, et al: Associations of insulin resistance andadiponectin with mortality in women with breast cancer. J Clin Oncol.29:32-39.

27. Goodwin PJ, Stambolic V, Lemieux J, et al: Evaluation of metformin inearly breast cancer: A modifıcation of the traditional paradigm for clin-ical testing of anti-cancer agents. Breast Cancer Res Treat. 126:215-220.

28. Zhang P, Li H, Tan X, et al: Association of metformin use with cancerincidence andmortality: Ameta-analysis.Cancer Epidemiol. Epub 2013Jan 23.

29. Patterson RE, Rock CL, Kerr J, et al: Metabolism and breast cancer risk:Frontiers in research and practice. J Acad Nutr Diet. 113:288-296.

30. Goodwin P, Esplen MJ, Butler K, et al. Multidisciplinary weight man-agement in locoregional breast cancer: Results of a phase II study.BreastCancer Res Treat. 1998;48:53-64.

31. Djuric Z, DiLaura NM, Jenkins I, et al. Combining weight-loss counsel-ing with the weight watchers plan for obese breast cancer survivors.Obes Res. 2002;10:657-665.

32. Pakiz B, Flatt SW, Bardwell WA, et al: Effects of a weight loss interven-tion on bodymass, fıtness, and inflammatory biomarkers in overweightor obese breast cancer survivors. Int J Behav Med. 18:333-341

33. Shaw C, Mortimer P, Judd PA. A randomized controlled trial of weightreduction as a treatment for breast cancer-related lymphedema.Cancer.2007;110:1868-1874.



34. Thomson CA, Stopeck AT, Bea JW, et al: Changes in body weight andmetabolic indexes in overweight breast cancer survivors enrolled in arandomized trial of low-fat vs. reduced carbohydrate diets.Nutr Cancer.62:1142-1152.

35. Segal R, Pond G, Vallis M: Randomized trial of a lifestyle interventionfor women with early-stage breast cancer (BC) receiving adjuvant hor-mone therapy: Initial results. J Clin oncol. 29:(abstr 512, 2011).

36. Chlebowski RT, Blackburn GL, Thomson CA, et al. Dietary fat reduc-tion and breast cancer outcome: Interim effıcacy results from theWom-en’s Intervention Nutrition Study. J Natl Cancer Inst. 2006;98:1767-1776.

37. Pierce JP, Natarajan L, Caan BJ, et al. Influence of a diet very high invegetables, fruit, and fıber and low in fat on prognosis following treat-ment for breast cancer: The Women’s Healthy Eating and Living(WHEL) randomized trial. JAMA. 2007;298:289-298.

38. Chlebowski R, Blackburn GL, Thomson C: Survival analyses from theWomen’s Intervention Nutrition Study (WINS) evaluating dietary fatreduction and breast cancer outcome, American Society of Clinical On-cology. J Clin Oncol:(Abstr. 522, 2008).

39. Morey MC, Snyder DC, Sloane R, et al. Effects of home-based diet andexercise on functional outcomes among older, overweight long-termcancer survivors: RENEW: A randomized controlled trial. JAMA. 2009;301:1883-1891.

40. Demark-Wahnefried W, Clipp EC, Lipkus IM, et al. Main outcomes ofthe FRESH START trial: A sequentially tailored, diet and exercisemailed print intervention among breast and prostate cancer survivors.J Clin Oncol. 2007;25:2709-2718.

41. Markes M, Brockow T, Resch KL: Exercise for women receiving adju-

vant therapy for breast cancer. Cochrane Database Syst Rev:CD005001,2006.

42. Galvao DA,Newton RU. Review of exercise intervention studies in can-cer patients. J Clin Oncol. 2005;23:899-909.

43. Demark-Wahnefried W, Peterson BL, Winer EP, et al. Changes inweight, body composition, and factors influencing energy balanceamong premenopausal breast cancer patients receiving adjuvant che-motherapy. J Clin Oncol. 2001;19:2381-2389.

44. Schmitz KH, Courneya KS, Matthews CE: American College of SportsMedicine roundtable on exercise guidelines for cancer survivors. MedSci Sports Exerc. 2010;42:1409-1426.

45. Mason C, Foster-Schubert KE, Imayama I, et al: Dietary weight loss andexercise effects on insulin resistance in postmenopausal women. Am JPrev Med. 41:366-375.

46. Campbell KL, Foster-Schubert KE, Alfano CM, et al: Reduced-caloriedietary weight loss, exercise, and sex hormones in postmenopausalwomen: Randomized controlled trial. J Clin Oncol. 30:2314-2326.

47. Ligibel JA, Campbell N, Partridge A, et al. Impact of a mixed strengthand endurance exercise intervention on insulin levels in breast cancersurvivors. J Clin Oncol. 2008;26:907-912.

48. Irwin ML, Varma K, Alvarez-Reeves M, et al. Randomized controlledtrial of aerobic exercise on insulin and insulin-like growth factors inbreast cancer survivors: The Yale Exercise and Survivorship study.Can-cer Epidemiol Biomarkers Prev. 2009;18:306-313.

49. Ballard-Barbash R, Friedenreich CM, Courneya KS, et al: Physical ac-tivity, biomarkers, and disease outcomes in cancer survivors: A system-atic review. J Natl Cancer Inst. 104:815-840.



BREAST CANCER

Optimizing Locoregional Treatmentin Early-Stage Breast Cancer

CHAIRStephen R. Grobmyer, MDCleveland ClinicCleveland, OH

SPEAKERSHiram S. Cody III, MDMemorial Sloan-Kettering Cancer CenterNew York, NY

Timothy Whelan, BM, BCh, MScJuravinski Cancer Centre at Hamilton Health SciencesHamilton, ON, Canada

Which Patients with Sentinel Node–Positive Breast Cancer CanAvoid Axillary Dissection?Alice Y. Ho, MD, MBA, and Hiram S. Cody III, MD

OVERVIEW

Sentinel lymph node (SLN) biopsy is standard care for patients with cN0 breast cancer. An extensive literature, including sevenrandomized trials, has established that patients with negative SLN do not require axillary dissection (ALND), that axillary localrecurrence after a negative SLN biopsy is rare, that disease-free and overall survival are unaffected by the addition of ALND to SLNbiopsy, and that the morbidity of SLN biopsy is substantially less than that of ALND. It is now clear that many patients with positiveSLN do not require ALND. In ACOSOG Z0011, 6-year locoregional control and survival were equivalent with versus without theperformance of ALND in cT1–2N0 patients with �2 positive SLN treated by breast conservation with whole breast radiation therapy.A small but growing body of data now suggests that ALND may not be required for selected patients outside the Z0011 eligibility criteria,specifically those treated by mastectomy (without post-mastectomy radiation therapy), by partial breast irradiation, and by neoad-juvant chemotherapy. Looking ahead, the principal goals of axillary staging, prognostication, and local control will be accomplished bySLN biopsy for a substantial majority of patients, and the role of ALND will continue to diminish.

The widespread adoption of sentinel lymph node (SLN)biopsy as standard care for axillary staging in cN0 breast

cancer is supported by the results of at least 69 observationalstudies,1 seven randomized trials,2 3 meta-analyses,2-4 anASCOGuideline,5 and an extensive literature covering all as-pects of the procedure. These studies establish that patientswith negative SLNdonot require axillary dissection (ALND),that axillary local recurrence (LR) after a negative SLNbiopsyis rare (0.3%),6 that disease-free (DFS) and overall survival(OS) are unaffected by the addition of ALND to SLN biopsy,and that the morbidity of SLN biopsy is less than that ofALND. The logical next question in the evolution of axillarystaging is to ask whether there are SLN-positive patients whocan avoid ALND, and it is clear that there are: 30% to 50% ofSLN-positive patients have disease limited to the SLN.1In a retrospective study of practice patterns in the United

States, drawing on the National Cancer Data Base, Bilimoriaet al7 reported on 97,314 SLN-positive patients (1998–2006);23% with SLN macrometastases (�2 mm, pN1) and 55%with SLNmicrometastases (0.2–2 mm, pN1mi) did not haveALND, yet for both pN1 and pN1mi SLN disease, axillarylocal recurrence and 5 year relative survival were the samewith or without ALND (Table 1). A similar study by Yi et al8from the SEER Database of 26,986 SLN-positive patients(1998–2004) found that 11% of those with SLNmacrometas-tases and 33% of those with SLN micrometastases did nothave ALND, and that OS was unaffected at a median

follow-up of 50 months. Both studies report a strong trendover time away from ALND for patients with SLNmicrome-tastases. Nine smaller retrospective studies9 comprising1,035 patients with positive SLN and no ALND report lowrates of axillary LR, most in the range of 0% to 2%, at 28–82months follow-up.The most defınitive data are from ACOSOG Z0011,10,11 a

prospective randomized trial in which 813 SLN-positive pa-tients with clinical stage T1–2N0 breast cancerwere random-ized to ALND compared with no further surgery. All wereSLN-positive by routine hematoxylin and eosin staining andall had breast conservation including whole-breast radiationtherapy (RT). Patients with three or more positive SLN (orwith matted nodes) were excluded, and axillary-specifıc RTwas not allowed. Additional positive nodes were found in27% of the patients who had ALND, but at 6 years’ follow-upthere were no differences between the ALND and no-ALNDarms in local (3.6% vs. 1.9%), regional (0.5% vs. 0.9%), oroverall locoregional recurrence (4.1% vs. 2.8%),10, nor werethere any differences in DFS or OS11 (Table 2).Critics of Z0011 have focused on issues of case selection

(arguing that youngwomen and thosewith estrogen receptor[ER]-negative tumors were under-represented), follow-up(arguing that 6.3 years is inadequate), and statistical power(arguing that Z0011 did not meet its planned accrual or sta-tistical endpoints). In defense of Z0011, Morrow and Giu-liano12 respond, rightly inmy opinion, as follows: 1) younger

From the Department of Radiation Oncology, Memorial Sloan-Kettering Cancer Center, New York, NY; Breast Service, Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York,NY; Weill Cornell Medical College, New York, NY.


Corresponding author: Hiram S. Cody III, MD, Memorial Sloan-Kettering Cancer Center, 300 East 66th St, #831, New York, NY 10065; email: [email protected].


AVOIDING AXILLARY DISSECTION


age is associated with higher rates of recurrence in the ipsi-lateral breast, but not in regional nodes, 2) ER-negative tu-mors are associated with early relapse but not with higherrates of axillary node involvement, 3)mostwomenwithin theZ0011 selection criteria are postmenopausal and ER-positive, 4) axillary recurrence is an early event (virtually alloccur within the fırst 5 years), and 5) Z0011 closed early(based on slow accrual and a lower-than-expected rate ofevents) but achieved its predefıned goal, showing with a highlevel of signifıcance that SLN biopsy alone was not inferior toALND.The principal implications of Z0011 are surgical, and over

the last 2 years many institutions and surgeons in the UnitedStates (and to a lesser extent in Europe and worldwide) havefound the results to be persuasive and practice-changing,incorporating into their treatment guidelines a policy of“no ALND” for SLN-positive patients who meet the Z0011

selection criteria. At our institution, we have done so since2010; for “Z0011-like” patients we have largely abandonedpreoperative axillary ultrasound, axillary fıne needle aspira-tion core biopsy, and intraoperative frozen section of SLN,with a substantial decline in the rate of completionALND forSLN-positive patients. Our observations match those of theUniversity of Texas MD Anderson Cancer Center,13 whichcompared their “pre-Z0011” verses “post-Z0011” practices,observing declines in the rates of intraoperative pathologicassessment from 69% to 26%, and of ALND for SLN-positivepatients from 85% to 24%. In a separate report, they esti-mated that 75% of their SLN-positive patients from the“pre-Z0011” era could have avoided ALND.14 Among re-spondents to a 2012 survey of the American Society of BreastSurgeons,15 97% expressed familiarity with Z0011% and 57%said they would not performALND in “Z0011-like” patients.The response to Z0011 worldwide, and especially in Europe,has been mixed; the 2011 St. Gallen Consensus16 acknowl-edged the results of Z0011 without making a straightforwardrecommendation, and cautioned that a policy of no-ALNDfor SLN-positive patients should adhere strictly to the Z0011selection criteria.What are the implications of Z0011 for the medical oncol-

ogist? Montemurro et al17 have suggested that the omissionof ALND could alter the choice of systemic therapy; usingpost hoc case review, they found that the information gainedfrom completion ALND could have changed the indicationfor systemic chemotherapy in 16% of their patients. Reassur-ingly, two large trials which randomized SLN-positive pa-tients to ALND verses no-ALND (ACOSOG Z001111) andALND verses axillary RT (EORTC AMAROS18) found nodifferences in the usage of chemotherapy, hormone therapy,or RT based on the performance of ALND.What are the implications of Z0011 for the radiation on-

cologist? Positive axillary nodes were left behind in a pre-sumed 27% of the Z0011 patients in the no-ALND arm, butonly 0.9% developed axillary LR. Although Z0011 did not al-low the use of supraclavicular or axillary fıelds, Haffty et al19

KEY POINTS

� Sentinel lymph node biopsy (SLN) is well-established asstandard care for virtually all patients with cN0 breastcancer.

� For SLN-negative patients, long-term locoregional controland survival of SLN biopsy alone versus SLN biopsy/axillarydissection (ALND) are comparable, and the morbidity ofSLN biopsy is less.

� For SLN-positive patients who meet the ACOSOG Z0011entry criteria (cT1–2N0 disease, �2 SLN positive, treatedby breast conservation with whole-breast radiationtherapy), locoregional control, and survival of SLN biopsyalone versus SLN biopsy/ALND are comparable, and ALND isnot required.

� For selected SLN-positive patients outside the Z0011 entrycriteria (i.e., mastectomy without post-mastectomyradiation therapy, partial-breast irradiation, or neoadjuvantchemotherapy) ALND may not be obligatory, but furtherstudy is required in well-characterized patient cohorts.

� The future role of ALND will increasingly be to treat locallypersistent or recurrent disease, not to prevent it.

TABLE 1. Results of SLN Biopsy with and withoutALND, by Size of SLN Metastasis, in SLN-PositivePatients: National Cancer Data Base (1998–2006)*

Axillary LocalRecurrence (5 yr)

RelativeSurvival (5 yr)

SLN micrometastases (�2 mm, pN0i�/N1mi)

SLN only (n � 802) 0.4% 99%

SLN/ALND (n � 2,357) 0.2% 98%

SLN macrometastases (�2 mm, pN1)

SLN only (n � 5,596) 1.0% 90%

SLN/ALND (n � 22,591) 1.1% 89%

Abbreviations: SNL ,sentinel lymph node; ALND, axillary lymph node dissection; yr, year.

*Adapted from Bilimoria et al7

TABLE 2. Results of ACOSOG Z0011, a RandomizedComparison of SLN Biopsy with and without ALND, inSLN-Positive Patients*

SLN Biopsy/ALNDn � 388

SLN Biopsy/No ALNDn � 425

Loco-regional recurrence (6.3 yr)

Local 3.6% 1.9%

Regional node 0.5% 0.9%

Local�Regional 4.1% 2.8% (P � NS)

Survival (6.3 yr)

Disease-free survival 82% 84% (p � NS)

Overall survival 92% 93% (p � NS)

Abbreviations: SNL ,sentinel lymph node; ALND, axillary lymph node dissection; yr, year;H&E, hematoxylin and eosin stain; RT, radiation therapy; NS, not significant.

*Adapted from Giuliano et al10,11: all patients were clinical stage T1–2N0, with �2 H&E-positive SLN, treated by breast conservation with whole-breast and without axillary-specificRT

HO AND CODY


have suggested that irradiation of the lower axillary nodeswith “high tangents” to the breast may have contributed tothe low rate of axillary LR. Modern computed tomography(CT)-guided treatment planning allows identifıcation of theaxillary nodes and treatment of at least part of the axilla byadjusting the superior and deep tangent borders. Resnik etal20 have used CT to estimate the proportion of the prescrip-tion dose to the breast given to the axilla by standard breasttangents (66% to level I, 44% to level II, and 31% to level III)versus “high” breast tangents (86%, 71%, and 73%, respec-tively), and recommend high tangents for axillary prophy-laxis. Schlembach et al 21 have described a technique for hightangents that allows treatment of the SLN region andmost oflevels I and II. An audit of the Z0011 treatment fıelds is on-going to clarify the degree to which participating radiationoncologists adjusted their tangent fıelds based on tumorcharacteristics and the extent of axillary surgery, but thesenuances of intentmay be diffıcult to extract in a retrospectivestudy.Some confusion comes from the recent results of theNCIC

Clinical Trials MA.20,22 a trial in which patients with high-risk node-negative and 1–3 node-positive breast cancer wererandomly assigned to whole breast RT with or without re-gional node irradiation (RNI). These patients, 85% of whomhad 1–3 positive nodes, overlap somewhat with the Z0011population, none of whom received RNI. At 5 years, MA.20observed notable improvements for the RNI arm in isolatedlocoregional DFS (by 2.3%), distant DFS (by 5.4%), DFS (by5.7%) and OS (by 1.6%, p � 0.07). These results require fur-ther study (the full paper has not yet been published) and areperplexing; inexplicably, the absolute gain in DFS exceededthe gain in local control, a result which stands in sharp con-trast to theOxfordmeta-analysis,23 which famously observedthat the prevention of four local recurrences was required tosave one life.Looking ahead, can we extend the success of Z0011 to

Z0011-ineligible patients, specifıcally those treated a) bymastectomy without RT, b) by partial breast RT (PBI), andc) by neoadjuvant chemotherapy (NAC)?Regardingmastectomy, we have recently reported24 on 535

SLN-positive patients from the pre-Z0011 erawho had eithermastectomy or breast conservation without other axillary-specifıc treatment: among 234 with N1mi or N1 disease,there were no differences at 4 years in regional node recur-rence between mastectomy (97 patients, 2.5%) and breastconservation (134 patients, 1.5%). On the plus side, this lowevent rate is encouraging but requires wider confırmation inprospective studies specifıc to mastectomy; on the minusside, a low event rate argues that a “Z0011-for-mastectomy”randomized trial would not be feasible. In lieu of a random-ized trial, well-characterized case series with adequatefollow-up of SLN-positivemastectomy patients treated with-out RT should be suffıcient to address this issue.Regarding PBI, the MammoSite Registry Trial (in which

PBI is delivered through an intracavitary balloon) has re-ported 5-year axillary LR of 0.8% in SLN-negative patients,25a result quite similar to that of negative SLNbiopsy in general

(0.3%). A more defınitive answer comes from the TARGITTrial,26 an international multicenter randomization of PBIgiven as a single intraoperative dose to the tumor site (in1,113 patients) verses conventional whole-breast RT (in1,119 patients); 4-year local recurrence rates were 1.20% and0.95%, respectively (p� 0.41), and amore recent analysis hasfound no differences between the treatment arms in the rateof axillary LR (J. S. Vaidya, oral communication February,2013). Both studies suggest that the effect of whole-breast RTon axillary LR in SLN-negative patients ismodest at best. An-other trial, NSABP B-39,27 compares PBI (delivered by intra-cavitary balloon, brachytherapy catheters, or postoperativeexternal beam) with whole-breast RT and has not yet re-ported outcomes. There are no data reporting outcome inSLN-positive patients treated with PBI and without ALND,but it is reasonable to assume that the results would be at leastas good as for mastectomy and, to the extent that the PBIpatients usually have earlier-stage disease, probably better.For the same reasons noted above, a “Z0011-for-PBI” trial isunlikely. Interestingly, in the American Society of BreastSurgeons survey,15 36% of respondents would not performALND in patients receiving PBI, and 27% would consideromitting ALND in patients not planned to receive RT at all.These results fıt the nationwide trends noted by Bilimoria etal7 in their survey of the National Cancer Data Base.Regarding neoadjuvant chemotherapy (NAC), the false-

negative rate of SLN biopsy after NAC (in 27 studies com-prising 2,148 patients) is roughly comparable to that ofSLN biopsy in general, 10.5%.28 Hunt et al29 have reportedthat for patients who were cN0 at the time of diagnosis, thefalse-negative rates of SLN biopsy done before versus afterchemotherapy are virtually identical (4.2% vs. 5.9%). Theperformance of SLN biopsy following NAC in patients withbiopsy-proven nodal metastases is the subject of a prospec-tive validation study, ACOSOG 1071, recently reported byBoughey et al30 at the San Antonio Breast Cancer Sympo-sium; 40% of patients had a pathologic complete response intheir lymph nodes and among 607 patients with cN1 diseasewho receivedNAC followed by a successful SLNbiopsy and aplanned backup ALND, the false-negative rate was 12.8%. Anew report by Mamounas et al31 on the 10-year patterns oflocoregional recurrence after NAC in National Surgical Ad-juvant Breast and Bowel Project (NSABP) trials B-18 andB-27 demonstrates the highest rates of regional node recur-rence in clinically node-positive patients whose nodes re-mained positive after NAC. Taken together, these datasuggest that ALND may not be required for cN0 or cN1 pa-tients whose SLN are negative after NAC, but should bestrongly considered whenever the SLN remain positive afterNAC. Many questions remain unanswered, and the decisionfor ALND (and/or nodal RT) following NAC can be com-plex; patient age, pretreatment T/N stage, biologic subtype oftumor, the clinical response to NAC, the operation (mastec-tomy vs. breast conservation), the pathologic response inbreast and nodes, and the anticipated rate of local recurrence.This complexity argues against a “Z0011-for-neoadjuvant”trial.



Looking further ahead, we must ask whether axillary stag-ing is necessary at all; the 21-gene recurrence score32 appearsto give better prognostication and prediction of response totherapy than conventional histopathology for node-negative33 and possibly for node-positive patients34 and is the

subject of large randomized trials (TAILORx35 and Rx-PONDER [swog.org/rxponder/]). It is quite possible that the“SLN biopsy of the future” may be no SLN biopsy at all andthat the “ALNDof the future” will be limited to the salvage oflocally persistent or recurrent disease.



References

1. Kim T, Giuliano AE, Lyman GH. Lymphatic mapping and sentinellymph node biopsy in early-stage breast carcinoma: a metaanalysis.Cancer. 2006;106:4-16.

2. KellM, Burke J, BarryM, et al.Outcomeof axillary staging in early breastcancer: Ameta-analysis. Breast Cancer Res Treat. 2010;120:441-447.

3. Wang Z, Wu LC, Chen JQ. Sentinel lymph node biopsy compared withaxillary lymph node dissection in early breast cancer: a meta-analysis.Breast Cancer Res Treat. 2011;129:675-689.

4. Pesek S, Ashikaga T, Krag LE, et al. The false-negative rate of sentinelnode biopsy in patients with breast cancer: a meta-analysis. WorldJ Surg. 2012;36:2239-2251.

5. Lyman GH, Giuliano AE, Somerfıeld MR, et al. American Society ofClinical Oncology guideline recommendations for sentinel lymph nodebiopsy in early-stage breast cancer. J Clin Oncol. 2005;23:7703-7720.

6. van der Ploeg IM, Nieweg OE, van Rijk MC, et al. Axillary recurrenceafter a tumour-negative sentinel node biopsy in breast cancer patients:A systematic review and meta-analysis of the literature. Eur J Surg On-col. 2008;34:1277-1284.

7. Bilimoria KY, Bentrem DJ, Hansen NM, et al. Comparison of sentinellymph node biopsy alone and completion axillary lymph node dissec-tion for node-positive breast cancer. J Clin Oncol. 2009;27:2946-2953.

8. Yi M, Giordano SH, Meric-Bernstam F, et al. Trends in and outcomesfrom sentinel lymph node biopsy (SLNB) alone vs. SLNB with axillarylymph node dissection for node-positive breast cancer patients: Experi-ence from the SEER database. Ann Surg Oncol. 2010;17 Suppl 3:343-351.Epub 2010 Sep 19.

9. Cyr A, Gao F, GillandersWE, et al. Disease recurrence in sentinel node-positive breast cancer patients forgoing axillary lymph node dissection.Ann Surg Oncol. 2012;19:3185-3191.

10. Giuliano AE, McCall L, Beitsch P, et al. Locoregional recurrence aftersentinel lymph node dissection with or without axillary dissection inpatients with sentinel lymph node metastases: the American College ofSurgeons Oncology Group Z0011 randomized trial. Ann Surg. 2010;252:426-433.

11. Giuliano AE, Hunt KK, Ballman KV, et al. Axillary dissection vs no ax-illary dissection inwomenwith invasive breast cancer and sentinel nodemetastasis: a randomized clinical trial. JAMA. 2011;305:569-575.

12. Morrow M, Giuliano AE. To cut is to cure: can we really apply Z11 inpractice? Ann Surg Oncol. 2011;18:2413-2415.

13. Caudle AS, Hunt KK, Tucker SL, et al. American College of SurgeonsOncology Group (ACOSOG) Z0011: impact on surgeon practice pat-terns. Ann Surg Oncol. 2012;19:3144-3151.

14. YiM, KuererHM,Mittendorf EA, et al. Impact of the AmericanCollegeof Surgeons Oncology Group Z0011 Criteria Applied to a Contempo-rary Patient Population. J Am Coll Surg. 2013;216:105-113. Epub 2012Nov 2.

15. Gainer SM, Hunt KK, Beitsch P, et al. Changing behavior in clinicalpractice in response to the ACOSOG Z0011 trial: a survey of the Amer-ican Society of Breast Surgeons. Ann Surg Oncol. 2012;19:3152-3158.

16. Goldhirsch A, Wood WC, Coates AS, et al. Strategies for subtypes-dealing with the diversity of breast cancer: highlights of the St. GallenInternational Expert Consensus on the Primary Therapy of Early BreastCancer 2011. Ann Oncol. 2011;22:1736-1747.

17. Montemurro F, Maggiorotto F, Valabrega G, et al. Omission of axillarydissection after a positive sentinel node dissection may influence adju-vant chemotherapy indications in operable breast cancer patients. AnnSurg Oncol. 2012;19:3755-3761.

18. StraverME,MeijnenP, vanTienhovenG, et al. Role of axillary clearanceafter a tumor-positive sentinel node in the administration of adjuvanttherapy in early breast cancer. J Clin Oncol. 2010;28:731-737.

19. Haffty BG, Hunt KK, Harris JR, et al. Positive sentinel nodes withoutaxillary dissection: implications for the radiation oncologist. J Clin On-col. 2011;29:4479-4481.

20. Reznik J, Cicchetti MG, Degaspe B, et al. Analysis of axillary coverageduring tangential radiation therapy to the breast. Int J Radiat Oncol BiolPhys. 2005;61:163-168.

21. Schlembach PJ, Buchholz TA, Ross MI, et al. Relationship of sentineland axillary level I-II lymph nodes to tangential fıelds used in breastirradiation. Int J Radiat Oncol Biol Phys. 2001;51:671-678.

22. Whelan T, Olivotto I, Ackerman I, et al. NCIC-CTGMA. 20: An inter-group trial of regional nodal irradiation in early breast cancer. J ClinOncol. 2011;29[18 suppl].

23. Clarke M, Collins R, Darby S, et al. Effects of radiotherapy and of dif-ferences in the extent of surgery for early breast cancer on local recur-rence and 15-year survival: an overview of the randomised trials.Lancet.2005;366:2087-2106.

24. Milgrom S, Cody H, Tan L, et al. Characteristics and outcomes of sen-tinel node-positive breast cancer patients after total mastectomy with-out axillary-specifıc treatment. Ann Surg Oncol. 2012;19:3762-3770.

25. Aburabia M, Roses RE, Kuerer HM, et al. Axillary failure in patientstreated with MammoSite accelerated partial breast irradiation. AnnSurg Oncol. 2011;18:3415-3421.

26. Vaidya JS, JosephDJ, Tobias JS, et al. Targeted intraoperative radiother-apy versus whole breast radiotherapy for breast cancer (TARGIT-A tri-al): an international, prospective, randomised, non-inferiority phase 3trial. Lancet. 2010;376:91-102.

27. No authors listed. NSABP B-39, RTOG 0413: A Randomized Phase IIIStudy of conventional whole breast irradiation versus partial breast ir-radiation for women with stage 0, I, or II breast cancer.Clin Adv Hema-tol Oncol. 2006;4:719-721.

28. van Deurzen CH, Vriens BE, Tjan-Heijnen VC, et al. Accuracy of sen-

HO AND CODY


tinel node biopsy after neoadjuvant chemotherapy in breast cancer pa-tients: a systematic review. Eur J Cancer. 2009;45:3124-3130.

29. Hunt KK, Yi M, Mittendorf EA, et al. Sentinel lymph node surgeryafter neoadjuvant chemotherapy is accurate and reduces the need foraxillary dissection in breast cancer patients. Ann Surg. 2009;250:558-566.

30. Boughey JC, Suman VJ, Mittendorf EA, et al. The role of sentinellymph node surgery in patients presenting with node positive breastcancer (T0-4, N1-2) who receive neoadjuvant chemotherapy: Resultsfrom the ACOSOG Z1071 trial. Paper presented at: 2012 San Anto-nio Breast Cancer Symposium; Abstract S2-1; December, 2012; SanAntonio, TX.

31. Mamounas EP, Anderson SJ, Dignam JJ, et al. Predictors of locoregionalrecurrence after neoadjuvant chemotherapy: results from combined

analysis of National Surgical Adjuvant Breast and Bowel Project B-18and B-27. J Clin Oncol. 2012;30:3960-3966.



34. AlbainKS, BarlowWE, Shak S, et al. Prognostic and predictive value of the21-gene recurrence score assay in postmenopausal women with node-positive, oestrogen-receptor-positivebreast canceronchemotherapy: a ret-rospective analysis of a randomised trial. Lancet Oncol. 2010;11:55-65.

35. Sparano JA. TAILORx: trial assigning individualized options for treat-ment (Rx). Clin Breast Cancer. 2006;7:347-350.



Optimizing Clinical Management of Surgical Margins inBreast-Conserving Therapy for Breast CancerStephen R. Grobmyer, MD, Michael S. Cowher, MD, and Joseph P. Crowe, MD

OVERVIEW

There has been, and continues to be, significant controversy over the definition of an “optimal” surgical margin in breast-conservingtherapy (BCT). The historic basis of this controversy stems from the original trials documenting the safety of BCT and many conflictingretrospective studies that have sought to define the association between surgical margin width and outcomes over the last 20 years.It is important to understand that margin assessment is an inexact science, and current laboratory approaches to surgical-marginassessment represent only a sampling of the surgical margin. Currently available evidence suggests that decisions regarding surgicalmargins in BCT should be made in the context of what is known about the biology of breast cancer, as well the interactions of tumorbiology, adjuvant treatment for breast cancer, and outcomes. Achieving consensus on management of surgical margins in BCT shouldbe a clinical priority as it offers the opportunity to reduce the burden of breast cancer treatment on patients without compromisingcancer-related outcomes.

Breast conserving therapy (BCT) consisting of lumpec-tomy followed by adjuvant breast irradiation and sys-

temic therapy has become a widely accepted approach to thetreatment of early-stage breast cancer.1 Numerous studieshave demonstrated very high rates of local recurrence (�20%at 5 years) of breast cancer in patients in whom microscopi-cally clear margins (“negative”) are not achieved withlumpectomy even in the setting of adjuvant radiation therapy(reviewed byKlimberg et al.).2 Complete surgical resection ofbreast cancer to negativemicroscopicmargins is therefore anessential component of BCT.2 There has been, and continuesto be, controversy regarding the value of obtaining “wide”microscopicmargins in patients undergoing BCT in an effortto improve outcomes and reduce recurrences.3Wider surgical margins are a goal of some surgeons and

treatment teams who seek to remove any residual disease inthe hopes of improving outcomes.4However, the goal of BCTand lumpectomy is to preserve the volume and shape of thebreast, and it is desirable to minimize the amount of normaltissue removed around a tumor that is associated with theminimal risk of recurrence.2 The “optimal” surgical marginfor a lumpectomy for the purposes of this article is defınedas the minimum margin of normal tissue around a breastcancer that is associated with the lowest achievable ratesof recurrence and highest rates of breast cancer related sur-vival in the setting of modern adjuvant therapies. Based onthis defınition, removal of more than optimal tissue has

no additional value in relation to patient outcomes andcan be detrimental in terms of cosmesis, increased rates ofre-excision, costs, patient anxiety, and delay in adjuvanttherapy.5Reporting the distance from the edge of the tumor to the

edge of the surgical specimen (“margin width”) at the time oflumpectomyhas become a frequently reported qualitymetricin breast cancer care. However, there has been no generalconsensus on how to use this measurement to optimallyguide further surgical therapy.6 There is a need to developconsensus guidelines in defıning the optimal surgical marginin BCT. Defıning the optimal surgical margin as this has thepotential to reduce overtreatment without negatively im-pacting the survival of patientswith early-stage breast cancer.There are no prospective randomized trials that directly ad-dress the influence of measured margin status on outcomesin BCT for invasive breast cancer or ductal carcinoma in situ(DCIS).In this article, we review the evidence concerning the rela-

tionship between margin status and outcomes in BCT. Thisdata will be reviewed with a specifıc focus on the limitationsof current approaches to margin assessment, other clinicaland biologic factors that govern clinical outcomes, as wellfactors associated with a residual-disease burden followinglumpectomy. This report will not address optimal surgicalmargins in patients with breast cancer undergoing lumpec-tomies who do not receive adjuvant radiotherapy.

From the Section of Surgical Oncology, Cleveland Clinic, Cleveland, OH.


Cooresponding author: Stephen R. Grobmyer, MD, Section of Surgical Oncology, Cleveland Clinic, 9500 Euclid Ave./A81, Cleveland, OH 44195; email: [email protected].


GROBMYER, COWHER, AND CROWE


THE HISTORIC BASIS OF CONCERNS OVER “CLOSE”MARGINS IN BCTInvasive Breast CancerTable 1 summarizes the surgical margin criteria used by in-vestigators in the large randomized trials that formed the ba-sis of modern BCT. These trials varied greatly in theirminimum requirements for surgical margins. All trials re-quired at least complete gross removal of the tumor at thetime of lumpectomy (Table 1). These trials, performed in the1970s and 1980s, and now with long-term follow-up, dem-onstrated that patients undergoing BCT had overall survivalcomparable with patients having mastectomy for breast can-cer. In only one of these trials, theNational Surgical AdjuvantBreast and Bowel Project B-06 reported by Fisher et al., didinvestigators record the margin width of the lumpectomyspecimens, but simply having a negativemargin was satisfac-tory for trial-entry requirements.1 When viewed together,these trials suggest that variation in surgical margins andin particular wide surgical margins are not a determinantof survival in patients undergoing BCT for invasive breastcancer.Subsequent to the publication of the prospective trials,

there have been many retrospective studies that have inves-tigated the concept of “close” margins and the effect of“close” margins on outcomes (reviewed in Singletary).7 Inthese studies, the defınition of “close” margin varied and wasdefıned by the investigators in each individual study (e.g.,�1mm or �2 mm). Some studies have suggested that “close”margins are associated with recurrence rates similar to thosefor patientswith positivemargins.8Other studies have shownintermediate outcomes for those with close margins com-pared with “positive” and “negative.” Others have shownthat “close” margins have no effect on relapse-free survivalcompared with widely negative margins. This uncertaintyhas prompted many to advocate greater initial tissue resec-

tion or re-excision for “close” margins in an effort to reducelocal recurrence (or improve relapse-free survival) for pa-tients with invasive breast cancer. Reported ranges of localrecurrence rates at 5 years for clear margins, close margins,and positive margins are 2%-3%, 2%-8%, and 10%-25%,respectively.7Houssami et al.9 recently reported the results of a meta-

analysis of 21 studies analyzing the relationship betweenmargin status and outcome for patients having BCT for in-vasive breast cancer. The authors demonstrated that al-though increasing margin width has a weak correlation withincreased risk for local recurrence, the effect is mitigatedwhen adjustments are made for administration of adjuvantendocrine and radiation therapy. The authors conclude,“adoption of wider margins, relative to narrower widths, fordeclaring negative margins is unlikely to have a substantialadditional benefıt for long-term local control in BCT.”9

Ductal Carcinoma In SituTwo large randomized trials involving radiation followinglumpectomy for DCIS (NSAPB 17 and NSAPB 24) defınedmargins as either pathologically positive or negative (i.e., notumor touching ink).10 In these trials, low rates of ipsilateralrecurrence have been demonstrated among patients receiv-ing radiation therapy, and the lowest rates of recurrencewereseen in those with negative margins. In another large ran-domized trial of radiation therapy following lumpectomy forDCIS, EORTC 10853, positive margins were defıned as tu-mor 1 mm or smaller from the margin.11 On multivariateanalysis in this trial, tumor-margin status and use of radia-tion were associated with reduced risk of local recurrence.Patientswith a positivemargin had a 10-year local recurrencerate of 39% without radiation and 24% with radiation.Rudloff et al.12 reported a retrospective series of 294 pa-

tients treated for DCIS with lumpectomy with or without ra-diotherapy with a median 11-year follow-up. This seriesdemonstrated higher recurrence rates in patients with nar-row margins. Ten-year recurrence rates were 42%, 27%, and21% for measured margins smaller than 1 mm, 1 mm to9 mm, and10 mm or more, respectively, in patients not re-ceiving radiation therapy. The authors also demonstrated an

KEY POINTS

� There are currently wide variations in the practice ofsurgical-margin management and re-excision ratesfollowing breast-conserving surgery for breast cancer.

� Achieving pathologically negative margins should be theprimary goal of breast-conserving surgery for invasivebreast cancer.

� Low-volume residual disease following lumpectomy iscommon and has no proven effect on clinical outcomes inthe setting of modern adjuvant systemic and radiationtherapies.

� Adjuvant radiation therapy and chemotherapy have beenassociated with reduced rates of ipsilateral breast tumorrecurrence.

� Decisions for surgical re-excision to obtain wider marginsin BCT should be based not solely on measured marginwidth but in the context of tumor biology and othertreatment-related factors.

TABLE 1. Randomized Trials of Mastectomy versusLumpectomy plus Radiation

StudyLumpectomy MarginRequirement

Follow-upTime(Years)

OverallSurvivalDifference

Jacobson, 199532 Gross tumor removal 10 NSˆ

Arriogada, 199633 2 cm excision margin 14 NS

Fisher, 20021 Histologically negative 20 NS

Van Dongen, 200034 Gross tumor removal* 13 NS

Veronesi, 200235 2–3 cm excision, skin,fascia

20 NS

Abbreviation: NS, not significant.*“In general, re-excision took place only when macroscopic (i.e., palpable) disease was leftbehind.”34

MARGINS IN BREAST CONSERVING THERAPY


associated reduction in ipsilateral recurrence in all patientsreceiving adjuvant radiation therapy.A meta-analysis performed by Dunne et al.13 analyzed

4,660 patients undergoing BCT (lumpectomy and adjuvantradiation therapy) for DCIS that demonstrated signifıcantlyhigher recurrence rates in patients with positive marginscomparedwith close or negativemargins (p� 0.01). Further,when considering specifıc margin thresholds in DCIS, theauthors found 2 mm was a threshold. Beyond that there waslittle additional value in reducing ipsilateral recurrence.

THERE ARE CURRENTLY WIDE VARIATIONS INDEFINING “OPTIMAL” MARGINS AND RE-EXCISIONRATES IN BCTDefıning optimal margin in BCT has been an area of contro-versy and debate among breast cancer specialists for manyyears.3,7 There is a current lack of consensus in defıning op-timal margins as demonstrated in several recently reportedsurveys of breast cancer physicians. There are also differencesnoted among radiation oncologists and breast cancer sur-geons, and geographical differences noted between physi-cians in North America and Europe. Taghian et al.4 recentlyreported the results of a survey of 1,137 radiation oncologistsregarding their approach to surgical margins in BCT. Therewere signifıcant variations in how the respondents defıned“negative” and “close”margins. Therewere signifıcant differ-ence in the responses of physicians in North America andEurope. Greater than 50% of respondents considered “nega-tive” margins in BCT as being further than 1 mm from theinked specimen margin. There was also signifıcant variationnoted relative to the recommendation for surgical re-excision of margins.In several studies, signifıcant variations among breast can-

cer surgeons regarding their recommendations for re-excision of margins as part of BCT have been documented.Blair et al. analyzed the results of a survey of 351 breast sur-geons regarding a minimal acceptable margin width withlumpectomy.14 The authors found signifıcant variations inthe recommendations of the minimal acceptable marginwidth for both invasive cancer and DCIS. Sixty-fıve percentof surgeons found 2 mm or wider margin acceptable, al-though 35% viewed amargin width of less than2mm accept-able. Similarly, Azu et al.15 found wide variations in therecommendations for safe margins, and further that subspe-cialization in breast cancer surgery were associated with arecommendation for a smaller acceptable margin. In aggre-gate, these series suggest signifıcant variations in the currentpractice of surgical margin management in BCT for both in-vasive cancer and DCIS.Several more recent studies have documented wide varia-

tions in the re-excision rates following lumpectomy forbreast cancer among hospitals and regions. McCahill et al.6reported that in 2,286 patients undergoing lumpectomy, 23%required additional surgery. In this study, variation in therates of re-excision noted for patients with a negative patho-logic margin ranged from 1.7% to 20.9%. Although 47.9% of

patients with a clear but smaller than 1 mm margin under-went re-excision, only 20.2% of patients with margins be-tween 1.0 and 1.9 mm underwent re-excision. Similarly,Jeevan et al.16 reported wide variations in the use of re-excision following lumpectomy for breast cancer in England.

MICROSCOPIC RESIDUAL DISEASE IN THE BREASTIS COMMON AFTER BCTIn now classic studies, Rosen et al.17 performed simulatedlumpectomies on mastectomy specimens from patients whohad mastectomy for presumed unifocal breast cancer. Inter-estingly, they found residual carcinoma (in situ and invasive)in other quadrants of the breast in a signifıcant number ofpatients. The likelihood of fınding residual disease in otherquadrants of the breast increasedwith increasing primary tu-mor size. Patients with tumors smaller than 2 cmwere foundto have residual disease in other quadrants in 26% of cases,and patientswith tumors larger than 2 cmwere found to haveresidual disease in other quadrants in 38% of cases. Thesestudies suggest that small-volume residual disease is com-mon after lumpectomy. Again, these fındings suggest thatoverly aggressive attempts to remove small volume residualdisease are not warranted and not clinically meaningful par-ticularly when radiation and systemic therapy are planned.Multiple studies have been performed that have analyzed

predictors of residual disease that may be useful in clinicaldecision making regarding the need for re-excision. Youngage, extensive intraductal component, and triple-negativephenotype have all been associated with increased risk of re-sidual disease following lumpectomy for invasive breast can-cer.18 Other studies have demonstrated that tumor size, highgrade, number of involved margins, and linear extent of dis-ease at the margin were all associated with increased risk ofresidual disease at the time of re-excision.19 Factors associ-ated with residual disease following lumpectomy for DCIShave included the presence of comedo necrosis, multifocal-ity, margin width, and lesion size.20 An understanding ofthese predictorsmay be useful in clinical decisionmaking re-garding re-excision to achieve wider surgical margins inBCT.

LIMITATIONS OF PATHOLOGIC ASSESSMENT OFMARGINSThere are practical limitations of pathologic assessment ofmargin assessment thatmust be appreciatedwhen discussing“optimal” surgical margins following lumpectomy and re-lated clinical decision making. There is currently no stan-dardized approach to pathologic assessment of surgicalmargins.21 It is critical to realize that all techniques ofmarginassessment are based on sampling the margin,5 and it is notpractical for a pathologist to sample the entire specimenmar-gin formicroscopic analysis.21 Specimen handling, specimenprocessing, and the technique of margin assessment havealso been shown to influence margin assessment.22 There-



fore, adherence to strict clinical decision algorithms basedentirely on measured margin widths do not seem warrantednor clinically justifıed. It is also imperative to focus on devel-oping standards toward uniform pathologic specimen han-dling and analysis, which will ultimately lead to betterdefıning the relationship between margins and outcomes.

ADJVUANT THERAPIES HAVE A ROLE IN REDUCINGIPSILATERAL BREAST TUMOR RECURRENCESeveral reports have documented a reduction in local recur-rence rates following BCT over time that are likely attribut-able to advances in adjuvant therapy for breast cancer.23Cabioglu et al.23 demonstrated signifıcant reductions in localrecurrence over time that onmultivariate analysis were asso-ciated not only with improving negative surgical-marginrates but also adjuvant hormonal and chemotherapy (p �0.001). Other studies have also demonstrated that an associ-ation between endocrine therapy and chemotherapy admin-istration on reductions in local recurrence in patientsundergoing BCT.24,25 Among HER2� patients, adjuvanttrastuzumab has been associated with a reduction in locore-gional recurrences.26The addition of boost radiation therapy also has a role in

optimizing local control rates in BCT. The EORTC 22881–10882 trial27 demonstrated a signifıcantly lower rate of localrecurrence in patients who received boost irradiation follow-ing adjuvant whole breast irradiation (6.2%) compared withthose receiving whole breast irradiation alone (10.2%).In aggregate, these observations suggest that various adju-

vant treatment factors have an influence on ipsilateral breasttumor recurrence that should be considered in the context ofclinical decision making regarding surgical margin manage-ment. Advances in adjuvant therapy and selection of patientsfor adjuvant therapy are associated with observed improve-ments in local recurrence over time.

EFFECT OF PATIENT CHARACTERISTICS ANDTUMOR BIOLOGY ON LOCAL RECURRENCEAlthough the presence of tumor at the inked margin hasclearly been associated with high rates of local recurrence,both patient and tumor characteristics have also been associ-ated with increased rates of local recurrence in patients un-dergoing BCT. These factors are inherent to the patient andunderlying tumor biology and their influence on outcomesare not mitigated by wider surgical margins.3Young patient age, an extensive intraductal component,

basal phenotype, HER2� cancers, high-grade tumors, andlymphovascular invasion have all been associated withhigher risk of recurrence in patients undergoing BCT.28 Re-cent studies have also identifıed gene-expression profıles thatassociated an increased risk of locoregional recurrence ofbreast cancer in patients undergoing BCT for estrogen-receptor–positive invasive breast cancer.29 Biologic features

including tumor grade, size, and the presence of comedo ne-crosis have also been shown to be important in predictingrecurrence in DCIS.30Recent advances in genomic sequencing of human breast

tumors have categorized four main breast cancer subtypescaused by different subsets of genetic and epigenetic abnor-malities31 Future studies elucidating the attributes of thesesubtypes regarding in-breast growth and recurrence patternshave the potential to influence the acceptable surgicalmargincriteria for each of these subtypes.These fındings suggest that factors beyond measured sur-

gical margin are important determinants of recurrence inbreast cancer and suggest a need to better understand the bi-ology and the relationship of tumor to recurrence and sys-temic treatments in optimizing decision clinical making inBCT.

MANAGING SURGICAL MARGINS FOLLOWING BCTThe available evidence outlined in this report suggests thathaving either invasive cancer or noninvasive cancer at a sur-gical margin is associated with a high rate of local recurrenceand should be avoided in patients undergoing BCT.7 For in-vasive breast cancer, we are in agreement with other authorsthat routine re-excision of margins beyond tumor touchingink when adjuvant radiation therapy and appropriate sys-temic therapy are planned as part of BCT is not necessary.3For patients with DCIS, the data suggest that routine re-excision of margins greater than 2 mm when radiation ther-apy follows lumpectomy is not necessary.3 Consensus onthese issues would likely reduce unnecessary re-excisions andundoubtedly improve cosmesis for patients undergoing BCT.These recommendations must be taken within the context

of individual patients, and surgeons need to appreciate thatmeasured margin width should be but one factor consideredin the decision to perform a re-excision as a component ofBCT.5 Decisions for re-excision beyond the above outlinedparameters may be clinically indicated based on an under-standing of the potential for fınding a signifıcant volume re-sidual disease within the remaining breast, tumor biology,patient characteristics, and plans for the omission of adju-vant therapy.

CHALLENGES FOR THE FUTURE IN MANAGINGSURGICAL MARGINS WITH BCTRegarding surgical margins, challenges for the future are tobetter understand the biologic basis for both tumor growthpatterns and recurrence. Also, it is desirable to better under-stand the interplay between tumor biology, treatments, andoutcomes. Doing so may allow development of more preciseand more personalized algorithms to guide treatment ofpatients with surgical and adjuvant therapy. It is also imper-ative to focus on developing standards toward uniformpathologic-specimen handling and analysis that will be ulti-



mately lead to better defıning the relationship between mar-gins and outcomes. This will streamline clinical decisionmaking and eliminate artifacts that may lead to a misinter-pretation of the true surgical margin. There is also a need toclarify “optimal” margins in patients being treated with par-tial breast irradiation and in those receiving neoadjuvantchemotherapy followed by BCT.

The ultimate goal of these endeavors should be to reducethe rate of re-excisions being performed as part of BCT andthe associated downstream negative sequelae associated withre-excisions. Reducing the amount of nondiseased tissue re-moved at the time of BCT to the optimal level will minimizepatient morbidity while also minimizing tumor recurrenceand maximizing patient outcomes.



References

1. Fisher B, Anderson S, Bryant J, et al. Twenty-year follow-up of a ran-domized trial comparing total mastectomy, lumpectomy, and lumpec-tomy plus irradiation for the treatment of invasive breast cancer.NEnglJ Med. 2002;347:1233-1241.

2. Klimberg VS, Harms S, Korourian S. Assessing margin status. Surg On-col. 1999;8:77-84.

3. Morrow M, Harris JR, Schnitt SJ. Surgical margins in lumpectomy forbreast cancer-bigger is not better. N Engl J Med. 2012;367:79-82.

4. Taghian A, Mohiuddin M, Jagsi R, et al. Current perceptions regardingsurgical margin status after breast-conserving therapy: results of a sur-vey. Ann Surg. 2005;241:629-639.

5. Ananthakrishnan P, Balci FL, Crowe JP. Optimizing surgicalmargins inbreast conservation. Int J Surg Oncol. 2012;2012:585670. Epub 2012 Dec 9.

6. McCahill LE, Single RM,Aiello Bowles EJ, et al. Variability in reexcisionfollowing breast conservation surgery. JAMA. 2012;307:467-475.

7. Singletary SE. Surgical margins in patients with early-stage breast cancertreated with breast conservation therapy.Am J Surg. 2002;184:383-393.

8. Smitt MC, Nowels KW, Zdeblick MJ, et al. The importance of thelumpectomy surgical margin status in long-term results of breast con-servation. Cancer. 1995;76:259-267.

9. Houssami N, Macaskill P, Marinovich ML, et al. Meta-analysis of theimpact of surgical margins on local recurrence in women with early-stage invasive breast cancer treated with breast-conserving therapy. EurJ Cancer. 2010;46:3219-3232.

10. Wapnir IL, Dignam JJ, Fisher B, et al. Long-term outcomes of invasiveipsilateral breast tumor recurrences after lumpectomy in NSABP B-17and B-24 randomized clinical trials for DCIS. J Natl Cancer Inst. 2011;103:478-488.

11. Julien JP, Bijker N, Fentiman IS, et al. Radiotherapy in breast-conserving treatment for ductal carcinoma in situ: fırst results of theEORTC randomised phase III trial 10853. EORTC Breast Cancer Co-operative Group and EORTC Radiotherapy Group. Lancet. 2000;355:528-533.

12. Rudloff U, Brogi E, Reiner AS, et al. The influence of margin width andvolume of disease near margin on benefıt of radiation therapy forwomen with DCIS treated with breast-conserving therapy. Ann Surg.2010;251:583-591.

13. Dunne C, Burke JP, Morrow M, et al. Effect of margin status on localrecurrence after breast conservation and radiation therapy for ductalcarcinoma in situ. J Clin Oncol. 2009;27:1615-1620.

14. Blair SL, Thompson K, Rococco J, et al. Attaining negative margins inbreast-conservation operations: Is there a consensus among breast sur-geons? J Am Coll Surg. 2009;209:608-613.

15. Azu M, Abrahamse P, Katz SJ, et al. What is an adequate margin for

breast-conserving surgery? Surgeon attitudes and correlates. Ann SurgOncol. 2010;17:558-563.

16. Jeevan R, Cromwell DA, Trivella M, et al. Reoperation rates after breastconserving surgery for breast cancer among women in England: retro-spective study of hospital episode statistics. BMJ. 2012;345:e4505.

17. Rosen PP, Fracchia AA, Urban JA, et al. “Residual” mammary carci-noma following simulated partial mastectomy.Cancer. 1975;35:739-747.

18. Wazer DE, Schmidt-Ullrich RK, Ruthazer R, et al. The influence of ageand extensive intraductal component histology upon breast lumpec-tomy margin assessment as a predictor of residual tumor. Int J RadiatOncol Biol Phys. 1999;45:885-891.

19. Cellini C, Hollenbeck ST, Christos P, et al. Factors associated with re-sidual breast cancer after re-excision for close or positive margins. AnnSurg Oncol. 2004;11:915-920.

20. Ratanawichitrasin A, Rybicki LA, Steiger E, et al. Predicting the likeli-hood of residual disease in women treated for ductal carcinoma in situ.J Am Coll Surg. 1999;188:17-21.

21. Morrow M. Breast conservation and negative margins: how much isenough? Breast. 2009;3:S84-86.

22. Graham RA, Homer MJ, Katz J, et al. The pancake phenomenon con-tributes to the inaccuracy of margin assessment in patients with breastcancer. Am J Surg. 2002;184:89-93.

23. Cabioglu N, Hunt KK, Buchholz TA, et al. Improving local control withbreast-conserving therapy: a 27-year single-institution experience.Can-cer. 2005;104:20-29.

24. Fisher B, Dignam J, Bryant J, et al. Five versus more than fıve years oftamoxifen therapy for breast cancer patients with negative lymph nodesand estrogen receptor-positive tumors. J Natl Cancer Inst. 1996;88:1529-1542.

25. Fisher B, Dignam J, Mamounas EP, et al. Sequential methotrexate andfluorouracil for the treatment of node-negative breast cancer patientswith estrogen receptor-negative tumors: eight-year results from Na-tional Surgical Adjuvant Breast and Bowel Project (NSABP) B-13 andfırst report of fındings from NSABP B-19 comparing methotrexate andfluorouracil with conventional cyclophosphamide, methotrexate, andfluorouracil. J Clin Oncol. 1996;14:1982-1992.


27. Bartelink H, Horiot JC, Poortmans PM, et al. Impact of a higherradiation dose on local control and survival in breast-conserving ther-apy of early breast cancer: 10-year results of the randomized boostversus no boost EORTC 22881-10882 trial. J Clin Oncol. 2007;25:3259-3265.



28. Nguyen PL, Taghian AG, Katz MS, et al. Breast cancer subtype approx-imated by estrogen receptor, progesterone receptor, andHER-2 is asso-ciatedwith local and distant recurrence after breast-conserving therapy.J Clin Oncol. 2008;26:2373-2378.

29. Nuyten DS, Kreike B, Hart AA, et al. Predicting a local recurrence afterbreast-conserving therapy by gene expression profıling. Breast CancerRes. 2006;8:R62.

30. Early Breast Cancer Trialists’ CollaborativeGroup, CorreaC,McGale P,et al. Overview of the randomized trials of radiotherapy in ductal carci-noma in situ of the breast. J Natl Cancer Inst Monogr. 2010;2010:162-177.

31. Cancer Genome Atlas Network. Comprehensive molecular portraits ofhuman breast tumours. Nature. 2012;490:61-70.

32. Jacobson JA, Danforth DN, CowanKH, et al. Ten-year results of a com-

parison of conservationwithmastectomy in the treatment of stage I andII breast cancer. N Engl J Med. 1995;332:907-911.

33. Arriagada R, MG, Rochard F, et al. Conservative treatment versus mas-tectomy in early breast cancer: Patterns of failure with 15 years offollow-up data. Institut Gustave-Roussy Breast Cancer Group. J ClinOncol. 1996;14:1558-1564.

34. van Dongen JA, Voogd AC, Fentiman IS, et al. Long-term results of arandomized trial comparing breast-conserving therapy with mastecto-my: European Organization for Research and Treatment of Cancer10801 trial. J Natl Cancer Inst. 2000;92:1143-1150.

35. Veronesi U, Cascinelli N, Mariani L, et al. Twenty-year follow-up of arandomized study comparing breast-conserving surgery with radicalmastectomy for early breast cancer. N Engl J Med. 2002;347:1227-1232.



BREAST CANCER

Pushing the Limits of UpfrontCare and Drug Development:Neoadjuvant Opportunities inBreast Cancer

CHAIRAngela DeMichele, MD, MSCEAbramson Cancer Center of the University of PennsylvaniaPhiladelphia, PA

SPEAKERSTatiana Prowell, MDU.S. Food and Drug AdministrationSilver Spring, MD

Charles E. Geyer Jr., MDStatewide Clinical Trials Network of TexasAddison, TX

Drug Development: Neoadjuvant Opportunities in Breast CancerPriya Rastogi, MD, Charles E. Geyer Jr., MD, Eleftherios P. Mamounas, MD, and Angela DeMichele, MD, MSCE

OVERVIEW

Preoperative therapy allows for a higher rate of breast conserving surgery and has been shown equivalent to adjuvant therapy.Preoperative therapy provides an opportunity to obtain insights into breast cancer biology and to accelerate the evaluation of newtherapies. Clinical trials have shown that women who achieve a pathologic complete response (pCR) have substantially improvedoutcomes compared with those who do not achieve a pCR. The U.S. Food and Drug Administration (FDA) meta-analysis demonstratedthat the association of pCR and long-term outcomes is greater in women with aggressive breast cancer subtypes. In patients withHER2� breast cancer, the addition of trastuzumab to chemotherapy in the neoadjuvant setting has doubled pCR and correlated withimproved outcomes. Clinical trials will evaluate tailoring the use of radiation therapy in patients who have received neoadjuvanttherapy. Trials have established neoadjuvant endocrine therapy as a valid treatment and research option for ER-rich breast cancer. Theneoadjuvant setting allows for evaluation of endocrine therapies in combination with newer targeted therapies in the appropriatepatient populations. The neoadjuvant setting provides opportunity to accelerate the evaluation of new agents, improve pCR rates, andidentify predictive biomarkers for response. This setting provides the opportunity for screening new agents in combination withchemotherapy while obtaining serial biopsies to understand biology of response and resistance. Although current standard therapiesprovide substantial benefits for patients with a pCR, patients with residual disease are at substantial risk for disease recurrence. Newagents are being evaluated in patients with high-risk residual disease following standard treatment regimens.

Preoperative or neoadjuvant therapy, initially used for lo-cally advanced breast cancers, has become more com-

monly used for patients with operable disease, particularlythose with larger primary tumors. By reducing the size of theprimary tumor, preoperative therapy allows a higher rate ofbreast conserving surgery.1-10 The neoadjuvant setting alsoprovides opportunities to obtain insights into breast cancerbiology and accelerate evaluation of new therapies.

CLINICAL RATIONALE FOR PREOPERATIVE(NEOADJUVANT) THERAPYHortobagyi and colleagues demonstrated in the 1980s thatwomen with locally advanced breast cancer who achieved apathologic complete response (pCR) after neoadjuvant che-motherapy (NACT) had substantially improved outcomescompared with historic controls.11 Use of NACT in operablebreast cancer was evaluated in the National Surgical Adju-vant Breast and Bowel Project (NSABP) Protocol B-18 clini-cal trial in which 1,523 women with operable breast cancerwere randomized to four cycles of AC (doxorubicin 60mg/m2,cyclophosphamide 600mg/m2) every 21 days either before orafter defınitive surgery.5-7,10 A pCR in the breast was docu-mented in 13% of patients, who had substantially improved

outcomes relative to the larger group without pCR (disease-free survival (DFS) hazard ratio [HR] � 0.47, p � 0.0001;overall survival (OS) HR � 0.32, p � 0.0001). Pathologicallynegative nodes were identifıed in 58% of women receivingthe neoadjuvant therapy compared with 42% among womenundergoing initial surgery (p � 0.0001). A higher rate ofbreast conserving surgery was also observed in women re-ceiving preoperative therapy (68% vs. 60%, respectively; p �0.001). There were no statistically signifıcant differences inDFS (HR � 0.93, p � 0.27) and OS (HR � 0.99, p � 0.90)between the two groups. This trial demonstrated that NACTwas safe in operable breast cancer and that pCR status wasassociated with a favorable outcome.

PATHOLOGIC RESPONSE AS A SURROGATE FORCLINICAL OUTCOMEThe NSABP B-27 trial evaluated the incorporation ofdocetaxel into the neoadjuvant setting.8-10 Women with op-erable breast cancer were randomized to receive (1) preoper-ative AC followed by surgery, (2) preoperative AC followedby four cycles of preoperative docetaxel followed by surgery,or (3) preoperative AC followed by surgery and then four cy-cles of postoperative docetaxel. Preoperative docetaxel fol-

From the University of Pittsburgh Cancer Institute, Pittsburgh, PA; University of Texas Southwestern Medical Center, Dallas, TX; MD Anderson Cancer Center Orlando, Orlando, FL; University ofPennsylvania, Philadelphia, PA.


Corresponding author: Priya Rastogi, MD, National Surgical Adjuvant Breast and Bowel Project (NSABP), East Commons Professional Building, Four Allegheny Center - 5th Floor, Pittsburgh, PA15212; email: [email protected].


NEOADJUVANT THERAPY IN BREAST CANCER


lowing AC increased the pCR rate (13% to 26%, p � 0.001),and the proportion of women with negative nodes (51% to58%, p � 0.001) relative to AC alone. Although this largergroup of women had a more favorable outcome (DFS HR �0.49, p � 0.0001 and OS HR � 0.36, p � 0.0001), there wereno statistically signifıcant differences in DFS andOS demon-strated among the three overall treatment groups (p across allarms � 0.76).Recently an international working group, in collaboration

with the FDA, conducted a large meta-analysis of 12,993 pa-tients treated on 12 early, randomized NACT trials withavailable long-term follow-up for DFS and OS. Improvedoutcomes in patients with pCR compared with those withoutwere again demonstrated (event-free survivalHR� 0.48, p�0.001 and OS HR � 0.36, p � 0.001). Moreover, the magni-tude of the association between pCR and event-free survivalwas greater in patients with aggressive breast cancer subtypes(hormone receptor (HR)�/HER2-HR� 0.49, p� 0.001 andHR�/HER2� HR � 0.58, p � 0.001vs. HR-/HER2� HR �0.25, p� 0.001 andHR-/HER2-HR� 0.24, p� 0.001).How-ever, correlation of improvement in long-term outcomeswith increased pCR rates could not be established. The au-thors suggested this may be due to relatively low pCR rates inthese early trials, the heterogeneity of the populations ac-crued, and the lack of targeted therapy in the majority of pa-tients.12The German Breast Group (GBG) and Arbeitsgemein-

schaft Gynakologische Onkologie-Breast Group (AGO-B)conducted a pooled analysis of seven randomized NACTsconducted by the groups, involving 4,193 patients, to deter-mine whether pCR predicts improved outcomes across theintrinsic subtypes. Estrogen receptor (ER), progesterone re-ceptor (PR), HER2 and histologic grade determined by localpathologists were used to classify patients by intrinsic sub-types according to clinicopathologic criteria recently recom-mended by the St. Gallen panelists, with the exception thatgrade was substituted for Ki-67.13 In the luminal B/HER2-,ER-/HER2� and triple negative groups, pCR was associated

with improved outcomes.However in luminal A and luminalB/HER2� patient groups, no correlations were identifıed,suggesting that inclusion of these “low-risk” patients mayhave had a substantial adverse dilutional effect on the abilityto demonstrate improved outcomes among the women withhighly proliferative tumors with pCR. Analysis of the I-SPYTRIAL 1 (Investigation of Serial Studies to Predict YourTherapeutic Response with Imaging andMolecular Analysis1) correlated pCR, residual cancer burden (RCB), intrinsicsubtypes, and three-year RFS; the trial demonstrated similarfındings in low-risk subsets defıned by a variety of predictiveprofıles. Low-risk subsets had low pCR rates but favorableoutcomes regardless of the degree of response.14The hypothesis that the lack of correlation between pCR

rates and long-term outcome could be on the basis of inclu-sion of low-risk patients is further supported by results of theNOAH (NeoAdjuvant Herceptin) trial.15 In this study, 235women with HER2� locally advanced or inflammatorybreast cancer were treated with NACT with or without tras-tuzumab. For women receiving chemotherapy and trastu-zumab, the pCR rate was 38%, compared with 19% forwomen receiving chemotherapy alone. The doubling of pCRin this more homogeneous population with high prolifera-tion rates correlated with improved three-year event-freesurvival (71% vs. 56%; HR � 0.59, p � 0.013). In aggregate,these fındings support the contention that pCR is ameaning-ful short-term surrogate of outcome in more aggressive tu-mors (particularly HER2� and triple negative tumors) andprovide a strong rationale for testing new agents targetinghighly proliferative tumors in the neoadjuvant setting usingpCR as a primary end point.

TAILORING LOCOREGIONAL THERAPY WITHNEOADJUVANT CHEMOTHERAPYResponse to NACT allows for less extensive surgery to thebreast andmayprovide similar benefıts in tailoring the extentof surgical resection in the axilla (with the use of sentinelnode biopsy), as well as decreasing the need and extent ofpostoperative radiotherapy in patientswho achieve pCR.Theutility of downstaging of axillary nodes with NACT has beendemonstrated in several studies that have evaluated the roleof node status in prognosis after neoadjuvant therapy. His-torically, defınitions of pCRvary across studies, ranging fromthemost stringent, ypT0 ypN0 (absence of invasive cancer inthe breast and axillary nodes and absence of DCIS), to ypT0/Tis ypN0 (absence of invasive cancer in the breast and axil-lary nodes; DCIS allowed), to the most liberal, ypT0/Tis(absence of invasive cancer in the breast and DCIS allowed;regardless of nodal involvement). The pooled analysis ofseven randomized neoadjuvant trials conducted byGBG andAGO found that patients with no invasive or noninvasivedisease in the breast and negative axillary nodes had signifı-cantly better DFS than the group with either residual nonin-vasive or microinvasive breast disease or residual positivenodes.13 Hazard ratios for DFS comparing patients with orwithout pCRwere lowest when defıned as no invasive and no

KEY POINTS

� Preoperative therapy provides an opportunity to obtaininsights into breast cancer biology and accelerate theevaluation of new therapies.

� The neoadjuvant setting provides an opportunity toevaluate new agents with chemotherapy, improve pCRrates, and identify predictive biomarkers for response.

� Clinical trials will evaluate tailoring the use of radiationtherapy in patients who have received neoadjuvanttherapy.

� The neoadjuvant setting allows for evaluation of endocrinetherapies in combination with newer targeted therapies inthe appropriate patient populations.

� New agents are being evaluated in patients with high-riskresidual disease following standard treatment regimens.

RASTOGI ET AL


in situ residuals (0.446) and increased monotonously whenin situ residuals (0.523), no invasive breast residuals but in-volved nodes (0.623), and focal-invasive disease (0.727) wereincluded in the defınition. However the FDA-led meta-analysis, which included the German studies, demonstratedno difference in the hazard of recurrence or death regardlessof presence or absence of residual DCIS in the breast.12 Thesedata suggest that clearing the nodes has prognostic signifı-cance, and therefore accurate assessment of the residualnodal burden is necessary.An additional advantage of rendering the axillary nodes

free of disease with NACT is the potential reduction in sur-gicalmorbidity and late lymphedema from the use of sentinelnode evaluation. However, this advantage is predicated onthe accuracy of sentinel node biopsy following NACT.Because NACT can have an adverse effect on the integrityof axillary nodal tissue (with or without eradication of dis-ease), the accuracy of sentinel lymph node evaluation inthe patient who has undergone NACT must be establishedindependently of its validation in patients who arechemotherapy-naive. Several retrospective and prospective,single-institution and multicenter trials, as well as two meta-analyses, have shown that sentinel node biopsy after NACThas lower identifıcation rates but similar false-negative ratesas sentinel node biopsy before NACT (about10%).16-18 It isimportant to emphasize that the available data on the perfor-mance of sentinel node biopsy after NACT are only applica-ble to patients who present with operable breast cancers(T1–3N0–1) and should not be extrapolated to thosewhopres-ent with locally advanced disease (T4, N2). Similarly, cautionis required for patients who present with documented tumorinvolvement of the axilla (by fıne needle aspiration or corebiopsy) and undergo NACT. Two prospective trials pre-sented at the 2012 San Antonio Breast Cancer Symposiumdemonstrated false-negative rates of 12.6% and 14.2%, re-spectively, slightly higher than in the upfront setting.19,20Moreover, the number of resected sentinel nodes inverselycorrelatedwith the false-negative rate, and the data suggestedthat the removal of at least two sentinel nodes in this setting isimportant to keep the accuracy of sentinel node biopsy sim-ilar to that in the upfront setting.A pCR to NACT may identify subsets of patients with

lower risk for locoregional recurrences (LRR) that may notjustify regional chest wall and/or regional nodal radiation.This approach is predicated on the demonstration thatdownstaging substantially lowers the rates of locoregional re-currence. TheNSABPB-18 andB-27 trials provide an impor-tant data set to address these questions. In both trials,regional nodal radiation after lumpectomy or chest wall andregional nodal radiation after mastectomy were not permit-ted per protocol, thus avoiding the confounding effects of se-lective use of radiation in patients with positive nodes orother high-risk features. Data from a combined analysis ofthese two trials were recently published.21 The 10-year cumu-lative incidence of LRR was 12.3% for patients treated withmastectomy (8.9% local; 3.4% regional) and 10.3% for thosetreated with lumpectomy plus breast radiation (8.1% local;

2.2% regional). Independent predictors of LRR in patientstreated with lumpectomy were age, clinical node status (be-fore NACT), and pathologic node status/pathologic breasttumor response. In patients treated with mastectomy, inde-pendent predictors of LRR were clinical tumor size (beforeNACT), clinical node status (before NACT), and pathologicnode status/pathologic breast tumor response. By using theseindependent predictors, risk of LRR can be tailored, and theneed for postoperative radiotherapy could be individualized.On the basis of these fındings, a prospective, randomized

phase III clinical trial led by NSABP and RTOG (NSABPB-51/RTOG1304) has been initiated to validate the approachof tailoring the use of radiation for patients who present withdocumented involvement of axillary nodes beforeNACT butare found to have histologically negative nodes afterward. Al-ternatively for patients who present with documented in-volvement of axillary nodes beforeNACTand are found to bepathologically node-positive based on sentinel node biopsyfollowingNACT, the Alliance for Clinical Trials inOncologyis conducting a phase III trial (A011202) randomizing pa-tients to completion axillary dissection plus regional RTcompared with regional RT alone.

DEVELOPING NEW AGENTS IN THE NEOADJUVANTSETTINGSequential taxane/anthracycline-based three-drug regimenshave become standard for neoadjuvant therapy of patientswho present with stage II or III breast cancer and are candi-dates for chemotherapy. Trastuzumab is incorporated intothese regimens in HER2� disease. These regimens result inpCR rates of 20% to 50% in patient populationswith subtypesof breast cancer associated with high proliferative rates. Al-though current standard therapies provide substantial bene-fıts for patients with a pCR at surgery, patients with residualdisease are at substantial risk for disease recurrence, with themajority of patients with HER2� or triple negative breastcancers who do not achieve pCR developing distant meta-static disease within the fırst three years after therapy.14 At-tempts to improve pCR rates with additional standardchemotherapeutic agents have not been successful.22 There-fore efforts to improve pCR rates and develop alternativetherapies forwomenwith residual disease following standardtherapy are clearly needed.Insights into the heterogeneity and biology of breast cancer

obtained through gene-sequencing and other technologieshave identifıed a large number of potential targets in thesetumors for newer, more selective investigational agents (so-called targeted therapies). Toxicity of many of these agentsappears to be substantially less problematic than with morenonselective agents, and combination with chemotherapyappears to be feasible. The neoadjuvant setting provides anopportunity to accelerate evaluation of these agents andidentify predictive biomarkers for response, as well as to gaininsights into mechanisms of resistance. Other interventions,including immunotherapy and complementary approaches



in the neoadjuvant setting, are also being used or are underconsideration.

SCREENING TRIALSBecause palpable or locally advanced breast cancer is readilyaccessible for serial biopsies and imaging assessment, theneoadjuvant setting provides opportunity for screening newagents in combination with chemotherapy, while also ob-taining serial biopsies to understand biology of response andresistance as well as to develop early predictors of response.The prototype for this important step in drug development isthe I-SPY TRIAL 2.23 This ambitious, multicenter, random-ized phase II effort among academic cancer centers across theUnited States employs a standard regimen of weekly pacli-taxel for 12 weeks followed by AC for four cycles. Trastu-zumab is administered with paclitaxel for HER2� breastcancer in the control arm and in several of the treatmentarms. Potential candidates undergo biopsy for screening,which includes the 70-gene Mammaprint (MP) along withER, PR, and HER2 assessment. Women with low-risk MPscore and ER�/HER2 negative disease are excluded from en-try and randomization.Several agents (up to eight) are under evaluation at any

time point, with assignment to control therapy or investiga-tional therapy on the basis of results of the molecular profıl-ing and an adaptive randomization design. The primaryobjective is to determine whether adding experimentalagents to standard NACT (with or without trastuzumab) in-creases the probability of pCR over standard NACT for eachbiomarker signature established at trial entry, and to deter-mine for each experimental agent used the predictive proba-bility of success in a subsequent phase III trial for eachpossible biomarker signature. A data safety monitoringboard (DSMB) meets monthly to assess toxicity and reviewperformance of the adaptive randomization, given the poten-tial curability of patients receiving investigational therapy. Abiopsy for pharmacodynamicmarkers is obtained after threeweeks of therapy, and serial imaging is performed through-out the trial using automated dynamic contrast enhancedMRI. As information accumulates on each agent, it is either“graduated” or “dropped for futility,” and new agents enterthe trial. The basic design of the study remains unchanged sothat additional agents and potential biomarkers can be morerapidly and effıciently evaluated. Although the I-SPY TRIAL2 will not provide defınitive information on a new investiga-tional agent, it is hoped that promising agents will be quicklyidentifıed andmoved into phase III trials with amuch higherprobability for success than has been seen in the past.

NEOADJUVANT TRIALS POWERED FOR DISEASE-FREE AND OVERALL SURVIVALIn the potentially curative situation, it is likely that regulatoryagencies will continue to require trials to conclusively dem-onstrate improvement in DFS and perhaps OS for approval

of new agents in this setting. The FDA released a draft guid-ance in May 201224,25 stating that the agency would considergranting accelerated approval on the basis of a surrogate endpoint that is likely to predict clinical benefıt, and it proposedpCR as such an end point in the neoadjuvant setting for high-risk early stage breast cancer. Confırmation of clinical benefıtwith an improvement inDFS orOSwould still be required orthe indication could be withdrawn if confırmatory trials donot show a benefıt.A potential approach would be to use a two-step design in

which only neoadjuvant patients were entered in the fırststage to determine whether a prespecifıed improvement inpCR could be demonstrated. If so, the second stage wouldallow entry of patients in both the neoadjuvant and adjuvantsetting to achieve the sample size needed for determiningDFS and OS. Interim analyses of pCR rates could be used todetermine whether an accrual hiatus between the two stageswas necessary. Accrual to this type of trial could be readilyaccomplished in cooperative group and community settings.

POSTNEOADJUVANT RESIDUAL DISEASE TRIALSBecause patients with pCR following standard adjuvant ther-apies have a low risk for recurrence, itmay not be appropriateto evaluate newer agents with moderate toxicities or whichdo not combine well with standard therapies in these pa-tients. In this situation an alternative development approachwould be to study new agents in patients with high-risk re-sidual disease following standard treatment regimens.The NSABP, GBG, and other investigators are collaborat-

ing on a global trial (KATHERINE NCT 01772474) evaluat-ing T-DM1 as an alternative to continuation of trastuzumabin women with residual disease following neoadjuvant che-motherapy combined with HER2 targeted therapy that in-cludes trastuzumab. Patients with residual disease followingNACT and trastuzumab have a three-year RFS of only 70%,26so they are appropriate candidates for evaluating promisingnew therapeutic agents such as T-DM1. If successful this trialwould establish a new approach for drug development.Two innovative trials are in progress to address residual

disease in patients with triple-negative breast cancer. TheHoosier Oncology Group is currently evaluating the benefıtof cisplatin with or without a PARP inhibitor, rucaparib(Clovis) in women with triple negative or ER�/BRCA-mutant breast cancer who have residual disease in the breastafter neoadjuvant chemotherapy with an anthracyclineand/or taxane-based regimen (NCT01074970).27 The pri-mary end point of this trial is two-yearDFS. TheABCDE trial(Adjuvant Bevacizumab, Metronomic Chemotherapy Dietand Exercise)28 being conducted within the TranslationalBreast Cancer Research Consortium (TBCRC) enrolls pa-tients with triple negative or high-stage ER� disease whohave received anthracycline and/or taxane-containing neo-adjuvant chemotherapy and have residual disease at surgery.In this 2 x 2 design, patients are initially randomly assigned toeither “observation” (standard of care, including endocrinetherapy), or “treatment” (six months of bevacizumab with

RASTOGI ET AL


metronomic chemotherapy, followed by an additional 18months of intermittent bevacizumab alone). Patients are ad-ditionally randomized to one of two telephone-based lifestyleinterventions: diet alone or diet plus exercise.

NEOADJUVANT ENDOCRINE THERAPY TRIALSDevelopment of multigene expression profıles to predict thedegree of benefıt from specifıc therapeutic agents has beenchallenging; to date, a profıle for defıning benefıt from spe-cifıc agents has not been established. However the OncotypeDX Recurrence Score assay derived from an algorithm basedon expression level of 16 breast cancer-related genes and fıvereference genes has been shown to consistently identify largesubsets of patients with early stage, ER� breast cancer whohave a low-risk for recurrence with endocrine therapyalone.29 The assay also has apparent utility for defıning po-tential benefıt from the addition of chemotherapy in node-negative ER� breast cancers.30 The utility of the assay in thispopulation will be better defıned by results of the TAILORxstudy, which has completed accrual and is in follow-up.31SWOG is leading an effort (RxSPONDER) to evaluate theutility of the Recurrence Score for defıning chemotherapybenefıt in women with node-positive, ER� breast cancers aswell.32 Studies evaluating the Recurrence Score as well asother predictive assays have shown many women with earlystage, ER� breast cancer derive little or no benefıt from theaddition of chemotherapy to endocrine therapy. In view ofthese fındings, it is not surprising that pCR rates in responseto chemotherapy are low in these populations and that out-come is not associated with pCR.Given these observations, the neoadjuvant setting is ideal

for evaluating endocrine therapies in combination withnewer targeted therapies in appropriate patient populations.Ellis and colleagues conducted a randomized phase II studyof four months of neoadjuvant endocrine therapy with ta-moxifen vs. letrozole and evaluated post-therapy pathologyfındings to derive a response-based preoperative endocrineprognostic index (PEPI) to defıne a subset of patients who dowell with endocrine therapy alone.33 The index was validatedin an independent cohort of patients from another neoadju-vant endocrine study, the IMPACT trial.34 The authors con-cluded that “patients with breast cancerwith pathologic stageI or 0 disease after neoadjuvant endocrine therapy and a low-risk biomarker profıle in the surgical specimen (PEPI score 0)have an extremely low risk of relapse and are therefore un-likely to benefıt from adjuvant chemotherapy.“ACOSOG conducted a follow-up randomized phase II

neoadjuvant trial (Z1031) comparing the three approvedaromatase inhibitors in postmenopausal women with ER–rich stage II to III breast cancers. They demonstrated an over-all clinical response rate of 63% and that breast-conservingsurgery could be performed in 50%of patientswhopresentedwith disease that would have requiredmastectomy at presen-tation. They also demonstrated that a favorable PEPI score

was more common in luminal A than luminal B tumors(27.1% v 10.7%; P � 0.004).35The ALTERNATE trial (ACOSOGZ11103) will be a phase

III study comparing anastrozole vs. fulvestrant vs. the com-bination as neoadjuvant endocrine therapy in postmeno-pausal women with ER-rich stage II and III breast cancer.The study will also prospectively evaluate the PEPI score as apredictive factor for favorable outcome with endocrine ther-apy alone. This innovative series of trials has established neo-adjuvant endocrine therapy as a valid treatment and researchoption for ER-rich early breast cancer and will likely be fur-ther exploited in the future for combination endocrine/tar-geted therapy trials.

CONCLUSIONNeoadjuvant therapy for breast cancer has an established rolein the management of women with inoperable disease andfor those in whom breast conservation is a goal of therapy.Early randomized trials established pCR as a surrogate forsurvival outcomes, and more recent trials have refıned ourunderstanding of this relationship for different molecularsubtypes of disease. Clinical management to minimize LRRcontinues to evolve, with recent studies evaluating the appli-cability of sentinel node evaluation after NACT, the relativecontribution of nodal response to predicting outcome, andongoing trials designed to optimize radiotherapy.From a research standpoint, the neoadjuvant setting pro-

vides several important opportunities for translational drugdevelopment. Phase II randomized “screening studies” toimprove pCR rates in high-risk patients with the addition oftargeted therapies to standard chemotherapy are ongoingand will contribute important information on both drug ac-tivity and biomarkers identifying those patients most likelyto respond. Such studies will contribute to the pipeline ofagents for testing in phase III trials, with the ability to poten-tially accelerate approval of successful therapies by using thenew FDA guidance that provides a roadmap for the use ofpCR as a surrogate end point and smaller, more focused trialdesigns in specifıc biologically-defıned subgroups to assesssurvival. Molecular and genetic profıling of residual diseaseafter neoadjuvant therapy is an active avenue of investigationinto potential targets of resistant disease, and trials of agentsagainst these targets have the potential to improve outcomefor patients with poor prognosis after standard therapy. Therecognition that ER� low-proliferation tumors are often in-trinsically unresponsive to chemotherapy (and thus showlow pCR rates) has led to a parallel pathway of investigationinto neoadjuvant endocrine therapy, in which Ki-67 is beingvalidated as a surrogate for response and outcome. Taken to-gether, the advances brought about by neoadjuvant therapyin both standard and investigational settings have ushered ina more personalized approach to breast cancer therapy thatcould result in a reduced burden of toxicities and recurrentdisease and in improved survival.





Employment or Leadership Position: None. Consultant or Advisory Role: Angela DeMichele, Millennium; Pfizer. Eleftherios P. Mamounas, Celgene;Genentech/Roche; Genomic Health, Inc. Stock Ownership: None. Honoraria: Eleftherios P. Mamounas, Genomic Health, Inc. Research Funding: AngelaDeMichele, Genentech; GlaxoSmithKline; Incyte; Millennium; Pfizer; Wyeth. Expert Testimony: None. Other Remuneration: None.

References

1. Hortobagyi GN, Blumenschein GR, Spanos W, et al. Multimodal treat-ment of locoregionally advanced breast cancer. Cancer. 1983;51:763-768.

2. Hortobagyi GN, Ames FC, Buzdar AU, et al. Management of stage IIIprimary breast cancer with primary chemotherapy, surgery, and radia-tion therapy. Cancer. 1988;62:2507-2516.

3. BonadonnaG, Veronesi U, Brambilla C, et al. Primary chemotherapy toavoid mastectomy in tumors with diameters of three centimeters ormore. J Natl Cancer Inst. 1990;82:1539-1545.

4. BonadonnaG,Valagussa P, BrambillaC, et al. Primary chemotherapy inoperable breast cancer: eight-year experience at theMilan Cancer Insti-tute. J Clin Oncol. 1998;16:93-100.

5. Fisher B, Brown A, Mamounas E, et al. Effect of preoperative chemo-therapy on local-regional disease inwomenwith operable breast cancer:fındings from National Surgical Adjuvant Breast and Bowel ProjectB-18. J Clin Oncol. 1997;15:2483-2493.

6. Fisher B, Bryant J, Wolmark N, et al. Effect of preoperative chemother-apy on the outcome of womenwith operable breast cancer. J Clin Oncol.1998;16:2672-2685.

7. WolmarkN,Wang J,Mamounas E, Bryant J, Fisher B. Preoperative che-motherapy in patients with operable breast cancer: nine-year resultsfrom National Surgical Adjuvant Breast and Bowel Project B-18. J NatlCancer Inst Monogr. 2001;30:96-102.

8. Bear H, Anderson S, Brown A, et al. The effect of tumor response ofadding sequential preoperative docetaxel to preoperative doxorubicinand cyclophosphamide: preliminary results fromNational Surgical Ad-juvant Breast and Bowel Project Protocol B-27. J Clin Oncol. 2003;21:4165-4174.

9. Bear HD, Anderson S, Smith RE, et al. Sequential preoperative or post-operative docetaxel added to preoperative doxorubicin plus cyclophos-phamide for operable breast cancer: National Surgical Adjuvant Breastand Bowel Project Protocol B-27. J Clin Oncol. 2006;24:2019-2027.

10. Rastogi P, Anderson SJ, Bear HD, et al. Preoperative chemotherapy: up-dates of National Surgical Adjuvant Breast and Bowel Project ProtocolsB-18 and B-27. J Clin Oncol. 2008;26:778-785.

11. FeldmanLD,HortobagyiGN,BuzdarAU,Ames FC, BlumenscheinGR.Pathological assessment of response to induction chemotherapy inbreast cancer. Cancer Res. 1986;46:2578-2581.

12. Cortazar P, Zhang L, UntchM, et al. Meta-analysis results from the col-laborative trials in neoadjuvant breast cancer (CTNeoBC). Presented atSan Antonio Breast Cancer SymposiumConference; December 2012;San Antonio, TX. Abstract S1-11.

13. von Minckwitz G, Untch M, Blohmer J, et al. Defınition and impact ofpathologic complete response on prognosis after neoadjuvant chemo-therapy in various intrinsic breast cancer subtypes. J Clin Oncol. 2012;30:1796-1804.

14. Esserman LJ, Berry DA, DeMichele A, et al. Pathologic complete re-sponse predicts recurrence-free survival more effectively by cancer

subset: results from the I-SPY 1 TRIAL-CALGB 150007/150012,ACRIN 6657. J Clin Oncol. 2012;30:3242-3249.

15. Gianni L, EiermannW, SemiglazovV, et al. Neoadjuvant chemotherapywith trastuzumab followed by adjuvant trastuzumab versus neoadju-vant chemotherapy alone, in patients with HER2� locally advancedbreast cancer (the NOAH trial): a randomized controlled superioritytrial with a parallel HER2-negative cohort. Lancet. 2010;375:377-384.

16. Mamounas EP, Brown A, Anderson S, et al. Sentinel node biopsy afterneoadjuvant chemotherapy in breast cancer: results fromNational Sur-gical Adjuvant Breast and Bowel Project Protocol B-27. J Clin Oncol.2005;23:2694-2702.

17. Kelly AM, Dwamena B, Cronin P, Carlos RC. Breast cancer sentinelnode identifıcation and classifıcation after neoadjuvant chemotherapy-systematic review and meta analysis. Acad Radiol. 2009;16:551-563.

18. Xing Y, Ding D, Cox D, et al. Meta-analysis of sentinel lymph node bi-opsy following preoperative chemotherapy in patients with operablebreast cancer. Proc Am Soc Clin Oncol. 2004;23:17. Abstr 561.

19. Boughey JC, SumanVJ,Mittendorf EA, et al. The role of sentinel lymphnode surgery in patients presenting with node positive breast cancer(T0-T4, N1-2) who receive neoadjuvant chemotherapy - results fromthe ACOSOG Z1071 trial. Cancer Res. 2012;72(24 suppl): Abstr S2-1.

20. Kuehn T, Bauerfeind IGP, Fehm T, et al. Sentinel lymph node biopsybefore or after neoadjuvant chemotherapy-fınal results from the Pro-spective German, multiinstitutional SENTINA-Trial. Cancer Res. 2012;72(24 Suppl.):Abstr S2-2.

21. Mamounas EP,Anderson SJ, Dignam JJ, et al. Predictors of locoregionalrecurrence after neoadjuvant chemotherapy: results from combinedanalysis of National Surgical Adjuvant Breast and Bowel Project B-18and B-27. J Clin Oncol. 2012;30:3960-3966.

22. Bear HD, Tang G, Rastogi P, et al. Bevacizumab added to neoadjuvantchemotherapy for breast cancer. N Engl J Med. 2012;366:310-320.

23. U.S. National Institutes of Health. I-SPY 2 Trial: neoadjuvant and Per-sonalized Adaptive Novel Agents to Treat Breast Cancer. http://clinicaltrials.gov/ct2/show/study/NCT01042379. Accessed February 11, 2013.

24. Prowell TM, Pazdur R. Pathological complete response and accelerateddrug approval in early breast cancer.NEngl JMed. 2012;366:2438-2441.

25. United States Food and Drug Administration. Guidance for Industry.PathologicComplete Response inNeoadjuvant Treatment ofHigh-RiskEarly-Stage Breast Cancer: Use as an Endpoint to Support AcceleratedApproval. http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/UCM305501.pdf. Published May2012. Accessed February 13, 2013.

26. Untch M, Fasching PA, Konecny GE, et al. Pathologic complete re-sponse after neoadjuvant chemotherapy plus trastuzumab predicts fa-vorable survival in human epidermal growth factor receptor2-overexpressing breast cancer: results from the TECHNO trial of theAGO and GBG study groups. J Clin Oncol. 2011;29:3351-3357.

27. U.S. National Institutes of Health. PARP Inhibition for Triple Negative

RASTOGI ET AL


Breast Cancer (ER-/PR-HER2-) With BRCA 1/2 Mutations. http://clinicaltrials.gov/show/NCT01074970. Accessed February 13, 2013.

28. U.S. National Institutes of Health. Bevacizumab, Metronomic Chemo-therapy (CM), Diet and Exercise After Preoperative Chemotherapyfor Breast Cancer (ABCDE). http://clinicaltrials.gov/ct2/show/NCT00925652?term�ABCDE�Trial&rank�1. Accessed February 13, 2013.



31. U.S. National Institutes of Health. Hormone therapy with or withoutcombination chemotherapy in treating women who have undergonesurgery for node-negative breast cancer (the TAILORx trial). http://clinicaltrials.gov/ct2/show/NCT00310180?term�tailorx&rank�1. Ac-cessed February 13, 2013.

32. U.S. National Institutes of Health. A Phase III randomized clinical trialof standard adjuvant endocrine therapy �/-chemotherapy in patients

with 1-3 positive nodes, hormone receptor-positive andHER2-negativebreast cancer with recurrence score (RS) of 25 or less. http://clinicaltrials.gov/ct2/show/NCT01272037?term�S1007&rank�1. Accessed Febru-ary 13, 2013.

33. Ellis MJ, Tao Y, Luo J, et al. Outcome prediction for estrogen receptor-positive breast cancer based on postneoadjuvant endocrine therapy tu-mor characteristics. J Natl Cancer Inst. 2008;100:1380-1388.

34. Smith IE, Dowsett M, Ebbs SR, et al. Neoadjuvant treatment of post-menopausal breast cancer with anastrozole, tamoxifen, or both in com-bination: the Immediate Preoperative Anastrozole, Tamoxifen, orCombined with Tamoxifen (IMPACT) multicenter double-blind ran-domized trial. J Clin Oncol. 2005;23:5108-5116.

35. Ellis MJ, Suman VJ, Hoog J, et al. Randomized phase II neoadjuvantcomparison between letrozole, anastrozole, and exemestane for post-menopausal women with estrogen receptor-rich stage 2 to 3 breast can-cer: clinical and biomarker outcomes and predictive value of thebaseline PAM50-based intrinsic subtype-ACOSOGZ1031. J ClinOncol.2011;29:2342-2349.



BREAST CANCER

Surveillance and Monitoring inBreast Cancer Survivors:Maximizing Benefit and MinimizingHarm

CHAIRMaxine S. Jochelson, MDMemorial Sloan-Kettering Cancer CenterNew York, NY

SPEAKERSDaniel F. Hayes, MDUniversity of Michigan Medical CenterAnn Arbor, MI

Patricia A. Ganz, MDUniversity of California, Los Angeles Schools of Medicine and Public HealthLos Angeles, CA

Surveillance and Monitoring in Breast Cancer Survivors:Maximizing Benefit and Minimizing HarmMaxine Jochelson, MD, Daniel F. Hayes, MD, and Patricia A. Ganz, MD

OVERVIEW

Although the incidence of breast cancer has increased, breast cancer mortality has decreased, likely as a result of both breast cancerscreening and improved treatment. There are well over two million breast cancer survivors in the United States for whom appropriatesurveillance continues to be a subject of controversy. The guidelines from the American Society of Clinical Oncology (ASCO) and theAmerican College of Physicians are clear: only performance of yearly screening mammography is supported by evidence. Althoughadvanced imaging technologies and sophisticated circulating tumor biomarker studies are exquisitely sensitive for the detection ofrecurrent breast cancer, there is no proof that earlier detection of metastases will improve outcome. A lack of specificity may leadto more tests and patient anxiety. Many breast cancer survivors are not followed by oncologists, and their doctors may not be familiarwith these recommendations. Oncologists also disregard the data. A plethora of both blood tests and nonmammographic imaging testsare frequently performed in asymptomatic women. The blood tests, marker studies, and advanced imaging techniques are expensiveand, with limited health care funds, may prevent funding for more appropriate aspects of patient care. Abnormal marker studies leadto additional imaging procedures. Repeated CT scans and radionuclide imaging may induce a second cancer because of the radiationdose, and invasive procedures performed as a result of these examinations also add risk to patients without clear benefits. Improvedadherence to the current guidelines can cut costs, reduce risks, and improve patient quality of life without adversely affecting outcome.

Society has been trained by the lay press—often stimulatedby the medical fıeld—to be aware of signs and symptoms

of cancer. Although this concept is usually applicable to un-affected people, the perception that earlier detection is betterhas been ingrained in most patients who have been diag-nosed with breast and other cancers. This concern raises is-sues about whether early detection of occult metastases inpatients with asymptomatic cancer has clinical utility, and ifso, what are the optimal tools to use for surveillance.

REASONS TO PERFORM SURVEILLANCE OFPATIENTS WITH ASYMPTOMATIC BREAST CANCEROne might monitor asymptomatic patients with breast can-cer for one of several reasons. Several prospective trials ofadjuvant therapy have demonstrated that early systemic ther-apy improves overall survival (OS) when compared withwaiting until a patient has symptomatic disease, detectableby standard radiographic techniques.1-5 This observationstimulated the hypothesis that monitoring patients forearly recurrence after standard adjuvant systemic therapymight permit early detection of impending metastases lead-ing to earlier treatment, whichwould result in improved out-comes. The main anticipated benefıt of such a strategy is to

administer systemic therapy before the cancer has developedinherent resistance and therefore to improve OS. A secondpurported reason to monitor asymptomatic patients mightbe “up-front palliation.” In this scenario, early detection ofimpending relapse might permit initiation of a new thera-peutic regimen that might delay the time to progression andcancer-related symptoms. Therefore, one might argue thateven if this strategy does not improveOS, the delay in onset ofcancer-related symptomswould outweigh the toxicities asso-ciated with the earlier treatment. A secondary hypothesis re-lated to this concept would be that identifıcation of animpending catastrophic complication of ametastasis, such asfracture, brain metastasis, or advanced visceral involvement,would be benefıcial to the patient. Finally, it has been arguedthat no evidence of recurrence on a surveillance test is reas-suring to patients who are otherwise doing well.

EVIDENCE REGARDING SURVEILLANCE OF PATIENTSWHO ARE ASYMPTOMATICTo address this hypothesis, two prospective trials were initi-ated in Italy in the 1980s inwhichwomenwith a prior historyof breast cancer and who were asymptomatic and free of dis-ease were randomly assigned to either routine follow-up

From the Memorial Sloan-Kettering Cancer Center, New York, NY; University of Michigan Medical Center, Ann Arbor, MI; Schools of Medicine and Public Health, Jonsson Comprehensive CancerCenter, University of California, Los Angeles, Los Angeles, CA.


Corresponding author: Maxine Jochelson, MD, Memorial Sloan-Kettering Cancer Center, 300 East 66th Street #711, New York, NY; email: [email protected].


SURVEILLANCE AND MONITORING IN BREAST CANCER


without any special testing or periodic evaluation for occultmetastases using standard techniques available at the time.6,7Neither of these two trials demonstrated any of the abovebenefıts. Indeed, in one of these trials serial quality-of-lifeanalysis suggested that the anxiety and false-positive fındingsassociated with surveillance were worse in the screened pa-tients.Neither of these trials encompassed what would be consid-

ered modern diagnostic techniques (such as circulating tu-mormarkers or new imaging techniques, includingCT, PET,or MRI). In this regard, in a Finnish trial conducted in the1990s, patients were randomly assigned to blood counts, sed-imentation rate, liver enzymes, and CA15–3 every 3 or 6months after primary treatment or to routine use of diagnos-tic examinations or use based on clinical grounds only if con-cerning clinical fındings were identifıed.8 This trial also failedto detect any difference in outcome among any of the arms.8A 2005 Cochrane Collaboration–sponsoredmeta-analysis ofthese trials concurred that there was no apparent benefıt tointensive surveillance of patients who were asymptomaticfollowing primary and adjuvant systemic therapy for breastcancer.9

CURRENT GUIDELINES REGARDING SURVEILLANCEOF ASYMPTOMATIC BREAST CANCER SURVIVORSBased on the results of the studies from two separate ASCOguidelines committees, as well as the Breast Committee of theNational Cancer Center Network (NCCN), these organiza-tions have strongly recommended against any sort of routinesurveillance for metastic (disease other than routine screen-ing for locally recurrent tumor within the breast or new pri-mary cancers with mammography and age-appropriatescreening for cervical and colorectal cancers).10-12 In sum-mary, the only routine surveillance recommended by ASCOand NCCN is standard breast imaging for detection of anew primary or in-breast recurrence. Mammography isconsidered the standard of care in this regard. Breast MRIis only recommended in genetically high-risk women orwomen who had received mediastinal radiation for Hodgkin

lymphoma who do not elect to have contralateral prophylac-tic mastectomy.13 Indeed, to address a challenge to medicalspecialties to identify diagnostic tests or treatments that arecommonly performed but of dubious meaningful benefıt topatients, ASCO’s Cost of Care Task Force included lack ofadherence to breast cancer surveillance guidelines as one ofthe “Top Five” list for oncology.14

WHY NOT PERFORM ROUTINE SURVEILLANCETESTING?Overuse of imaging is associated with certain risks. These in-clude expense, radiation risk, anxiety, and the risk of addi-tional procedures.15

ExpenseThe cost of health care is amajor problem in this country andhas become a major political issue as well. The United Statesspendsmore on health care per capita thanmost other coun-tries without better results. The cause ismultifactorial but in-cludes cultural issues and fear of lawsuits. Sometimes it is justeasier to say yes to a breast cancer survivorwith no additionalrisk factors and fatty breasts who insists on having a yearlybreast MRI for her remote breast cancer than to show herwhy it is not indicated. The current charge for a bilateralbreast MRI can be more than $4,000.00. In a cohort of morethan 25,000 women diagnosed between 1998 and 2003 withstage I to II breast cancer from the SEER-Medicare linkeddatabase who survived more than 48 months, Panageas et alreported the cost of theMRIs alonewould be almost $1.5mil-lion, and even in the best hands there are many false-positivefındings leading to short interval follow-up MRI, core biop-sies, and occasionally surgery, making it even more expen-sive.16As noted above, although occasionally only a chest X-ray is

performed for surveillance, advanced imaging surveillancefor metastatic disease is often performed with either CT ofthe chest, abdomen, and pelvis and bone scan or more re-cently PET/CT.Charges for aCTandbone scan can runup to$6,000 andPET/CT is evenmore.Womenmay havemultipleexaminations each year. There will be a number of positivefındings, which will generate additional imaging, biopsies, oreven surgery. It has been shown that intensive surveillancetesting adds an additional $260million to $630million to theannual cost of caring for breast cancer survivors withoutdemonstrated benefıt compared with a strategy that followsthe guidelines. 17

Radiation RiskThere has been a great deal of consternation regarding thepossibility of cancers developing as a result of imaging, par-ticularly from the increasing use of CT scans. Much of thedata regarding the radiation risk from imaging studies comesfrom studies of survivors of atomic bombs dropped in Hiro-shima and Nagasaki as well as from 400,000 radiation work-ers in the nuclear industry.18 There are various ways to defıne

KEY POINTS

� Screening mammography is the only imaging test thatshould be performed for surveillance after treatment forbreast cancer.

� Early detection of occult metastases in asymptomaticwomen with a history of breast cancer has not beendemonstrated to improve outcome.

� There is no role for obtaining circulating tumor markers orimaging tests to evaluate for metastatic disease.

� Both oncologists and PCPs routinely order fewermammograms and more tumor marker studies and imagingexams than are recommended by ASCO and NCCNguidelines.

JOCHELSON, HAYES, AND GANZ


radiation exposure, but it is most frequently defıned as theeffective dose of the radiation delivered (measured in sieverts[Sv] ormilli sieverts [mSv]), particularly as it refers to specifıcorgan dose. Long-term survivors of atomic bomb exposureshad an increased risk for cancer, having received 10–100mSv. The Hiroshima and nuclear worker data are based onwhole-body exposure. Although CT exposure measured inthe current literature is to limited portions of the body, a sin-gle CT scan can deliver an equivalent radiation exposure.Brenner and Hall suggested that 1.5% to 2% of cancers in theUnited Statesmight be attributable to CT scans.19 Berringtonde Gonzalez et al used risk models to estimate future cancerrisks from CT scan use in the United States according to age,gender, and scan type. They determined that approximately29,000 future cancers could be the result of the 72million CTscans performed in the United States in 2007, predominantlyfrom CTs of the abdomen and pelvis. As demonstrated inmultiple other studies, the projected number of cancers de-creased with increasing age at exposure. Although pediatricpatients are at highest risk, in terms of absolute numbers, thehighest potential effect is in women age 35 to 54.20 A portionof the breast cancer survivor population falls into this cate-gory.In trying to understand patients’ actual radiation exposure

patients, Smith-Bindman et al conducted a retrospectivestudy evaluating the doses of the 11 most common types ofCT in 1,119 adult patients.21 Not surprisingly, doses variedamong various studies. However, within each study type,they detected a 13-fold variation between the highest andlowest doses, making it very diffıcult to assess individualrisks. More recently, a great deal of effort has been made toreduce the dose of radiation delivered by CT scans, and thereare now requirements for documentation of delivered dosefor each scan performed. However, the most effective dosereduction would be to eliminate unnecessary scans.

FALSE-POSITIVE FINDINGS GENERATE ADDITIONALEXPENSIVE AND RISKY PROCEDURESAs mentioned previously, additional imaging exams gener-ate additional procedures. The lower the prevalence of dis-ease is in a population, the higher the false-positive rate willbe. In a population of women who have early-stage breastcancer, recurrence rates are low. For example, Drotman et alretrospectively reviewed 2,426 pelvic CT scans in patientswith breast cancer and demonstrated that not only was therea low yield of cancer detection of disease in the pelvis as theonly site of metastasis, but also performance of pelvic CT ledto 204 additional tests. These included 186 pelvic ultrasoundsand 50 surgical procedures. These additional procedureswere normal, benign, or indeterminate in 84.6%.22

DO PHYSICIANS STILL MONITOR PATIENTS WHOARE ASYMPTOMATIC TO DETECT OCCULTMETASTASES?Despite these data and guideline recommendations, manyphysicians continue to order either circulating tumormarker

studies—which, if abnormal, lead to imaging studies—or im-aging studies including chest radiographs; MRI; bone scans;CT scans of the chest, abdomen, and pelvis; and PET/CTscans, whereas mammography surveillance is performed lessfrequently than expected.17,23 In a cohort ofmore than 11,000Canadian women followed for 5 years, one-quarter had lessthan the recommended mammographic surveillance,whereas one-half had more than recommended imaging formetastatic disease.23 Han et al surveyed 1,098 primary carephysicians (PCPs) andmedical oncologists, and 84%of PCPsand 72%of oncologists reported beliefs consistentwith bloodtest overuse, and 50% of PCPs and 27% of oncologists hadbeliefs consistent with imaging test overuse.24From a recent national survey of PCPs and oncologists—

the Survey of PhysicianAttitudes Regarding the Care of Can-cer Survivors (SPARCCS)—conducted in 2009, we learnedthat both physician groups recommended unnecessary tu-mor marker and imaging studies in response to a vignette-based case of a 4-year asymptomatic breast cancer survivor.25Specifıcally, 51% of PCPs and 31% of oncologists recom-mended tumormarkers; 42%of PCPs and 22%of oncologistsrecommended chest X-rays; and 23% of PCPs and 3% of on-cologists recommended bone scans.25 Although 85% of on-cologists reported being “very confıdent” about theirknowledge related to follow-up care, only 40% of PCPs re-ported high confıdence in perceived follow-up skills (p �0.001). Thus, there appears to be a need to improve knowl-edge among PCPs regarding the follow-up care of breast can-cer survivors.The exact number of “higher technology” studies per-

formed in the earlier investigations of breast cancer surveil-lance cannot be determined from the available data. Theseinclude CT, MRI, PET, and bone scans. In the cohort ofwomen reported by Panageas, 40% had at least one high-techexamination, and 30% had CT scans. Over time, use of thesetests increased from34%ofwomendiagnosed in 1998 to 43%in women diagnosed in 2003.The increasing use of high-techimaging continued through 2005.16

IS THERE A NEED TO MONITOR PATIENTS WHO AREASYMPTOMATIC?A major concern regarding the recommendations to notscreen asymptomatic patients is that they are based on out-dated technology, lacking sensitivity and specifıcity. In addi-tion, outcome data could possibly be improved in the futurewith the use of new antineoplastic agents causing less toxicityand better antitumor activity (possibly resulting in improvedOS compared with older therapies). For example, althoughtrastuzumab is now used routinely as adjuvant therapy forpatients with HER2-positive breast cancer, it was shown in aprospective randomized trial to provide an OS benefıt forwomen with metastatic breast cancer.26 The results of thisstudy suggest that earlier treatment of women with HER2-positive metastatic disease is superior to waiting for latertreatment, since the OS benefıt was observed even thoughmany patients were permitted to cross over from the



chemotherapy-only arm to receive trastuzumab. Therefore,onemight extrapolate that data to suggest that treatment of apatient who is asymptomatic with impending relapse mightbe more effective than waiting for development of clinicallyapparent metastases.Although these are cogent arguments, they have not been

demonstrated with any level of evidence and are only theo-retical at present. Investigators in the North American coop-erative groups are currently considering a new prospectivetrial that would incorporate modern circulating tumor bio-markers and imaging techniques.

CIRCULATING TUMOR BIOMARKERSSeveral studies have demonstrated that circulating routineliver and bone enzymes have very poor sensitivity and spec-ifıcity for detection of metastases.11 Circulating soluble tu-mor biomarkers—in particular MUC-1 protein andcarcinoembryonic antigen (CEA)—have been shown to havemore accurate performance characteristics in this setting.Two assays for circulating MUC-1 protein (CA15–3 andCA27.29) provide nearly identical results. MUC1 is elevatedin approximately 75% of women with symptomatic metasta-ses and approximately 50% of women with asymptomaticmetastases.27 CEA is elevated in approximately 50% ofwomen with symptomatic metastases and 25% of womenwith asymptomatic metastases. The two are complementary,raising the sensitivity in both settings by approximately 10%.However, both result in false-positive fındings, which areusually associated with benign inflammatory conditions ofthe gut, lung, pleura and peritoneum, and liver.27 Dependingon the cutoffs used to distinguish a positive from a negativefınding, the positive predictive value (PPV) of an elevated cir-culating tumor biomarker ranges from 50% to 95%. As ex-pected, the PPV is particularly high if one applies especiallystringent cutoffs for “positivity” to patients with high-riskdisease, such as T3 primary lesions or positive lymph nodes.Regardless, the median lead time from rising tumor bio-marker levels (again, depending on the criteria used for “pos-itive”) ranges from 3 to 12 months.27The negative predictive value (NPV) of these markers for

recurrence within the next 6 to 12 months in an asymptom-atic patient is quite high, but that is based more on the rela-tively low risk of recurrence for any such patient, especially ifshe had a relatively favorable prognosis and/or is severalyears away from her original diagnosis.27,28 In other words,the NPV for recurrence by history and physical examinationis quite high and is probably not improved by negative resultsof circulating biomarkers.Recently, circulating tumor cells (CTC) have been shown

to be associatedwith aworse prognosis in patientswith eitherearly-stage or metastatic breast cancer.29,30 However, thereare no data regarding use of CTC tomonitor patients who areasymptomatic.

EDUCATING ONCOLOGISTS AND PRIMARY CAREPHYSICIANS ABOUT THE LACK OF EVIDENCE FORSURVEILLANCE TESTINGOngoing surveillance despite the recommendations raisesthe question as to why physicians continue to pursue thisstrategy. In a subsequent analysis from the SPARCCS study,Han et al examined the beliefs about breast cancer surveil-lance testing among PCPs and oncologists in response to thevignette case of a 4-year asymptomatic and disease-freebreast cancer survivor.24 Both groups of physicians reportedbeliefs consistent with overuse of blood and imaging teststhat were not indicated for surveillance in this setting. Inmultivariable models, they examined physician and practicesetting factors that might be associated with the beliefs inoveruse of these tests and procedures. Factors signifıcantlyassociated with PCP beliefs related to overuse of blood testsincludedworking in a smaller practice setting and being a fullor part owner of the practice. For imaging tests, having alower volume of patients with breast cancer and being in asmallermetropolitan area (� 1million population) were sig-nifıcantly associated with greater overuse beliefs. In the caseof oncologists, factors signifıcantly associated with greateroveruse beliefs for both blood and imaging tests includedolder age, being internationally trained, and having lowerself-effıcacy (confıdence in knowledge about surveillancetesting) andhigher perceived ambiguity about testing recom-mendations. Blood test overuse beliefs were signifıcantlygreater among Asian oncologists than white oncologists andwere lower among employed oncologists than practice own-ers. Patient with breast cancer volume was signifıcantly asso-ciated with overuse beliefs for imaging tests, with a nonlinearpattern that was diffıcult to interpret.24The differences in factors associated with overuse beliefs

between PCPs and oncologists are interesting and point todifferent strategies that should be considered to improveknowledge and reduce overuse of nonindicated testing ineach group of physicians. Although the SPARCCS study didnot fınd a signifıcant relationship between the receipt of asurvivorship care plan and overuse beliefs among PCPs, it ispossible that this strategy has not yet had an effect on surveil-lance care because of the lack of self-effıcacy and ambiguityregarding these tests noted among oncologists who would bethe individuals expected to prepare a survivorship care plan.For oncologists, it appears that reducing ambiguity aboutsurveillance guidelines and improvement in self-effıcacymayhave value in addressing their use of unnecessary tests. Al-though the surveillance guidelines for patients with breastcancer have been available from ASCO since 2006, with re-cent reissue, adherence to these guidelines has not beenregularly monitored in most clinical practice settings.11 Per-haps the selection of this guideline as one of ASCO’s recom-mendations for the American Board of Internal MedicineFoundation’s Choosing Wisely Campaign will provide animpetus for health care organizations and oncology practicestomonitor the overuse of blood tests (including tumormark-ers) and advanced imaging tests in breast cancer survivors.31



Providing feedback to oncologists regarding their use ofthese tests in breast cancer survivors could help them appre-ciate what their own practice patterns are and demonstratetheir lack of awareness of the guidelines. Oncologists whowork together in a practice setting should be encouraged todiscuss their ambivalence about the guidelines among theirpeers to establish standards for their practice settings.

COMMUNICATING TO OUR PATIENTS THERECOMMENDATIONS TO NOT SCREEN FORASYMPTOMATIC METASTASESGiven the common misconceptions about the benefıts ofearly detection of recurrent breast cancer among physiciansdescribed earlier, it is not surprising that in practice, rela-tively few oncologists fully adhere to the ASCO and NCCNguidelines regarding lack of value of tumor markers and im-aging studies in the follow-up of patients with breast cancerand survivors. When patients present for regular follow-upvisits after adjuvant chemotherapy and radiation therapy,they deserve a good history and physical examination, as wellas a comprehensive symptom assessment that will elicit lin-gering side effects from past treatment as well as ongoing en-docrine therapy. These efforts also serve to identify newonsetproblems that might herald a recurrence of the disease. Un-fortunately, cognitive services such as this have been de-valued in the procedure-oriented world of fee-for-servicemedicine. Many providers and patients do not believe this issuffıcient, leading the patient to request and the provider toorder blood tests and scans. Instead, this timewith the patientshould be spent educating them about the natural history ofbreast cancer, the individual woman’s prognosis (whichmight be excellent), the harms of surveillance testing (e.g.,false-positives, false-negatives, radiation exposure, andcosts), and the lack of benefıt from testing.Fear of recurrence is extremely common, and patients do

not understand that these tests provide false reassurance thateverything is okay. Instead, the oncology provider, in the

cognitive exchange, must reassure the patient that theyshould report any symptoms or concerns when they occur,and a prompt diagnostic work-up will be undertaken. Psy-chosocial services should be provided to help the patient copewith ongoing uncertainties and the challenges of adjusting tothe “new normal” of survivorship. The use of treatment sum-maries and care plans can facilitate this conversationwith useof the patient information sheet that ASCO has providedalong with its guidelines on surveillance testing after early-stage breast cancer treatment.32 The treatment summary andcare plan should be shared with the patient’s PCP to ensureunnecessary testing does not occur in that setting. A focus onhealth promotion and well-being—including physical activ-ity, energy balance, and social engagement—should be em-phasized rather thanusing testing as the default activity in thefollow-up of the breast cancer survivor.

CONCLUSIONIn summary, contrary to the perceptions held by many pa-tients and their physicians, available data do not support rou-tine surveillance for occultmetastases asymptomatic patientswith a prior history of breast cancer. It is important to con-tinue breast imaging for a patient with intact breasts. Pro-spective randomized trials have failed to demonstrateimprovement in any clinically meaningful endpoints for sur-veillancewith routine blood tests, circulating tumormarkers,or other imaging. Surveillance screening is associated withfalse-positive fındings, induction of anxiety, risk of exposureto radiation, and higher medical expenses. Conveyance ofthis information with a thoughtful recommendation that apatient not undergo screening is an important part of clinicalmanagement of breast cancer survivors. The availability ofmore sophisticated screening techniques and more effective,less toxic therapies raises the possibility that surveillancemight be effective using these newer modalities, but this hy-pothesis must be demonstrated in prospective trials before itis introduced into standard care.


Relationships are considered self-held and compensated unless otherwise noted. Relationships marked “L” indicate leadership positions. Relationships marked “I” are those held by animmediate family member; those marked “B” are held by the author and an immediate family member. Relationships marked “U” are uncompensated. Authors marked with an asterisk (*) areparticipants in ASCO’s Disclosure Management System Pilot; their disclosure is not limited to subject matter under consideration in this article and includes payments to themselves, animmediate family member (I), and/or their institutions (Inst). For information on the pilot program, or to provide feedback, please visit coipilot.asco.org.

Employment or Leadership Position: Patricia A. Ganz, Intrinsic LifeSciences (I). Consultant or Advisory Role: Patricia A. Ganz, Alnylam (I); Pieris (I);Xenon (I). *Daniel F. Hayes, Biomarker Strategies, Inbiomotion. Stock Ownership: Patricia A. Ganz, Intrinsic LifeSciences (I). *Daniel F. Hayes,OncImmune, Medivation. Honoraria: None. Research Funding: Patricia A. Ganz, Amgen (I). *Daniel F. Hayes, Pfizer, Novartis, Veridex, JanssenOncology. Expert Testimony: None. Other Remuneration: *Daniel F. Hayes, Veridex (Intellectual Property Interests).

References

1. Early Breast Cancer Trialists Collaborative Group T. Comparisons be-tween different polychemotherapy regimens for early breast cancer:meta-analyses of long-termoutcome among 100,000women in 123 ran-domized trials. Lancet. 2005;379:432-444.

2. Peto R, Davies C, Godwin J, et al. Comparisons between different poly-chemotherapy regimens for early breast cancer: meta-analyses of long-term outcome among 100,000 women in 123 randomised trials. Lancet.2012;379:432-444.



3. Davies C, Godwin J, Gray R, et al. Relevance of breast cancer hormonereceptors and other factors to the effıcacy of adjuvant tamoxifen: patient-level meta-analysis of randomised trials. Lancet. 2011;378:771-784.


5. Romond EH, Perez EA, Bryant J, et al. Trastuzumab plus adjuvant che-motherapy for operable HER2-positive breast cancer. N Engl J Med.2005;353:1673-1684.

6. GIVIO Investigators. Impact of follow-up and testing on survival andhealth-related quality of life in breast cancer patients: amulticenter ran-domized controlled trial. JAMA. 1994;271:1587-1593.

7. Roselli Del Turco M, Palli D, Cariddi A, et al. Intensive diagnosticfollow-up after treatment of primary breast cancer: A randomized trial.JAMA. 1994;271:1593-1597.

8. Kokko R, Hakama M, Holli K. Follow-up cost of breast cancer patientswith localized disease after primary treatment: a randomized trial.Breast Cancer Res Treat. 2005;93:255-260.

9. Rojas MP, Telaro E, Russo A, et al. Follow-up strategies for womentreated for early breast cancer. Cochrane Database Syst Rev. 2005;CD001768

10. Harris L, Fritsche H, Mennel R, et al. American Society of Clinical On-cology 2007update of recommendations for the use of tumormarkers inbreast cancer. J Clin Oncol. 2007;25:5287-5312.

11. Khatcheressian JL,Wolff AC, Smith TJ, et al. American Society of Clin-ical Oncology 2006 update of the breast cancer follow-up and manage-ment guidelines in the adjuvant setting. J Clin Oncol. 2006;24:5091-5097.

12. Carlson RW, Allred DC, Anderson BO, et al. Invasive breast cancer.J Natl Compr Canc Netw. 2011;9:136-222.

13. SaslowD, Boetes C, BurkeW, et al. American Cancer Society guidelinesfor breast screeningwithMRI as an adjunct tomammography.CACan-cer J Clin. 2007;57:75-89.

14. Brody H. Medicine’s ethical responsibility for health care reform-thetop fıve list. N Engl J Med. 2010;362:283-285.

15. US Preventative Services Task Force. Screening for breast cancer: U.S.Preventive Services Task Force recommendation statement.Ann InternMed. 2009;151:716-726, W-236

16. Panageas KS, Sima CS, Liberman L, et al. Use of high technology imag-ing for surveillance of early stage breast cancer. Breast Cancer Res Treat.2012;131:663-670.

17. Keating NL, Landrum MB, Guadagnoli E, et al. Surveillance testingamong survivors of early-stage breast cancer. J Clin Oncol. 2007;25:1074-1081.

18. Cardis E, Vrijheid M, Blettner M, et al. The 15-country collaborative

study of cancer risk among radiation workers in the nuclear industry:estimates of radiation-related cancer risks. Radiat Res. 2007;167:396-416.

19. Brenner DJ, Hall EJ. Computed tomography-an increasing source ofradiation exposure. N Engl J Med. 2007;357:2277-2284.

20. Berrington de Gonzalez A, Mahesh M, Kim KP, et al. Projected cancerrisks from computed tomographic scans performed in theUnited Statesin 2007. Arch Intern Med. 2009;169:2071-2077.

21. Smith-Bindman R, Lipson J, Marcus R, et al. Radiation dose associatedwith common computed tomography examinations and the associatedlifetime attributable risk of cancer. Arch Intern Med. 2009;169:2078-2086.

22. DrotmanMB, Machnicki SC, Schwartz LH, et al. Breast cancer: Assess-ing the use of routine pelvic CT in patient evaluation. AJR Am J Roent-genol. 2001;176:1433-1436.

23. Grunfeld E, HodgsonDC, Del GiudiceME, et al. Population-based lon-gitudinal study of follow-up care for breast cancer survivors. J OncolPract. 2010;6:174-181.

24. Han PKJ, Klabunde CN,Noone, A, et al. Physician’s beliefs about breastcancer surveillance testing are consistent with test. Med Care. Epub2012 Dec 23.

25. PotoskyAL,HanPK, Rowland J, et al. Differences between primary carephysicians’ and oncologists’ knowledge, attitudes and practices regard-ing the care of cancer survivors. J Gen Intern Med. 2011;26:1403-1410.

26. Slamon DJ, Leyland-Jones B, Shak S, et al. Use of chemotherapy plus amonoclonal antibody against HER2 for metastatic breast cancer thatoverexpresses HER2. N Engl J Med. 2001;344:783-792.

27. Stearns V, Yamauchi H, Hayes DF. Circulating tumormarkers in breastcancer: accepted utilities and novel prospects. Breast Cancer Res Treat.1998;52:239-259.

28. Gion M, Barioli P, Mione R, et al. Tumor markers in breast cancer fol-low-up: A potentially useful parameter still awaiting defınitive assess-ment. Forza Operativa Nazionale sul Carcinoma Mammario(FONCaM). Ann Oncol. 1995;6:31-35.

29. Hayes DF, Smerage J. Is there a role for circulating tumor cells in themanagement of breast cancer? Clin Cancer Res. 2008;14:3646-3650.

30. Lucci A, Hall CS, Lodhi AK, et al. Circulating tumour cells in non-metastatic breast cancer: a prospective study. Lancet Oncol. 2012;13:688-695.

31. Schnipper LE, Smith TJ, RaghavanD, et al. American Society of ClinicalOncology identifıes fıve key opportunities to improve care and reducecosts: the top fıve list for oncology. J Clin Oncol. 2012;30:1715-1724.

32. Ganz PA, Hahn EE. Implementing a survivorship care plan for patientswith breast cancer. J Clin Oncol. 2008;26:759-767.



BREAST CANCER

Treatment Algorithms forHormone Receptor-PositiveAdvanced Breast Cancer

CHAIRStephen R. D. Johnston, MA, MD, PhDThe Royal Marsden Hospital NHS Foundation TrustLondon, United Kingdom

SPEAKERSTodd W. Miller, PhDDartmouth CollegeHanover, NH

Stephen K. L. Chia, MDBritish Columbia Cancer AgencyVancouver, BC, Canada

Treatment Algorithms for Hormone Receptor-PositiveAdvanced Breast Cancer: Applying the Results from RecentClinical Trials into Daily Practice—Insights, Limitations, andMoving ForwardSheridan Wilson, MBChB, and Stephen K. Chia, MD

OVERVIEW

Hormone receptor–positive (HR�) breast cancer is the most prevalent subtype of breast cancer in both early- and advanced-stagedisease. Thus, the treatment of HR� breast cancer has had the greatest global influence in improving clinical outcomes overall.Although the first-line metastatic breast cancer (MBC) trials comparing a third-generation aromatase inhibitor (AI) to tamoxifen havefavored the AI, one of the challenges in translating these findings into clinical practice stems from the influence of prior adjuvantendocrine therapy, particularly the increasing use of adjuvant AIs today, on the choice of endocrine agent in the advanced settingbecause of the development of acquired resistance. Because the majority of patients enrolled into these studies were eitherendocrine-treatment naïve or exposed to tamoxifen only, the “real-life” applicability of the evidence is unclear. Because a superiordose of the selective estrogen receptor (ER) downregulator fulvestrant has now been established, its role as first-line therapy is beingre-established. We are now starting to see the promise realized with blocking cross-talking growth factor pathways in addition to theER pathway. The greatest efficacy is seen with the mammalian target of rapamycin (mTOR) inhibitor everolimus in combination withexemestane and, perhaps to a lesser extent, anti-HER2–directed therapy in combination with an AI. Future gains will likely involve agreater understanding of the redundancy and compensation induced by blocking these pathways, trials involving blocking multiplepathways in addition to hormonal agents, and the molecular interrogation of the individual’s tumor in search of predictive biomarkersand “actionable” genomic aberrations.

Hormone receptor-positive (HR�) breast cancer is themost prevalent subtype of breast cancer in both early-

and advanced-stage disease. Moreover, the treatment ofHR� breast cancer has had the greatest global influence inimproving disease-free survival and overall survival (OS) inearly-stage disease, and in prolonging progression-free sur-vival (PFS) and potentially extending OS in advanced-stagedisease. Since Beatson fırst demonstrated surgical oophorec-tomy induced regressions in inoperable breast cancer over acentury ago, the developmental strategies of hormone ther-apy for the treatment of breast cancer have led to the classesof selective estrogen-receptor modulators, selective estrogen-receptor downregulators, and AI. Only in the past few yearshave we started to see the promise of additional pathway-targeted therapy in combination with hormone therapy havean effect on clinical outcomes. In the current climate HR hasbecome synonymous with ER. If renewed interest in target-ing the androgen receptor succeeds in establishing clinicalutility for this approach, then the accepted meaning of HR�breast cancermayneed to be redefıned. This reviewwill focus

on the current treatment strategies and algorithms in themanagement of HR� MBC on the basis of prior and recentrandomized clinical trials and attempt to integrate the resultsinto standard clinical practice.

FIRST-LINE HORMONE THERAPY IN HR� MBCFollowing seminal publications in the 1970s, tamoxifen be-came established as the preferred mode of endocrine manip-ulation for the treatment of HR� MBC on the basis of itssuperior effıcacy and toxicity profıle. More than 25 yearslater, third-generation AIs have now demonstrated greaterclinical effıcacy in head-to-head trials with tamoxifen. How-ever interpreting the results of these and other fırst-line trialsin the context of current practice can be challenging becauseof changing patterns and prevalence of adjuvant hormonetherapy use, ER reporting, and possible discordant receptorstatus relative to the primary tumor, and increasingsubsequent-line treatment options available. In two largestudies comparing anastrozole with tamoxifen, there werenotable differences in clinical outcomes. PFS (median) was

From the British Columbia Cancer Agency, Vancouver, British Columbia, Canada.


Corresponding author: Stephen K. Chia, MD, British Columbia Cancer Agency, 600 West 10th Ave., Vancouver, BC V5Z 4E6, Canada; email: [email protected].


WILSON AND CHIA


signifıcantly longer for anastrozole (11.1 vs. 5.6 months) inthe smaller NorthAmerican trial, although no difference wasobserved in the larger European TARGET study (8.3monthsvs. 8.2 months).1-2 It should be noted that no OS differencewas demonstrated in either of the trials. A higher percentageof known HR� patients and patients who had received ad-juvant tamoxifen in the North American study may have se-lected for an outcome favoring the AI. The two trials werealso prospectively designed to facilitate a combined analysisin which no difference in PFS was demonstrated.3In the PO25 study, letrozole was associated with a signifı-

cantly improved median time to progression (TTP) (9.4months vs. 6 months for tamoxifen, hazard ratio � 0.72 p �0.0001).4 The study design built in cross-over at the time ofprogression. There was no signifıcant difference in medianOS (hazard ratio� 0.96; 95%CI 0.84–1.09). The steroidalAI,exemestane, has also been shown to improve median PFSover tamoxifen (9.9 vs. 5.8 months), without an OS benefıt(hazard ratio � 1.13;95% CI 0.85–1.5).5Fulvestrant, an estrogen-receptor downregulator, was ini-

tially approved for use in the second-line setting at a dose of250 mgmonthly. Over the last 5 years there has been an evo-lution of the recommended dose and of its position withinthe treatment algorithm forHR�MBC. An updated analysisof the FIRST trial (a randomized phase II study), comparinga higher dose (HD) of fulvestrant (500 mg days 1 and 14, fol-lowed by 500 mg monthly thereafter) with anastrozole, con-fırmed superior TTP for fulvestrant (23.4 months vs. 13.1month) compared with anastrozole.6 It is noteworthy that a

high proportion of patients in both study arms wereendocrine-treatment naïve (71.6% fulvestrant arm and77.7% anastrozole arm), which may make interpretation ofthese results confounded in those patients relapsing on anadjuvant hormonal regimen. Both treatments were well tol-erated with no important differences in prespecifıed events.Notwithstanding the limitations of this as a phase II open-label study, the role of fulvestrant as fırst-line treatment forHR� MBC is promising and the basis for a phase III ran-domized control trial (FALCON).Historically, combination regimens have failed to demon-

strate a benefıt over single endocrine agents. Furthermore, inthe adjuvant setting the combination arm of the ATAC trialfailed to show effıcacy over tamoxifen alone and was discon-tinued.7 A pharmacokinetic interaction whereby anastrozoleconcentrations were lower in the combination with tamox-ifen has been proposed as an explanation for this fınding. Inpreclinical studies fulvestrant is active in a low estrogen en-vironment and has demonstrated effıcacy in combinationwith an AI. Fulvestrant competes with estrogen for bindingof the ER, and reducing estrogen levels might enhance effı-cacy by allowing increased fulvestrant-ER binding.Clinical studies evaluating the effıcacy of fulvestrant by us-

ing a loading dose (500mg day 1, 250mg days 14 and 28, andevery 28 days thereafter) in combination with anastrozolecompared with anastrozole alone have reported mixed re-sults. In the SWOG S0226 study8 a PFS and OS benefıt infavor of the combination arm were seen (PFS 15 vs. 13.5months, hazard ratio � 0.80; 95% CI 0.68–0.94; p � 0.007,OS 47.7 vs. 41.3 months, hazard ratio � 0.81; 95% CI 0.65–1.0; p� 0.049). In contrast, the similarly designed FACT trialfound no statistically signifıcant difference in either TTP orOS (TTP hazard ratio � 0.99; 95% CI, 0.81–1.20; OS hazardratio, 1.0; 95% CI, 0.76–1.32).9 The discordance betweenthese studies may be partly explained by a higher proportionof patients with previous exposure to tamoxifen in the FACTstudy (67% compared with 40% in the SWOG trial) and thelarge proportion of patients with de novo metastatic disease(38.9%) in the SWOG study.In the case of premenopausal women combining a gonad-

otropin-releasing–hormone analog with tamoxifen wasshown to bemore effective than either agent alone in a three-arm study of 161 patients randomly assigned to either buse-relin, tamoxifen (40 mg daily) or both.10 Less than 3% of thepatients enrolled had received adjuvant tamoxifen. The com-bination arm demonstrated superior median PFS comparedwith the buserelin and tamoxifen arms (9.7, 6.3, and 5.6months, respectively; p � 0.03). Median OS was also greaterfor the combination (3.7, 2.5, and 2.9 years; p � 0.01). Evi-dence for combining a gonadotropin-releasing–hormoneanalogwith anAI is less substantial (single-armphase II trialsonly11,12) and therefore this approach is best regarded as asecond-line option.The use of fulvestrant in premenopausal women is yet to be

established. An observational study of fulvestrant in combi-nationwith ovarian suppression forMBCdescribed a clinicalbenefıt rate of 58%, suggesting that further investigation of

KEY POINTS

� The treatment of HR� breast cancer has the greatestglobal influence on breast cancer burden.

� It is important to confirm hormone-receptor presence inthe relapsed tumor and take into consideration priorexposure and duration of exposure of prior adjuvanthormonal agent(s) when selecting hormone-basedtreatment options in the advanced setting.

� At present it does not appear the sequence of hormonalagents in advanced HR� breast cancer affects overallsurvival (OS), rather using all of the agents available in ahormonally sensitive population is the important factor.

� We are now starting to see the promise realized withblocking cross-talking growth factor pathways in additionto the ER pathway, with the greatest efficacy seen withthe mTOR inhibitor everolimus in combination withexemestane and perhaps to a lesser extent anti-HER2–directed therapy in combination with an AI.

� Future gains will likely involve a greater understanding ofthe redundancy and compensation induced by blockingthese pathways, trials involving blocking multiple pathwaysin addition to hormonal agents, and the molecularinterrogation of the patient’s tumor in search of predictivebiomarkers and “actionable” genomic aberrations toestablish that individual’s treatment algorithm.

TREATMENT ALGORITHMS FOR HORMONE RECEPTOR–POSITIVE ADVANCED BREAST CANCER


this combination is indicated.13 Modern trials includingovarian suppressionwith a gonadotropin-releasing hormonehave commonly employed a 28-day schedule of goserelin orbuserelin at 6mg 6-weekly for two treatments and 8mg every8 weeks thereafter. There is no clinical data assessing the rel-ative effıcacy of various gonadotrophin-releasing-hormoneanalogs or different dosing schedules. Despite this, it seemsreasonable to substitute a long-acting agent once responsehas been established.In the absence of impeding visceral crisis or knowledge of

primary endocrine resistance, endocrine strategies remainthe preferred fırst approach for the management of HR�MBC. Recent studies have sought to establish which agentshould be considered for this pole-position. One of the chal-lenges in translating these fındings into clinical practicestems from the influence of prior adjuvant endocrine ther-apy, particularly the increasing use of adjuvant AIs, on thechoice of endocrine agent in the advanced setting. Becausethe majority of patients enrolled in the studies discussedabove were either endocrine-treatment naïve or exposed totamoxifen only, the “real-life” applicability of the evidence isunclear. Nonetheless the literature supports an advantage forAIs over tamoxifen in terms of the clinical end points of TTPand clinical benefıt, although not inOS. Current data suggestfulvestrant may be a reasonable fırst-line option; however,further clinical trials are needed (which are ongoing) tomoreconfıdently establish its position.

SECOND-LINE HORMONE THERAPY IN HR� MBCIn trials investigating third-generation AIs compared withmegestrol for the treatment of HR� MBC progressive on ta-moxifen, the benefıt for AIs appeared small, but their toxicity

profıle compared favorably, particularly with respect to theunintentional weight gain associated with megestrol.14 Withlonger-term follow-up a survival benefıt for anastrozole overmegestrol emerged.15 Although these agents have individu-ally established effıcacy in both fırst- and second-line trials,there is less data to support superiority of one AI agent overanother. Preclinical data indicate that letrozole is a more po-tent inhibitor of aromatization.16 In a head-to-head trialcomparing letrozole with anastrozole following progressionon antiestrogen therapy there was no difference in TTP orOS,17 and as such clinically meaningful superiority for letro-zole over the other AIs has not been defınitively established.Early second-line trials of fulvestrant in patients who had

progressed on prior antiestrogen therapy demonstratedsimilar response rates and survival compared with AIs.18,19A LD regimen (500 mg day 1, 250 mg day 14 and 28, andevery 28 days thereafter) was developed to hasten attainmentof steady-state concentrations and was tested against ex-emestane, following failure of a nonsteroidal AI, in theEFECT study.20 60% of those enrolled had received two ormore prior endocrine therapies. Median TTP was equivalent(3.7 months, hazard ratio � 0.963; CI, 0.819–1.133; p �0.6531) as was the clinical benefıt rate (32.2% and 31.5%, p�0.853). In the three-arm SoFEA trial a similar evaluation wasperformed comparing fulvestrant (LD) alone, with fulves-trant in combinationwith anastrozole or exemestane alone.21Median PFS was 4.8 months, 4.4, and 3.4 months, respec-tively, thus failing to demonstrate any added benefıt for thecombination arm, and confırming the results of the fulves-trant and exemestane arms from EFECT.Two different doses of fulvestrant (250 mg every 28 days

compared with 500mg day 1, 14, 28, and every 28 days there-after) was tested in the CONFIRM study.22 Results support

TABLE 1. First-Line Hormone Therapy Trials in HR� MBC

Bonneterre2000(TARGET)2

Nabholtz20001

Bonneterre2001(Combined)3

Mouridsen2007(PO25)4 Paridaens 20085

Robertson2009(FIRST)6

Mehta 2012 (SWOGS0226)8

Bergh 2012(FACT)9

Tam Ana Tam Ana Tam Ana Let Tam Exe Tam Ful Ana Ful � Ana Ana Ful � Ana Ana

N � 340 328 171 182 511 510 453 454 182 189 102 103 349 345 258 256

CBR% 56 56 59 46 57 52 49 38 Notdone

Notdone

72.5 67 73 70 55 55.1

Median TTP 8.2 8.3 11.3 5.6 8.5 7 9.4 6 9.9 (PFS) 5.8 (PFS) 23.4 13.1 15 (PFS) 13.5 (PFS) 10.8 10.2

Cross-over 51%crossover(double-blind crossoverpermitted)

41% cross over tocombination arm atprogression. Cross-over from either armto Ful 500 mgallowed atprogression.

hazard ratio� 0.9995% CI:0.86-NR; p� 0.941

hazard ratio� 1.4495% CI,1.16-NR; p� 0.005

hazard ratio� 1.13 95%CI, 1.00-NRp � 0.103

hazard ratio� 0.72 p� .0001

Log-rank p � .121 (p� 0.028 usingWilcoxon test)

hazard ratio� 0.6695% CI0.47–0.92P � .01

p � 0.007 hazard ratio �0.99, 95% CI0.81 to 1.20, p� 0.91

HR� % 45 89 60 66 93 100 100 100

Abbreviations: Ana, anastrozole; Exe, exemestane; CBR, clinical benefit rate; Ful, fulvestrant; HR�, hormone-receptor positive; Let, letrozole; MBC, metastatic breast cancer; NR, not reported; PFS;progression-free survival; Tam, tamoxifen; TTP, time to progression.

WILSON AND CHIA


the preferential use of the 500mg regimen that demonstratedsuperior PFS (6.5months vs. 5.5months, hazard ratio� 0.80;95% CI, 0.68–0.94; p � 0.006) without a signifıcant increasein adverse events. A recent update from the study suggestedan OS benefıt for the higher dose of fulvestrant. How andwhere fulvestrant fıts remains an area of active research. Al-though earlier trials demonstrated equivalence, more recentdata has demonstrated that the 500 mg dose is the more effı-cacious dose therefore calling into question the validity ofearlier studies done with a lesser-dose regimen.Without specifıc sequencing trials it is not possible to es-

tablish the optimal order of exposure to hormonal agents inthe advanced setting. Large PFS differences are seldom seenin either fırst- or second-line trials, and in the absence of sig-nifıcant OS benefıts, the specifıc sequencemay be less impor-tant that ensuring exposure to each class of agent over time aslong the tumor(s) remain hormonally sensitive.

THIRD-LINE AND BEYOND HORMONE THERAPY INHR� MBCThere is little data on which to establish evidence-based rec-ommendations for third-line hormonal treatment and be-yond. On the basis of the EFECT study20 and the results of areview of cross-resistance between steroidal and nonsteroi-dal AI’s,23 both fulvestrant and exemestane are reasonableoptions following progression on tamoxifen and a nonsteroi-dal AI as monohormone therapy. The preferred option, inmedically fıt individuals, would be the combination of ex-emestane and everolimus (an mTOR inhibitor)—see below.Endocrine therapy with either progestins or high-dose estro-gen can be considered in selected cases.24 Suitability for fur-ther endocrine therapy following progression on two ormore prior lines must be assessed with consideration of ear-lier individual drug exposures, disease burden, and previousduration of response (clinical benefıt).

HORMONE THERAPY IN COMBINATION WITHADDITIONAL TARGETED AGENTSIt is likely we have realized the majority of the benefıt fromour current hormonal agents given as monotherapy for the

treatment of MBC. The major focus over the past decade hasbeen the investigation of blocking signaling pathways thatpotentially cross-talk to the ER pathway and are thought tobe involved in either intrinsic or acquired resistance. Thepathways that have led to randomized clinical trials includethe HER2 pathway, epidermal growth factor receptor(EGFR) pathway, insulin growth factor receptor pathway,and the phosphatidylinositol 3-kinase (PI3K)-Akt-mammalian target of mTOR pathway.Despite a quantitative inverse relationship between HER2

and ER (and progesterone receptor), approximately half ofall HER2� breast cancers are alsoHR�.25 HER2 overexpres-sion confers a worse prognosis in breast cancer, regardless ofthe accompanying hormone-receptor status. Two phase IIIrandomized controlled trials have been performed with theaddition of targeted anti-HER2–therapy in combination withhormone therapy in HER2� HR� MBC. The TAnDEMstudy (207 patients) combined trastuzumab with anatrozolecomparedwith anastrozole alone as fırst- or second-line hor-mone therapy in advanced stage disease.26 Prior tamoxifen asadjuvant or hormone therapy for MBC was allowed, thoughapproximately 35% to 40% of patients were hormone-therapy naïve on enrollment. Although the addition of tras-tuzumab to an AI did have a statistically signifıcant effect inimproving the hazard ratio for PFS (hazard ratio� 0.63; 95%CI, 0.47–0.84), the absolute improvement wasmodest at best(median PFS 4.8 months vs. 2.4 months; log-rank p �0.0016).Moreover, therewas nodifference inOSbetween thearms, with the authors stating a likely reason being that 70%of patients on the anastrozole-alone arm crossed over to re-ceive a trastuzumab-containing regimen at some point intime postprogression. The second randomized controlledtrial to mention is the phase III trial that compared the com-bination of letrozole plus lapatinib with letrozole with a pla-cebo as fırst-line treatment of HR� MBC, which included apopulation of knownHER2� tumors.27 In theHER2�HR�cohort (219 patients), the addition of lapatinib to letrozoleimproved PFS (hazard ratio � 0.71; 95% CI, 0.53–0.96), butagainwith no difference as of yet seen inOS (though less than50% of deaths have occurred at time of analysis). Although

TABLE 2. Second-Line Hormone Therapy Trials in HR� MBC

Buzdar 199815Dombernowsky199836

Kaufmann200037

Rose200317

Chia200820 Johnston 201221 Di Leo 201022

Ana Meg Let 2.5 mg Meg Exe Meg Let Ana Ful Exe Ful Ful � Ana Exe Ful 500 Ful 250

n 263 253 174 189 366 403 356 357 351 342 231 243 249 362 374

CBR 35 32 37.4 34.6 27 23 32.2 31.5 Not available 45.6 39.6

TTP months 4.8 4.6 5.6 5.5 4.7 3.9 5.7 5.7 3.7 3.7 4.8 4.4 3.4 6.5 5.5

Significance hazard ratio� 0.94; p� .49)(0.76, 97.5%CI, 1.16)

p � 0.07 p � 0.0370 p � 0.653 p � 0.98; p � 0.56 hazardratio�0.80; 95%CI, 0.68 to0.94 p �0.006

Known HR � 26% 42% 68% 49% 98% 100% 100% 100%

Abbreviations: Ana, anastrozole; Exe, exemestane; Ful, fulvestrant; HR�, hormone-receptor positive; Let, letrozole; Meg, megestrol; MBC, metastatic breast cancer.



numerically the hazard ratio was less on this trial comparedwith the TAnDEM trial, the absolute improvement in PFSwas greater (median PFS 8.2 months vs. 3.0 months). Takentogether, it would appear reasonable to offer a postmeno-pausal woman with HR� HER2� advanced breast cancernot medically fıt to receive chemotherapy with trastuzumab,a combination of either lapatinib or trastuzumab with an AI,with the goal of prolonging PFS but uncertain of improvingOS. The other reasonable scenario would be in patients onmaintenance trastuzumab (and possibly pertuzumab as well)postcombination chemotherapy to add a hormonal agent tothe antibody(s) with hopes again of prolonging PFS and de-laying the time to the next chemotherapy regimen.Preclinical data has suggested that cross-talk between

growth-factor receptors and the ER pathway are involved inthe development of endocrine resistance. Increased expres-sion of EGFR, HER2 and IGF-1 receptors is associated ta-moxifen resistance via activation of downstream signalingpathways.28 Two randomized phase II studies have been per-formed of hormone therapy in combination with either anoral EGFR tyrosine kinase inhibitor (TKI) (gefıtinib) or a pla-cebo in HR� MBC.29-30 In the smaller of the two studies (93patients), the PFS was longer in the anastrozole and gefıtinibarm comparedwith the anastrozole and placebo arm (hazardratio � 0.55; 95% CI, 0.32–0.94) with a median PFS of 14.7months versus 8.4 months, respectively.29 Numerically theclinical benefıt rate was also higher in the gefıtinib arm, butno results were provided in regards to the effect onOS. In theother study in which gefıtinib or the placebo was combinedwith tamoxifen (n� 290), there was no improvement in PFS(hazard ratio � 0.84; 95% CI, 0.59–1.18).30 What was an in-teresting observation was that in the cohort of patients thatwere endocrine-therapy naïve, the hazard ratio was 0.78(95% CI, 0.52–1.15) in favor of the gefıtinib arm; however, inpatients treated with prior tamoxifen or prior AIs, the hazardratio in fact favored the placebo arms (hazard ratio � 1.45,95% CI 0.63–3.45; hazard ratio � 1.16, 95% CI 0.69–1.93,respectively). Because these studies having conflicting results(to some degree), the phase II statistical design, relatively

small sample sizes, and no demonstration of an effect onOS—it would not be advisable to prescribe gefıtinib or anyother EGFR TKI in combination with hormone therapy inHR� MBC.In the previously mentioned study by Johnston et al.,27

there was a preplanned hypothesis to explore the role of lapa-tinib to overcome endocrine resistance in the cohort of HR�and HER2- patients experiencing a relapse on or less than 6months since discontinuation of adjuvant tamoxifen. Thepremise being preclinical models demonstrating growth-factor activity is enhanced in association with endocrine re-sistance. Because lapatinib is a dual TKI against EGFR andHER2, the combination of lapatinib and letrozole would beideal to test this hypothesis. In fact, therewas a nonsignifıcanttrend toward a prolonged PFS for lapatinib-letrozole in thiscohort (200 patients), with a hazard ratio of 0.78 (95% CI,0.57–1.07; p � 0.117) and median PFS of 8.3 months versus3.1 months, respectively. Although it may be tempting toconsider this combination in tamoxifen resistant scenarios,this was only an exploratory subgroup analysis, there aregreater grade 3 and 4 toxicities (primarily diarrhea and rash)for the combination, and there is no evidence of a survivalbenefıt to recommend this therapeutic strategy.Perhaps the greatest promise realized so far in attempting

to combine targeted agents to hormone therapy has been inblocking the PI3K-Akt-mTORpathway. Preclinical data sug-gests a close interaction between this pathway and ER signal-ing, with a substrate of the mTOR complex 1 (mTORC1)having the ability to directly activate the ER in a ligand inde-pendent manner. In the landmark BOLERO-2 study, 724postmenopausal womenwithHR�MBCandprior exposureto a nonsteroidal AI were randomly assigned to eithereverolimus (a mTORC1 inhibitor) and exemestane or a pla-cebo and exemestane.31 Despite the patient population beingrelatively heavily pretreated (100%prior nonsteroidal AI, ap-proximately 50% prior tamoxifen, and 68% had receivedprior chemotherapy) there was an improvement in PFS withamedianPFSof 6.9monthswith everolimus plus exemestanecompared with 2.8 months for the placebo and exemestane

TABLE 3. Targeting Additional Pathways in Addition to Hormone Therapy in HR� MBC

Kaufman 200926 Johnston 200927 Cristofanilli 201029 Osborne 201130 Baselga 201231 Finn 201234

Ana Ana �Tras

Let �Placebo

Let �Lapatinib

Ana �Placebo

Ana �Gefitinib

Tam �Placebo

Tam �Gefitinib

Exe �Placebo

Exe �Everolimus

Let Let �PD0332991

n 104 103 108 111 50 43 101Stratum1

105Stratum1

239 485 81 84

CBR 27.9% 42.7% 29% 48% 34% 49% 45.5% 50.5% NR NR 44% 68%

PFS (months) 3.8 5.6* 3.0 8.2* 8.4 14.7* 8.8 10.9* 2.8 6.9* 7.5 26.2*

OS (months) 23.9 28.5 NR NR NR NR NR

Significance* hazard ratio �0.63 95% CI:0.47–0.84

hazard ratio � 0.7195% CI: 0.53–0.96

hazard ratio � 0.5595% CI: 0.32–0.94

hazard ratio � 0.8495% CI: 0.59–1.18

hazard ratio � 0.4395% CI: 0.35–0.54

hazard ratio �0.32 95% CI:0.19–0.56

Target HER2 HER2 EGFR EGFR mTOR CDK 4/6

Abbreviations: Ana, anastrozole; CBR, clinical benefit rate; EGFR, epidermal growth factor receptor; Exe, exemestane; Ful, fulvestrant; HR�, hormone-receptor positive; Let, letrozole; Meg,megestrol; MBC, metastatic breast cancer; mTOR, mammalian target of rapamycin; NR, not reported; PFS, progression-free survival; Tras, trastuzumab.

WILSON AND CHIA


(hazard ratio � 0.43; 95% CI, 0.35–0.54; p � 0.001). Thisclinically relevant improvement in PFS unfortunately comesat the expense of associated greater toxicities in the everoli-mus arm, in particular increased grade 3 or 4 stomatitis (8%),anemia (6%), dyspnea (4%), fatigue (4%), hyperglycemia(4%), and pneumonitis (3%). Fortunately though, time to de-terioration of performance status and quality of life (�5%)was not statistically different between the treatment groupsas measured on the study. OS data are not yet mature, withresults expected within the next year. Subgroup analysesdemonstrated the consistency of the results across all sub-groups assessed. With these compelling results, it would beappropriate for clinicians to consider the combination ofeverolimus and exemestane as a standard option in HR�postmenopausal MBC with prior exposure to nonsteroidalAI inmedically fıt individuals who would be followed closelyby their oncologist, and whom are aware of the toxicity spec-trum andmanagement strategies associated with this mTORinhibitor.

WHERE WILL PROGRESS BE POTENTIALLY MADE INTHE NEAR FUTURE?Technology today is allowing us to perform genomic andtranscriptomic analyses to high resolution in breast cancersin a timely and less-costly manner than ever before. We havenow recognized from a molecular landscape not all HR�breast cancers are identical, to now using subtype designa-tions of luminal A and luminal B breast cancers.32 In a recentintegrated clustering analysis of copy number and gene ex-pression of close to 2,000 breast tumors, 10 novel subtypes(or clusters) were identifıed with distinct clinical outcomes.33This included the discovery of a high-risk ER� subgroupcomposed of 11q13/14 cis-acting luminal tumors. Severalknown and putative driver genes reside in this region, such asCCND1, EMSY, PAK1 andRSF1. A recent randomized phaseII trial of an oral inhibitor of cyclin-dependent kinase 4/6(PD 033299) in combination with letrozole demonstrated asignifıcant improvement in PFS (p � 0.001) compared withletrozole alone, with preclinical studies suggesting a greatersensitivity in tumors with CCND1 amplifıcation.34 The hopeis that as the capacity increases to molecularly interrogatebreast tumors on a real-time basis we can understand themutational evolution that occurs with disease progression,

identify potentially “actionable” genomic aberrations, andultimately use the information to better select therapeuticagents more likely to be effıcacious.ER biology is inextricably linked to multiple cell-signaling

pathways with known cross-talk and regulatory feedbackloops betweenpathways.As a further example of linked path-way cross-talk, both preclinical and a clinical studies havedemonstrated that blockade of mTOR can induce AKT acti-vation and activation of IGF-1R (via induction of IRS-1) as aresult of loss of a potent intrinsic-negative feedback loop.35Thus further preclinical studies with human tumor samplesand tissue-based pharmacodynamic studies will need to beperformed to better understand the complexity of blocking apathway and the resultant alterations in other linked path-ways to move this fıeld further forward. Only with this infor-mationwill we have the knowledge of whether to block singleor multiple pathways together with hormonal agents todelay or prevent the development of resistance to hormonetherapy.

CONCLUSIONHormonal-based treatment strategies are and will almostcertainly remain the mainstay of therapy in the majority ofbreast cancers today and into the foreseeable future. At pres-ent it does not appear the sequence of hormonal agents effectOS rather using all of the agents available in a hormonallysensitive population is the important factor. It is likely priorexposure (duration of exposure and time from exposure) toadjuvant hormonal agent(s) and a greater understanding ofthe pathways involved in primary and acquired hormonal re-sistance will dictate the choice of the actual hormonal agentused. We are now starting to see the promise realized withblocking cross-talking growth factor pathways in addition tothe ER pathway, with the greatest effıcacy seen with themTOR inhibitor everolimus in combination with exemes-tane and perhaps to a lesser extent anti-HER2–directed ther-apy in combination with an AI. Future gains will likelyinvolve a greater understanding of the redundancy and com-pensation induced by blocking these pathways, trials involv-ing blocking multiple pathways in addition to hormonalagents, and the molecular interrogation of the patient’s tu-mor in search of predictive biomarkers and “actionable”genomic aberrations to establish that individual’s treatmentalgorithm.



Employment or Leadership Position: None. Consultant or Advisory Role: Stephen K. Chia, AstraZeneca; Genomic Health; Novartis; Roche. StockOwnership: None. Honoraria: Stephen K. Chia, Roche. Research Funding: Stephen K. Chia, AstraZeneca; Genomic Health; Novartis; Roche. ExpertTestimony: None. Other Remuneration: None.



References

1. Nabholtz JM, Buzdar A, Pollak M, et al. Anastrozole is superior to ta-moxifen as fırst-line therapy for advanced breast cancer in postmeno-pausal women: results of a North American multicenter randomizedtrial. Arimidex Study Group. J Clin Oncol. 2000;18:3758-3767.

2. Bonneterre J, Thurlimann B, Robertson JF, et al. Anastrozole versus ta-moxifen as fırst-line therapy for advanced breast cancer in 668 post-menopausal women: results of the Tamoxifen Or ArimidexRandomized Group Effıcacy And Tolerability Study. J Clin Oncol. 2000;18:3748-3757.

3. Bonneterre J, Buzdar A, Nabholtz JM, et al. Anastrozole is superior totamoxifen as fırst-line therapy in hormone receptor positive advancedbreast carcinoma. Cancer. 2001;92:2247-2258.

4. Mouridsen HT. Letrozole in advanced breast cancer: The PO25 trial.Breast Cancer Res Treat. 2007; 105 Suppl 1:19-29.

5. Paridaens RJ, Dirix LY, Beex LV, et al. Phase III study comparing ex-emestane with tamoxifen as fırst-line hormonal treatment of metastaticbreast cancer in postmenopausal women: The European Organizationfor Research and Treatment of Cancer Breast Cancer CooperativeGroup. J Clin Oncol. 2008;26:4883-4890.

6. Robertson J, Lindemann J, Llombart-Cussac A, et al. Abstract S1-3: Acomparison of fulvestrant 500 mg with anastrozole as fırst-line treat-ment for advanced breast cancer: Follow-up analysis from the ‘FIRST’study. Cancer Res 2010;70(24 Suppl):Abstract nr S1-3.

7. BaumM, Budzar AU, Cuzick J, et al. Anastrozole alone or in combina-tion with tamoxifen versus tamoxifen alone for adjuvant treatment ofpostmenopausal women with early breast cancer: fırst results of theATAC randomised trial. Lancet. 2002;359:2131-2139.

8. Mehta RS, Barlow WE, Albain KS, et al. Combination anastrozole andfulvestrant inmetastatic breast cancer.NEngl J Med. 2012;367:435-444.

9. Bergh J, PE, Lidbrink EK, et al. FACT: An open-label randomized phaseIII study of fulvestrant and anastrozole in combination compared withanastrozole alone as fırst-line therapy for patientswith receptor-positivepostmenopausal breast cancer. J Clin Oncol. 2012;30:1919-1925.

10. Klijn JG, Beex LV,Mauriac L, et al. Combined treatment with buserelinand tamoxifen in premenopausal metastatic breast cancer: a random-ized study. J Natl Cancer Inst. 2000;92:903-911.

11. Carlson RW, Theriault R, Schurman CM, et al. Phase II trial of anastro-zole plus goserelin in the treatment of hormone receptor-positive, met-astatic carcinoma of the breast in premenopausal women. J Clin Oncol.2010;28:3917-3921.

12. Park IH, Ro J, Lee KS, et al. Phase II parallel group study showing com-parable effıcacy between premenopausal metastatic breast cancer pa-tients treated with letrozole plus goserelin and postmenopausal patientstreated with letrozole alone as fırst-line hormone therapy. J Clin Oncol.2010;28:2705-2711.

13. Bartsch R, Bago-Horvath Z, Berghoff, et al. Ovarian function suppres-sion and fulvestrant as endocrine therapy in premenopausal womenwith metastatic breast cancer. Eur J Cancer. 2012;48:1932-1938.

14. Carlini P, Bria E, Giannarelli D, et al. New aromatase inhibitors assecond-line endocrine therapy in postmenopausal patients with meta-static breast carcinoma: a pooled analysis of the randomized trials.Can-cer. 2005;104:1335-1342.

15. Buzdar AU, Jonat W, Howell A, et al. Anastrozole versus megestrol ac-etate in the treatment of postmenopausal women with advanced breastcarcinoma: results of a survival update based on a combined analysis ofdata from two mature phase III trials. Arimidex Study Group. Cancer.1998;83:1142-1152.

16. Geisler J. Differences between the non-steroidal aromatase inhibitors

anastrozole and letrozole-of clinical importance? Br J Cancer. 2011;104:1059-1066.

17. Rose C, Vtoraya O, Pluzanska A, et al. An open randomised trial ofsecond-line endocrine therapy in advanced breast cancer. comparisonof the aromatase inhibitors letrozole and anastrozole. Eur J Cancer.2003;39:2318-2327.

18. Osborne CK, Pippen J, Jones SE, et al. Double-blind, randomized trialcomparing the effıcacy and tolerability of fulvestrant versus anastrozolein postmenopausal women with advanced breast cancer progressing onprior endocrine therapy: results of a North American trial. J Clin Oncol.2002;20:3386-3395.

19. Howell A, Robertson JF,QuaresmaAlbano J, et al. Fulvestrant, formerlyICI 182,780, is as effective as anastrozole in postmenopausal womenwith advanced breast cancer progressing after prior endocrine treat-ment. J Clin Oncol. 2002;20:3396-3403.

20. Chia S, Gradishar W, Mauriac L, et al. Double-blind, randomized pla-cebo controlled trial of fulvestrant compared with exemestane afterprior nonsteroidal aromatase inhibitor therapy in postmenopausalwomen with hormone receptor-positive, advanced breast cancer: Re-sults from EFECT. J Clin Oncol. 2008;26:1664-1670.

21. Johnston S, Kilburn LS, Ellis P, et al. 2LBA fulvestrant alone or withconcomitant anastrozole vs exemestane following progression onnon-steroidal aromatase inhibitor - fırst results of the SoFEa trial(CRUKE/03/021 & CRUK/09/007) (ISRCTN44195747). Eur J Cancer.2012;3:S2.

22. Di Leo A, Jerusalem G, Petruzelka L, et al. Results of the CONFIRMphase III trial comparing fulvestrant 250 mg with fulvestrant 500 mg inpostmenopausal women with estrogen receptor-positive advancedbreast cancer. J Clin Oncol. 2010;28:4594-4600.

23. Beresford M, Tumur I, Chakrabarti J, et al. A qualitative systematic re-view of the evidence base for non-cross-resistance between steroidaland non-steroidal aromatase inhibitors inmetastatic breast cancer.ClinOncol (R Coll Radiol). 2011;23:209-215.

24. Lønning PE, Taylor PD,AnkerG, et al. High-dose estrogen treatment inpostmenopausal breast cancer patients heavily exposed to endocrinetherapy. Breast Cancer Res Treat. 2001;67:111-116.

25. Konecny G, Pauletti G, Pegram M, et al. Quantitative association be-tweenHER-2/neu and steroid hormone receptors in hormone receptor-positive primary breast cancer. J Natl Cancer Inst. 2003;95:142-153.

26. Kaufman B, Mackey JR, Clemens MR, et al. Trastuzumab plus anastro-zole versus anastrozole alone for the treatment of post-menopausalwomen with HER2� HR� metastatic breast cancer: results from therandomized phase III TAnDEMstudy. J ClinOncol. 2009;27:5529-5537.

27. Johnston S, Pippen J Jr, Pivot X, et al. Lapatinib combinedwith letrozoleversus letrozole and placebo as fırst-line therapy for post-menopausalhormone receptor positive metastatic breast cancer. J Clin Oncol. 2009;27:5538-5546.

28. Musgrove EA, Sutherland RL. Biological determinants of endocrine re-sistance in breast cancer. Nat Rev Cancer. 2009;9:631-643.

29. Cristofanilli M, Valero V, Mangalik A, et al. Phase II, randomized trialto compare anastrozole combined with gefıtinib or placebo in post-menopausal women with hormone receptor positive metastatic breastcancer. Clin Cancer Res. 2010;16:1904-1914.

30. Osborne CK, Neven P, Dirix LY, et al. Gefıtinib or placebo in combina-tion with tamoxifen in patients with hormone receptor positive meta-static breast cancer: a randomized phase II study.Clin Cancer Res. 2011;17:1147-1159.

31. Baselga J, CamponeM, PiccartM, et al. Everolimus in post-menopausal

WILSON AND CHIA


hormone receptor positive advanced breast cancer.N Engl J Med. 2012;366:520-529.

32. Perou CM, Sørlie T, Eisen MB, et al. Molecular portraits of humanbreast tumours. Nature. 2000;406:747-752.

33. Curtis C, Shah SP, Chin SF, et al. The genomic and transcriptomic ar-chitecture of 2,000 breast tumours reveals novel subgroups. Nature.2012;486;346-352

34. Finn RS, Crown JP, Boer K, et al. Results of a randomized phase II studyof PD 0332991, a cyclin dependent kinase 4/6 inhibitor in combinationwith letrozole vs letrozole alone for fırst line treatment of ER�/HER2-advanced breast cancer. In: The 2012 Proceedings of the San AntonioBreast Cancer Symposium. Abstract S1-6

35. Di Cosimo S, Scaltriti M, Val D, et al. The PI3-K/AKT/mTOR pathway

as a target for breast cancer therapy. J Clin Oncol. 2007;25 (suppl; abstrabstract 3511). a footnote to explain the asterisks that appears in the PFSrow.

36. Dombernowsky P, Smith I, Falkson G, et al. Letrozole, a new oralaromatase inhibitor for advanced breast cancer: double-blind ran-domized trial showing a dose effect and improved effıcacy and tol-erability compared with megestrol acetate. J Clin Oncol. 1998;16:453-461.

37. Kaufmann M, Bajetta E, Dirix LY, et al. Exemestane is superior tomegestrol acetate after tamoxifen failure in postmenopausal womenwith advanced breast cancer: results of a phase III randomized double-blind trial. The Exemestane study Group. J Clin Oncol. 2000;18:1399-1411.



Treatment Algorithms for Hormone Receptor–PositiveAdvanced Breast Cancer: Going Forward in EndocrineTherapy—Overcoming Resistance and Introducing New Agents

Stephen R. D. Johnston, MA, PhD, and Gaia Schiavon, PhD

OVERVIEW

Overcoming de novo or acquired endocrine resistance remains critical to further enhancing the benefit of existing endocrine therapies.Recent progress has been made in understanding the molecular biology associated with acquired endocrine resistance, includingadaptive “cross-talk” between ER and various growth factor receptor and cell-signaling pathways. Strategies that combine endocrinetherapy with targeted inhibitors of growth factor receptors or cell-survival pathways to further enhance first-line response have largelybeen disappointing, suggesting that any attempts to prevent endocrine resistance by blocking specific pathways from the outset willbe futile. In contrast, success has been seen by selecting patients with acquired endocrine resistance and enhancing response to furtherendocrine therapy by the addition of mTOR antagonists. Numerous other therapeutics are being evaluated in combination with endocrinetherapies based on varying levels of preclinical science to support their use, including inhibitors of PI3K, HDAC, Src, IGFR-1, and CDK4/6.Enriching trial recruitment by molecular profiling of different ER� subtypes will become increasingly important to maximize anyadditional benefit that these new agents may bring to current endocrine therapies for breast cancer.

Endocrine therapies are widely used in the treatment ofrecurrent/metastatic estrogen receptor–positive (ER�)

advanced breast cancer, given their proven effıcacy and gen-erally favorable side-effect profıle.1 Unfortunately, not all pa-tients respond to fırst-line endocrine treatment due toprimary de novo resistance, while others may initially re-spond but eventually progress with secondary acquired resis-tance.2 Recent years have witnessed major efforts tounderstand the various biologic mechanisms responsible forthe development of endocrine resistance, with the aim ofidentifying new clinical therapeutic strategies to enhance theeffıcacy of current treatment strategies for hormone recep-tor–positive breast cancer.3,4 Not only has progress beenmade in understanding how endocrine resistance develops inlaboratory models, but also new clinical strategies haveemerged including the addition of mTOR antagonists to en-docrine therapy for patients refractory to prior therapy.However, not all strategies that initially looked promising inexperimental models have proved successful in the clinic.This article provides an update on some of the current ther-apeutic strategies that are being investigated as a means tofurther enhance the benefıt of existing endocrine therapiesand examines some of the clinical trial design issues that needto be considered for this approach to be successful.

REVERSING OR PREVENTING ENDOCRINERESISTANCE: WHAT IS THE SCIENCE?To enhance the benefıt of existing endocrine therapies andimprove outcomes in ER� advanced breast cancer, it has be-come clear that understanding key molecular pathways inER� breast cancer is central to developing effective clinicalstrategies that might reverse or prevent endocrine resistanceoccurring. Acquired endocrine resistance develops as a con-sequence of a series of complex adaptive changes occurringin breast cancer cells, during the selective pressure of long-term endocrine treatment.5 Activation of various intracellu-lar signaling pathways leads to endocrine resistance inpreclinical models, and increasing evidence suggest that tar-geting a number of these pathways could be a valid strategy toreverse resistance to endocrine therapy. There is evidence forincreased “cross-talk” between various growth factor recep-tor signaling pathways and ER at the time of relapse on en-docrine therapy, with ER often becoming activated andsupersensitized by a number of different intracellular ki-nases, includingmitogen-activated protein kinases (MAPKs)that are activated either by peptide growth factors such ashuman epidermal growth factor receptors (i.e., EGFR,HER2,HER3), the insulin-like growth factor (IGF) signaling path-way, or the fıbroblast growth factor receptor (FGFR).6-8 As

From the Department of Medicine, Royal Marsden NHS Foundation Trust, Chelsea, London, United Kingdom.


Corresponding author: Stephen R. D. Johnston, MA, PhD, Department of Medicine, Royal Marsden NHS Foundation Trust, Fulham Road, Chelsea, London, SW3 6JJ, United Kingdom; email:[email protected].


JOHNSTON AND SCHIAVON


such, ER-mediated gene transcription and endocrine-resistant cell growth in these cells can then be interrupted byusing a number of different approaches to abrogate upstreamsignaling, including the EGFR tyrosine kinase inhibitor(TKI) gefıtinib, anti-HER2 therapies, andMEK inhibitors.9,10This has provided the clinical rationale for combining vari-ous growth factor therapies with endocrine therapy, and sev-eral different trials have now been completed with mixedresults as discussed below.Likewise, ER can become involved with the phosphatidyl-

inositol 3-kinase (PI3K)/Akt/mammalian target of rapamy-cin (mTOR) pathway in breast cancer cells, with bothgenomic and nongenomic “cross talk” between this signalingpathway and ER. Because of its role in cell survival, there isevidence that the pathway becomes activated in acquiredhormone-resistant breast cancer and accounts for survival ofcells despite the presence of continued endocrine blockade.5Preclinical models of both ER� hormone-sensitive and re-sistant breast cancer have been used to examine the effects ofcombining mTOR antagonists with endocrine therapy, withclear evidence for additive/synergistic effects11,12 These labo-ratory data provided a strong rationale for combining endo-crine therapy with mTOR inhibition as a means of treatingendocrine resistance, which to date has proved the most suc-cessful strategy to be translated into the clinical setting –however, as discussed below it is unclear whether this ap-proach can only be used to treat established endocrine resis-

tance, or whether it could be used as fırst-line therapy toprevent or delay resistance developing in hormone-sensitivebreast cancer.

TREATING ENDOCRINE RESISTANCE WITHTARGETED THERAPIES IN THE CLINIC: HAVE WEMADE PROGRESS RECENTLY?Given the involvement of both peptide growth factor signal-ing and cell-survival pathways in models of endocrine resis-tance, a number of recent trials have been conducted eitherwith targeted therapies such as EGFR/HER2 inhibitors ormTOR antagonists given in combination with endocrinetherapy.13 Some of these studies were conducted in patientswith established hormonal resistancewhere activated growthfactor or cell-survival pathways may already be operative,and it was hoped that combined endocrine and targeted ther-apy could be more effective than using another endocrinetherapy given alone. Other randomized trials were con-ducted in the fırst-line advanced breast cancer setting withthe expectation that combined endocrine and targeted ther-apymight delay the time to disease progression on endocrinetherapy alone by blocking from the outset a key resistancemechanism (i.e., peptide growth factor signaling) that oper-ates in ER� breast cancer cells to cause the development ofacquired resistance. Although some of these approaches haveyielded success, improving the response to existing endo-crine therapy is not necessarily that simple.Two randomized trials have assessed whether co-targeting

EGFRandER could enhance the benefıt of endocrine therapyin the fırst-line advanced breast cancer setting (Table 1).These studies set out to prove the hypothesis that combina-tion therapy could delay the onset of acquired resistance toendocrine therapy, as demonstrated in xenograft models invivo.14 A double-blind placebo-controlled phase II trial of ta-moxifen with/without gefıtinib included patients with endo-crine naïve disease or who had developed greater than a yearafter completion of adjuvant tamoxifen (Stratum1, n� 206),or had developed during or after AI therapy (Stratum 2 n �84).15 In the endocrine naïve patients, there was a numericalincrease in progression-free survival (PFS) from 8.8 to 10.9months, while patients who had been pre-exposed to AIs didnot gain any benefıt from the combination. A second ran-domized trial of gefıtinib and anastrozole compared withanastrazole alone in a similar fırst-line patient population ofwomen with ER�ve–advanced breast cancer reported a sig-nifıcant prolongation of progression-free survival fromame-dian of 8.2 months with anastrozole to 14.6 months with thecombination (HR 0.55, 95% CI 0.32–0.94).16 Although thenumber of patients in this second study was only 93, a sub-sequent combined analysis of both clinical trials suggestedthat the benefıt for the combination was seen exclusively inthose patients who were endocrine-therapy naïve, includingno prior endocrine therapy in the adjuvant setting. On thebasis of these results, a prospectivemulticenter study (MINT,NCT 01151215) was conducted with a novel tyrosine kinase

KEY POINTS

� Endocrine resistance (either primary de novo or secondaryacquired) limits the ultimate effectiveness of current first-line endocrine treatment for ER� advanced breast cancer.

� Significant progress in both laboratory and translationalstudies has been made in identifying key peptide growthfactor and cell-survival signaling pathways that account forendocrine resistance.

� The addition of either EGFR or HER2 targeted therapies toendocrine therapy in hormone-sensitive breast cancer doesnot significantly improve efficacy, although co-targeting ofknown ER� HER2� disease is an effective strategy.

� The addition of mTOR antagonists to further endocrinetherapy in endocrine-resistant breast cancer significantlyimproves clinical outcomes and is an important advance inthe second-line setting, although to date there is nobenefit for the combination in the first-line setting.

� Based on scientific evidence for each signaling pathwayand the availability of appropriate therapeutics, numeroustargeted therapies are now being evaluated in variousendocrine combinations studies in ER� advanced breastcancer.

� Appropriate clinical trial design, patient selection, andmolecular profiling of ER� breast cancer (includingbiopsies of metastatic disease) will be increasinglyimportant to ensure success for future trials of combinedendocrine and targeted therapy.



inhibitor AZD8931, a potent inhibitor of EGFR, HER2, andHER3, to test the hypothesis that combined therapy ofgrowth factor blockade togetherwith anastrozole could delaytime to progression compared with anastrozole alone in en-docrine therapy naïve metastatic–breast cancer. As such, theMINT study will be the defınitive test of this concept thatblocking this pathway could be a strategy to delay/preventendocrine resistance from the outset.Co-targeting HER2 in hormone receptor positive–breast

cancer has also been explored as ameans of improving endo-crine responsiveness, given the evidence that HER2 expres-sion results in breast cancer models is associated withprimary resistance to endocrine therapy. Three randomizedtrials,171819 have confırmed that this approach can treat thisform of endocrine resistance in known ER� HER2� ad-vanced disease (Table 1). In the large EGF30008, there werean additional 952 patients with ER� HER2-negative tumorswhere the hypothesis that development of acquired resis-tance to letrozole due to adaptive EGFR or HER2 upregula-tion could be prevented/delayed by dual targeting. However,there was no improvement in PFS for the combination inthese patients, which implies that in ER� endocrine–sensi-tive breast cancer specifıc co-targeting of HER2 (as opposedto EGFR) together with ER from the outset does not delayresistance. Indeed, this result is consistent with experimentalmodels that showed the failure of trastuzumab and letrozolecombined together from the outset to delay endocrine resis-tance in hormone receptor–positive xenografts, in contrastto combined therapy that was very effective once resistanceto letrozole had developed.9 As discussed below, this impor-tant lesson needs to be borne in mind when designing othercombined therapy trials in the fırst-line setting, as improvingon the effıcacy of existing endocrine therapy in this settingmay not be that straightforward if targeted blockade simplyallows other resistance mechanisms to evolve over a similartime frame to fırst-line aromatase inhibition.

In contrast to the relative disappointment of fırst-line stud-ies of endocrine therapy combined with EGFR/HER2-targeted therapies, a more successful approach has been theaddition of the mTOR antagonist everolimus to endocrinetherapy for postmenopausal women with ER�ve MBC whohave already received prior endocrine therapy. In these stud-ies, patients had often developed endocrine resistance, and assuch might be expected to have acquired either cell-survivalor adaptive-signaling pathways that could be susceptible toan appropriate therapeutic agent. Tamoxifen plus everoli-mus was investigated in patients with AI-resistant metastaticbreast cancer (MBC) in the phase II TAMRAD (tamoxifenplus everolimus) study and showed that the combinationtherapy gave an improvement in TTP (8.6 vs. 4.5 months),6-month CBR (61% vs. 42%) and median overall survival(OS) comparedwith tamoxifen alone.20 Importantly, the trialdesign included stratifıcation according to type of resistanceto previous treatment with AIs and showed that the greatestclinical benefıt for the combination arm occurred in patientswith acquired secondary resistance.These clinical data would support the hypothesis that tu-

mors that initially respond toAIs and then develop resistancemay utilize the PI3K/Akt/mTOR cell-survival pathway andthat a combined approach should be most effective in thosepatients with ER�ve advanced disease that progresses dur-ing or after nonsteroidal-aromatase-inhibitor therapy.21 Thiswas confırmed in the Breast Cancer Trials of Everolimus-2(BOLERO-2) study, a large randomized phase III trial thatassigned 724 postmenopausal patients with ER-positivemet-astatic breast cancer in a 2:1 ratio to either exemestane aloneor the combination of exemestane and everolimus.22 All pa-tients had progressed on a nonsteroidal AI, and importantly84% of them had demonstrated prior hormone-sensitive dis-ease defıned as “at least 24 months of endocrine therapy be-fore recurrence in the adjuvant setting, or a response orstabilization for at least 24 weeks of endocrine therapy for

TABLE 1. Randomized Clinical Trials of EFGF/HER2 Inhibitors Plus Endocrine Therapy to Overcome EndocrineResistance in MBC

Study Phase Arms n ORR%MedianTTP or PFS (mo)

MedianOS (mo)

Cristofanilli et al.16 II RCT anastrozole � placebo 50 12 8.4 nr

in ER� MBC anastrozole � gefitinib 43 2 14.7 nr

Osborne et al.15 II RCT tamoxifen � placebo 136 14.9 8.8 nr

in ER� MBC tamoxifen � gefitinib 153 12.4 10.9 nr

TAnDEM17 III RCT anastrozole 104 6.8 2.4 23.9

in ER� HER2� MBC anastrozole � trastuzumab 103 20.3* 4.8* 28.5

EGF3000819 III RCT letrozole � placebo 108 15 3 32.3

in ER� HER2� MBC letrozole � lapatinib 111 28* 8.2* 33.3

eLEcTRA18 II RCT letrozole 31 13 3.3 nr

in ER� HER2� MBC letrozole � trastuzumab 26 27 14.1 nr

Abbreviations: MO, months; NR, not reported; ORR, objective response rate; OS, overall survival; PFS, progression-free survival, RCT, randomized controlled trials; TTP, time-to-progression; ER,estrogen receptor; HER2, human epidermal growth factor receptor-2; MBC, metastatic breast cancer.* Statistically significant difference.



advanced disease.”22 In BOLERO-2, there was a statisticallysignifıcant and clinically relevant improvement in PFS forthe combination (median 7.8 vs. 3.2 months, HR � 0.45,p�0.0001).23 The clinical benefıt was primarily due to bettercontrol of the disease, although there was a signifıcant im-provement in tumor response rates from only 0.4% in theexemestane-alone group to 9.5% in the everolimus/exemes-tane group (p � 0.001).22A key question remains as to whether the combination of

an mTOR inhibitor with endocrine therapy will only be ef-fective for endocrine-resistant breast cancer, or whether thisis a new option for endocrine-sensitive MBC in the fırst-linesetting that could delay or prevent endocrine resistance de-veloping. A large fırst-line phase III study (HORIZON) re-cently reported the effıcacy for the oral mTOR antagonisttemsirolimus (30mg orally for 5 days every 2 weeks) in com-bination with letrozole vs. letrozole/placebo in 1,112 patientswith AI-naïve ER� advanced–breast cancer.24 In contrast toBOLERO-2, the population in this larger study was mainlytotally endocrine therapy naïve (approximately 60%), andhad received no prior AI therapy for locally advanced/meta-static disease. In HORIZON, there was no improvement inPFS overall (median 9 months, HR � 0.90, p � 0.25), or inthe 40% patient subset who had received prior adjuvant en-docrine therapy. These data suggest that as fırst-line therapythe combination may not be any better than an AI alone.

FUTURE STRATEGIES TO CO-TARGET ENDOCRINEAND OTHER SIGNALING PATHWAYSThe emergence of endocrine resistance during prolongedtherapy is complex, and it is unlikely that any single mecha-nism is operative. Although the EGFR/HER2 and mTORpathways have been studied extensively, numerous other sig-naling pathwaysmay also be implicated. Both preclinical andearly-phase clinical research is now trying to identify variousother strategies to overcome endocrine resistance, based onthe availability of targeted therapeutics that can be combinedwith endocrine therapy. Some of the key areas are discussedbelow, together with a list of current randomized trials thatare investigating various signal-transduction inhibitors incombination with endocrine therapy in ER� advanced–breast cancer (Table 2).

Agents Targeting PI3K/AKT/MTOR PathwayDespite the encouraging results reported for the combina-tion of everolimus plus exemestane in the BOLERO-2 trial, ata molecular level, two key regulatory loops may limit the ef-fectiveness of currentmTOR inhibitors. A negative-feedbackloop exists downstream in the PI3K/Akt/mTOR pathwaywhereby the mTOR activated kinase S6K1 phosphorylatesand de-stabilizes the IRS1 and IRS2 proteins in insulin-likegrowth factor (IGF) responsive cells.25 In these cells, mTORinhibition leads to a reduction in S6K1 activity, which inturn allows IRS1/2 expression to be increasedwith associated

enhanced activation of IGFR-1 dependent Akt activity.26,27Clinically phosphorylated Akt is upregulated in both tumorand skin biopsies of patients treated with everolimus,28 andas such this loss of negative feedback may overcome theantitumor effectiveness of mTOR blockade. In addition, apositive regulatory loop exists involving the mTORC2 com-plex that can be activated more directly by growth factorsand activated phosphorylated Akt. Therefore the inability ofsome rapamycin derivatives to block mTORC2 could resultin increased Akt signaling that result in ER phosphorylationon Serine 167 negating the effect of combined aromataseinhibition.29,30Although these two mechanisms may limit the benefıt of

the current generation of mTOR inhibitors in combinationwith endocrine therapy, several other drugs that target thePI3K/AKT pathway upstream of mTOR are currently beingtested in phase I/II trials in patients with advanced ER�breast cancer in the hope that may prove more specifıc andeffective than current mTOR inhibitors. These include panor isoform-specifıc PI3K inhibitors, dual PI3K/mTOR inhib-itors, and AKT inhibitors (Table 3). For example, BKM120 isa potent oral pan-PI3K inhibitor that when given either con-tinuously or intermittently in combinationwith letrozole in aphase I study has been demonstrated to be safe, with evidenceof antitumor effıcacy as assessed by FDG-PET scans.31 Thecombination of BKM120 with fulvestrant has also been in-vestigated, and a randomized phase III study of BKM120with fulvestrant in patients with HR�/HER2-negative lo-cally advanced/metastatic breast cancerwho have progressedafter prior AI therapy (BELLE-2) is recruiting a second-linepatient population very similar to that in the BOLERO-2 trial(NCT 00863655). Given the likely increased use of everoli-mus in combination with exemestane in the second-line set-ting, a further trial will assess the role of BKM-120 withfulvestrantwhohave progressed onor aftermTOR inhibitors(BELLE-3) (NCT01633060).Another approach is to develop drugs that target PI3K and

mTOR together, and two pharmaceutical companies haveset up studies comparing these dual inhibitors with pan-inhibitors of PI3K, both in combinationwith endocrine ther-apy compared with endocrine therapy alone. For exampleeither XL147 (inhibitor of PI3K) or XL765 (dual inhibitor ofPI3K and mTOR) are being combined with letrozole in aphase I/II trial (NCT01082068) in ER� advanced–breastcancer. Likewise, FERGI is a multicenter, international, ran-domized, double-blinded, placebo-controlled phase II trialrecruiting patients with advanced or MBC who have previ-ously received treatment with an AI, randomized to receiveeither GDC-0941 (pan PI3K inhibitor) � fulvestrant orGDC-0980 (dual inhibitor of PI3K and mTOR) � fulves-trant, or placebo � fulvestrant (NCT01437566). Whetherthese dual-targeted drugs are more effective than pan-isoform PI3K–inhibitors remains to be seen, together withearly assessments of toxicities that sometimes can be greaterfor drugs with broader target specifıcities.



Histone Deacetylase Inhibitors (HDACI)Another possible approach to reverse hormone resistance isthe use of histone deacetylase inhibitors (HDACI) to re-sensitize breast cancer cells to hormone manipulation. It hasbeen shown that in some breast cancers, expression of ER canbe repressed/lost by epigenetic modifıcations such as meth-ylation and histone deacetylation, and this could be a mech-anism for endocrine resistance. Entinostat is an HDACI thathas been shown to increase expression of both ER and theenzyme aromatase in a dose-dependentmanner both in vitroand in vivo, which then sensitized breast cancer cells to es-trogen and subsequent inhibition by the aromatase-inhibitorletrozole.32 Furthermore, in xenograft experiments, the com-bination of letrozole plus entinostat was signifıcantly moreeffective at inhibiting xenograft growth than either therapyalone. In a randomized phase II trial (ENCORE 301,NCT00676663), entinostat in combination with exemestanewas compared with exemestane/placebo in patients who hadreceived prior hormone therapy.33 This trial showed prolon-gation of median PFS (4.3 vs. 2.3 months) and extension ofOSbenefıt (26.9 vs. 19.8months), and a randomized phase III

trial is being planned. Similarly, a phase II study testing vori-nostat and tamoxifen in 43 patients with ER-positive–MBCprogressing on endocrine therapy showed an overall re-sponse rate (ORR) of 19% and amedian response duration of10.3 months.34 Correlative studies suggest that HDAC2 ex-pression could be a predictive biomarker, and that histonehyper-acetylationmay be a valid pharmaco-dynamicmarkerfor the effıcacy of this combination.

Antiangiogenic AgentsPreclinical data35 and retrospective clinical data36 suggestthat high vascular endothelial growth factor (VEGF) levels inbreast tumors are associatedwith a decreased response to en-docrine therapy. As several phase II studies had suggested thefeasibility and activity of the combination of bevacizumabwith endocrine agents,37,38 a randomized phase III study(LEA) was conducted to test the hypothesis that anti-VEGFtreatment with bevacizumab could prevent resistance to hor-mone therapy (either letrozole 2.5 mg/day or fulvestrant 250mg/4 weeks) given as fırst-line therapy in endocrine-respon-sive–advanced breast cancer.39 The PFS was better with the

TABLE 2. Randomized Clinical Trials Investigating Signal Transduction Inhibitors (STIs) Plus Endocrine Agents inMBC

Target Agent Stage and Study NumberEstimatedEnrollment (n pts)

PI3K/AKT/mTOR XL147 (inhibitor of PI3K) or XL765 (dual inhibitorof PI3K and mTOR) plus letrozole

Phase I/II (NCT01082068) 99

GDC-0941 � fulvestrant or GDC-0980 �fulvestrant or placebo � fulvestrant

Phase II (NCT01437566) 270

BKM120 (pan-PI3K inhibitor) plus fulvestrant vs.placebo plus fulvestrant

Phase III (NCT01633060) 615

MK-2206 (Akt inhibitor) plus anastrozole, orletrozole, or exemestane, or fulvestrant

Phase I (NCT01344031) 54

Histone deacetylase (HDAC) entinostat (SNDX-275) plus exemestane versusplacebo plus exemestane

Phase II (NCT00676663) 125

Vascular endothelial growth factor/angiogenesis Bevacizumab plus tamoxifen or letrozole vs.tamoxifen or letrozole alone


Bevacizumab plus letrozole or fulvestrant vs.letrozole or fulvestrant alone


BMS-690514 (inhibitor of EGFR, HER2, and VEGFreceptor kinases) plus letrozole versuslapatinib plus letrozole

Phase II (NCT01068704) 140

Proteasome (NF-kB pathway) Bortezomib plus fulvestrant vs. fulvestrantalone

Phase II (NCT01142401) 118

Src kinase Dasatinib plus fulvestrant vs. fulvestrant alone Phase II (NCT00754325) 100

Dasatinib plus exemestane vs. exemestanealone

Phase II (NCT00767520) 157

Fibroblast growth factor receptor (FGFR) AZD4547 plus fulvestrant vs. fulvestrant alone Phase I/II (NCT01202591) 120

Insulin-like growth factor type 1 (IGF-1) MEDI-573 (Dual IGF-I/II-neutralizing antibody)plus AI vs. AI Alone

Phase Ib/II (NCT01446159) 193

BMS-754807 plus letrozole vs. BMS-754807 alone Phase II (NCT01225172) 59

MM-121 plus exemestane vs. exemestane alone Phase II (NCT01151046) 130

Cyclin dependent kinase (CDK) 4/6 PD-0332991 plus letrozole vs. letrozole alone Phase I/II (NCT00721409) 177

PD-0332991 plus letrozole vs. placebo plusletrozole


Abbreviations: AI, aromatase inhibitor; mTOR, mammalian target of rapamycin; N, number; PI3K, phosphatidylinositol-3-kinase; PTS, patients.



combination of bevacizumab plus endocrine therapy thanwith endocrine monotherapy (18.4 vs. 13.8 months), but thiswas not statistically signifıcant. The combination had a sig-nifıcantly higher incidence of hematologic and nonhemato-logic toxicities and does not appear to be a promisingapproach to enhance fırst-line therapy.

Agents Targeting Src KinaseResults from preclinical studies showed that ER-Src kinaseaxis play an important role in promoting hormonal resis-tance by proto-oncogenes such as HER2, PELP1, and thatblocking this axis prevents the development of hormonal in-dependence in vivo.40 Dasatinib is a potent, broad spectrumATP–competitive inhibitor of Src tyrosine kinase. However,the addition of dasatinib to fulvestrant in a randomized phaseII study in ER-positive postmenopausal MBC patients whohad progressed after a NSAI did not improve PFS, CBR, orOS.41 Similarly, 157 patients were randomized in a double-blind phase II trial (CA180–261) to receive dasatinib (100mgdaily) or matched placebo in combination with exemestane(25 mg daily). Although the PFS difference was not signifı-cant in overall study population, a higher clinical benefıt rate(CBR) in the dasatinib arm and higher PFS in patients withsymptomatic bonemetastasis (HR� 0.68) suggested that da-satinib may have effıcacy in a subset of patients.42

Agents Targeting FGFR PathwaySeveral studies have shown that the fıbroblast growth factorreceptor-1 gene (FGFR1) is amplifıed in approximately 10%of all breast cancers, correlating with increased FGFR1mRNA or protein expression.43 Amplifıcation of FGFR1 is

enriched in up to 20% of ER-positive breast cancers. Ampli-fıcation and over-expression of FGFR1may be a major con-tributor to poor prognosis in luminal-type B breast cancers,driving anchorage-independent proliferation and endocrine-therapy resistance.43 AZD4547 is a potent selective inhibitorof FGFR-1, 2, and 3 receptor tyrosine kinases (enzyme andcellular phosphorylation endpoints) and has a signifıcantlylower potency for inhibition of IGF1R and KDR.44 The co-administration of an FGFR inhibitor and exemestane has thepotential to improve outcome for patients with aggressivedisease or resistance to endocrine therapy. Therefore,GLOWis a randomized double-blind phase IIa study (with phase Icombination safety run-in) designed to assess the safety andeffıcacy of AZD4547 in combination with exemestane vs. ex-emestane alone in patients with ER-positive and FGFR1 am-plifıed (FISH � 4) breast cancer who have failed treatmentwith one prior endocrine therapy (adjuvant or fırst-line met-astatic) (NCT01202591).

Agents Targeting Insulin-Like Growth Factor Type 1(IGF-1)The role of the insulin-like growth factor (IGF) system inendocrine-resistant breast cancer has been studied, and in-hibitors of this pathway are currently in clinical trials inER�ve patients who have progressed on prior endocrinetherapy. MEDI-573 is a dual-targeting human antibody thatneutralizes IGF-I/-II ligands and inhibits insulin-like growthfactor receptor 1 (IGF-1R) and insulin receptor-A (IR-A) sig-naling pathways, which play a role in breast and other epithe-lial cancers. By sparing insulin receptor-B (IR-B) and itshybrid receptors, MEDI-573 is expected to achieve antitu-mor activity without perturbing glucose homeostasis and hasshowed acceptable safety and favorable PK profıles withoutsignifıcant changes in glucose levels.45 A biomarker-richphase Ib/II study of MEDI-573 with an aromatase inhibitorin patients with advanced ER�ve breast cancer is ongoing(NCT01446159).

Inhibitor of Cyclin-Dependent Kinase (CDK) 4/6Modulating the cell cycle has always been an attractive ther-apeutic target in cancer, and previously published data havesuggested that CDK 4/6 inhibition may play a key role in thetreatment of subsets of breast cancers.46,47 PD 0332991 is anovel oral selective inhibitor of cyclin-dependent kinase(CDK) 4/6, which prevents cellular DNA synthesis by block-ing cell-cycle progression fromG1 to S phase. Recently, it wasreported that the combination of PD 0332991 and letrozolesignifıcantly improvedmedian PFS in a randomized phase IIstudy in patients with advanced ER-positive–breast cancer,including those with identifıed cyclin D1 amplifıcationand/or p16 loss in whom CDK 4/6 inhibition is expected tobe most effective.48 A PFS of 26.1 month was observed forpatients in the combination arm compared with 7.5 monthsfor patients treated with letrozole alone. In patients withmeasurable disease, an improved response rate was seen(45% compared with 31%), and the toxicity profıle for thecombination was favorable with the most common adverse

TABLE 3. Novel Agents under InvestigationTargeting the PI3K/AKT/mTOR Pathway

Target(s) Drug Pharmaceutical Company

PI3K� BYL719 Novartis

PI3K� GDC-0032 Genentech

PI3K� MLN-1117 Millenium

PI3Kd CAL-101 Calistoga

Pan-PI3K XL-147 Exelixis/Sanofi

Pan-PI3K BKM120 Novartis

Pan-PI3K GDC-0941 Genentech

Pan-PI3K PKI-587 Pfizer

PI3K/mTOR XL-765 Exelixis/Sanofi

PI3K/mTOR BEZ235 Novartis

PI3K/mTOR GDC-0980 Genentech

PI3K/mTOR PF-4691502 Pfizer

TORC1/2 MLN-128 Millenium

TORC1/2 OSI-027 OSI Pharma

TORC1/2 AZD2014 AstraZeneca

AKT (catalytic) AZD5363 AstraZeneca

AKT (allosteric) MK-2206 Merck

AKT (catalytic) GDC-0068 Genentech



events being (uncomplicated) neutropenia, leukopenia, ane-mia, and fatigue. On the basis of this extremely promisingresult, a randomized, multicenter, double-blind fırst-linestudy of PD-0332991 plus letrozole comparedwith letrozole/placebo in postmenopausal women with ER� HER2- MBCwho have not received any prior systemic anticancer treat-ment for advanced disease will open to recruitment soon.(NCT01740427).As such, CDK 4/6 inhibition seems one of the more prom-

ising approaches to enhance endocrine response in ER� en-docrine-sensitive–breast cancer and could potentiallyproduce that quantum leap in response to fırst-line endo-crine therapy that to date has eluded this area of clinical re-search in ER� advanced breast cancer. However,appropriate clinical trial design, patient selection, and bio-marker research remains crucial to enhancing the chance ofsuccess.

ISSUES FOR FUTURE CLINICAL TRIAL DESIGN INER� BREAST CANCERIt is becoming clear that improving endocrine therapy by theaddition of targeted therapy is not that simple.49 Undoubt-edly, the signifıcant effıcacy for the combination of themTOR antagonist everolimus and exemestane in those pa-tients refractory to prior AI therapy is a major advance inproviding greater clinical benefıt compared with the use ofjust further endocrine therapy alone for these patients, whichmay spare the use of palliative chemotherapy for a period oftime. As for identifying untreated ER� MBC patients thatwould benefıt from a given combined targeted and endocrinetherapy in the fırst-line setting, this appears much tricki-er—so far it would appear that “blind” co-treatment of hor-mone sensitive ER� MBCwith a given drug combination inthe hope of delaying resistance and improving the benefıt al-ready obtained with fırst-line AI therapy does not work. It ispossible that some ER� tumors are inherently primed to re-spond to a given combination, as shown by expression ofPIK3CA mutations in a neoadjuvant study of letrozole andeverolimus,50 or co-expression ofHER2 andER in the studiescited above (Table 1). However, the challenge in daily clinicalpractice is to identify the relevant pathways that are operativein individual patients with ER�ve MBC, which may or maynot benefıt from a given targeted-therapy combination.Although overexpression of any given oncogene or mo-

lecular target could identify the best group of ER� patientsto treat with any given novel-targeted therapeutic in a com-bination strategy to enhance endocrine responsiveness, itremains likely that only a proportion will gain benefıt be-cause of other coexisting mutations and molecular altera-tions within the complex web of inter-related signalingnetworks that will determine response/resistance to inhibi-tion of any key target (e.g., either loss of the PTEN tumorsuppressor gene or activating mutation of PI3-kinase may

modulate response to trastuzumab in HER2�ve breast can-cer). As such, translational studies in real-time metastaticsamples from patients will remain crucial for optimizingclinical benefıt from these new therapies, whether it is theoriginal primary tumor or the relapsed sample where themolecular profıle may have changed.In future, all clinical studies shouldmake a greater effort to

enrich their trial population with the most appropriate pa-tients, with a tumor molecular profıle likely to benefıt fromtargeting the given pathway. In the absence of the ability toidentify the molecular target in samples of metastatic disease(i.e., only the original primary tumor is available, as often isthe case), understanding the inherent biology in the primarytumor that accounts for relapse/resistance may become cru-cial is determining which population to select for these stud-ies. Genomic profıling in ER�ve breast cancer may helpidentify those more likely to develop resistance to endocrinetherapy, or indeed the pathways that these tumors are mostlikely to utilize as escape mechanisms, which in turn mayguide appropriate selection of target therapies to add in at thetime of relapse. Recent studies have started to show commononcogenic pathways that intrinsic subtypes of breast cancerwill utilize, thus allowing strategies to be developed for com-binations of various signaling agents to be used in an attemptto enhance responsiveness to current therapies.51 In particu-lar, in ER�ve breast cancer, gene expression profıling hasidentifıed that in the luminal B subtype activation of growthfactor signaling pathways occurs, often independent ofHER2 overexpression, thus contributing to their poorerprognosis.52 Selection of this subgroup for future combina-tion strategies may yield answers faster than treating a moreheterogeneous unselected group of patients with ER� breastcancer.

CONCLUSIONThere are now amultitude of targeted therapeutics in variousstages of clinical development for breast cancer (Table 2) thatare based on a rationale that the target in question is valid inthe pathogenesis of ER� breast cancer. Although the inte-gration of targeted therapies with conventional therapeuticsin breast cancer has been pioneered by the combination oftrastuzumab with chemotherapy, substantial research is on-going into combining targeted therapeutics with endocrinetherapy to enhance responsiveness and delay resistance.13The emphasis is now shifting toward targeting networks andpathways with combinations of signaling drugs, either inparallel (so-called combined horizontal blockade) or in series(combined vertical blockade). Selection of which approach isvalid depends on key preclinical studies that need to be un-dertaken in various relevant models, in order to guide whichcombinations need testing in early phase clinical trials. Thiswill ensure rapid and effıcient transition from proof of con-cept studies into pivotal effıcacy studies, thusmaximizing thelikelihood of success.





Employment or Leadership Position: None. Consultant or Advisory Role: Stephen R. Johnston, AstraZeneca; GlaxoSmithKline; Roche/Genentech. StockOwnership: None. Honoraria: Stephen R. Johnston, GlaxoSmithKline; Novartis. Research Funding: None. Expert Testimony: None. Other Remuneration:None.

References

1. Tamoxifen for early breast cancer: an overview of the randomised trials.Early Breast Cancer Trialists’ Collaborative Group. Lancet. 1998;351:1451-1467

2. Ring A, Dowsett M. Mechanisms of tamoxifen resistance. Endocr RelatCancer. 2004;11:643-658.

3. Ali S, Coombes RC. Endocrine-responsive breast cancer and strategiesfor combating resistance. Nat Rev Cancer. 2002;2:101-112.

4. Osborne CK, Schiff R. Estrogen-receptor biology: continuing progressand therapeutic implications. J Clin Oncol. 2005;23:1616-1622.

5. Miller TW, Balko JM, Arteaga CL. Phosphatidylinositol 3-kinase andantiestrogen resistance in breast cancer. J Clin Oncol. 2011;29:4452-4461.

6. ShimWS,ConawayM,Masamura S, et al. Estradiol hypersensitivity andmitogen-activated protein kinase expression in long-term estrogen de-prived human breast cancer cells in vivo. Endocrinology. 2000;141:396-405.

7. StephenRL, ShawLE, LarsenC, et al. Insulin-like growth factor receptorlevels are regulated by cell density and by long term estrogen depriva-tion in MCF7 human breast cancer cells. J Biol Chem. 2001;276:40080-40086.

8. Campbell RA, Bhat-Nakshatri P, Patel NM, et al. Phosphatidylinositol3-kinase/AKT-mediated activation of estrogen receptor alpha: a newmodel for anti-estrogen resistance. J Biol Chem. 2001;276:9817-9824.

9. SabnisG, SchavowitzA,GoloubevaO, et al. Trastuzumab reverses letro-zole resistance and amplifıes the sensitivity of breast cancer cells to es-trogen. Cancer Res. 2009;15;69(4):1416-1428.

10. Martin LA, Farmer I, Johnston SR, et al. Enhanced estrogen receptor(ER) alpha, ERBB2, and MAPK signal transduction pathways operateduring the adaptation ofMCF-7 cells to long term estrogen deprivation.J Biol Chem. 2003;278:30458-30468.

11. deGraffenried LA, Friedrichs WE, Russell DH, et al. Inhibition ofmTOR activity restores tamoxifen response in breast cancer cells withaberrant Akt Activity. Clin Cancer Res. 2004;10:8059-8067.

12. Boulay A, Rudloff J, Ye J, et al. Dual inhibition of mTOR and estrogenreceptor signaling in vitro induces cell death inmodels of breast cancer.Clin Cancer Res. 2005;11:5319-5328.

13. Johnston SR. Enhancing the effıcacy of hormonal agents with selectedtargeted agents. Clin Breast Cancer. 2009;9(suppl 1:S28-36).

14. Massarweh S, Osborne CK, Creighton CJ, et al. Tamoxifen resistance inbreast tumors is driven by growth factor receptor signaling with repres-sion of classic estrogen receptor genomic function.Cancer Res. 2008;68:826-833.

15. Osborne CK, Neven P, Dirix LY, et al. Gefıtinib or placebo in combina-tion with tamoxifen in patients with hormone receptor positive meta-static breast cancer: a randomized phase II study.Clin Cancer Res. 2011;17:1147-1159.

16. Cristofanilli M, Valero V, Mangalik A, et al. Phase II, randomized trialto compare anastrozole combined with gefıtinib or placebo in post-

menopausal women with hormone receptor-positive metastatic breastcancer. Clin Cancer Res. 2010;16:1904-1914.

17. Kaufman B, Mackey JR, Clemens MR, et al. Trastuzumab plus anastro-zole versus anastrozole alone for the treatment of postmenopausalwomen with human epidermal growth factor receptor 2-positive, hor-mone receptor-positive metastatic breast cancer: results from the ran-domized phase III TAnDEM study. J Clin Oncol. 2009;27:5529-5537.

18. Huober J, Fasching PA, BarsoumM, et al. Higher effıcacy of letrozole incombination with trastuzumab compared to letrozole monotherapy asfırst-line treatment in patients with HER2-positive, hormone-receptor-positive metastatic breast cancer - results of the eLEcTRA trial. Breast.2012;21:27-33.

19. Johnston S, Pippen J Jr, Pivot X, et al. Lapatinib combinedwith letrozoleversus letrozole and placebo as fırst-line therapy for postmenopausalhormone receptor-positive metastatic breast cancer. J Clin Oncol. 2009;27:5538-5546.

20. Bachelot T, Bourgier C, Cropet C, et al. Randomized phase II trial ofeverolimus in combination with tamoxifen in patients with hormonereceptor-positive, human epidermal growth factor receptor 2-negativemetastatic breast cancer with prior exposure to aromatase inhibitors: aGINECO study. J Clin Oncol. 2012;30:2718-2724.

21. Rugo HS, Keck S. Reversing hormone resistance: have we found thegolden key? J Clin Oncol. 2012;30:2707-2709.

22. Baselga J, Campone M, Piccart M, et al. Everolimus in postmenopausalhormone-receptor-positive advanced breast cancer.NEngl JMed. 2012;366:520-529.

23. Piccart M. Final Progression-free survival analysis of BOLERO-2: aphase III trial of everolimus for postmenopausal women with advancedbreast cancer [SABCS abstract P6-04-02]. Cancer Res. 2012;72(suppl24).

24. Wolff AC, Lazar AA, Bondarenko I, et al. Randomized phase IIIplacebo-controlled trial of letrozole plus oral temsirolimus as fırst-lineendocrine therapy in postmenopausal women with locally advanced ormetastatic breast cancer. J Clin Oncol. 2013;31:195-202.

25. Sabatini DM. mTOR and cancer: insights into a complex relationship.Nat Rev Cancer. 2006;6:729-734.

26. O’Reilly KE, Rojo F, She QB, et al. mTOR inhibition induces upstreamreceptor tyrosine kinase signaling and activates Akt. Cancer Res. 2006;66:1500-1508.

27. Sun SY, Rosenberg LM,Wang X, et al. Activation of Akt and eIF4E sur-vival pathways by rapamycin-mediated mammalian target of rapamy-cin inhibition. Cancer Res. 2005;65:7052-7058.

28. Tabernero J, Rojo F, Calvo E, et al. Dose- and schedule-dependent in-hibition of the mammalian target of rapamycin pathway with everoli-mus: a phase I tumor pharmacodynamic study in patients withadvanced solid tumors. J Clin Oncol. 2008;26:1603-1610.

29. Sarbassov DD, Guertin DA, Ali SM, et al. Phosphorylation and regula-



tion of Akt/PKB by the rictor-mTOR complex. Science. 2005;307:1098-1101.

30. Jacinto E, Loewith R, Schmidt A, et al.Mammalian TOR complex 2 con-trols the actin cytoskeleton and is rapamycin insensitive. Nat Cell Biol.2004;6:1122-1128.

31. Mayer I. SU2C phase Ib study of pan-PI3K inhibitor BKM120 withletrozole in ER�/HER2- metastatic breast cancer (MBC). J Clin Oncol.2012;30 (suppl; abstr 510).

32. Sabnis G, Goloubeva O, Chuman S, et al. Functional activation of theestrogen receptor-alpha and aromatase by the HDAC inhibitor, enti-nostat, sensitizes ER negative tumors to letrozole. Cancer Res. 2011;71(5):1893-1903.

33. KleinP.Characterization of the overall survival benefıt in ENCORE301,a randomized placebo-controlled phase II study of exemestanewith andwithout entinostat in ER� postmenopausal women with metastaticbreast cancer, ASCO Annual Meeting, J Clin Oncol. 2012;30 (suppl 27;abstr 128).

34. Munster PN, Thurn KT, Thomas S, et al. A phase II study of the histonedeacetylase inhibitor vorinostat combined with tamoxifen for the treat-ment of patients with hormone therapy-resistant breast cancer. Br JCancer. 2011;104:1828-1835.

35. De la Haba J: Combined antiangiogenic and anti-estrogen therapy inbreast cancer. Molecular mechanisms involved [SABC abstract P2-05-07]. Cancer Res. 2011;71(suppl):24.

36. Manders P, Beex LV, Tjan-Heijnen VC, et al. Vascular endothelialgrowth factor is associated with the effıcacy of endocrine therapy in pa-tients with advanced breast carcinoma. Cancer. 2003;98:2125-2132.

37. Yardley DA, Burris HA 3rd, Clark BL, et al. Hormonal therapy plus be-vacizumab in postmenopausal patients who have hormone receptor-positive metastatic breast cancer: a phase II Trial of the Sarah CannonOncology Research Consortium. Clin Breast Cancer. 2011;11:146-152.

38. Traina TA, Rugo HS, Caravelli JF, et al. Feasibility trial of letrozole incombination with bevacizumab in patients with metastatic breast can-cer. J Clin Oncol. 2010;28:628-633.

39. Martin M: Phase III trial evaluating the addition of bevacizumab to en-docrine therapy as fırst-line treatment for advanced breast cancer: fırsteffıcacy results from the LEA study [SABC S1-7]. Cancer Res. 2012;72(suppl):24.

40. Vallabhaneni S, Nair BC, Cortez V, et al. Signifıcance of ER-Src axis inhormonal therapy resistance. Breast Cancer Res Treat. 2011;130:377-385.

41. Wright GL: Randomized phase II trial of fulvestrant with or withoutdasatinib in postmenopausal patients with hormone receptor-positivemetastatic breast cancer previously treated with an aromatase inhibitor[SABC PD01-01]. Cancer Res. 2011;71(suppl):24.

42. Llombart A: Randomized phase II study of dasatinib vs placebo in ad-dition to exemestane in advanced ER/PR-positive breast cancer [BMSCA180-261 Study] [SABC PD01-02]. Cancer Res. 2011;71(suppl):24.

43. Turner N, Pearson A, Sharpe R, et al. FGFR1 amplifıcation drives endo-crine therapy resistance and is a therapeutic target in breast cancer.Can-cer Res. 2010;70:2085-2094.

44. Gavine PR,Mooney L, Kilgour E, et al. AZD4547: an orally bioavailable,potent, and selective inhibitor of the fıbroblast growth factor receptortyrosine kinase family. Cancer Res. 2012;72:2045-2056.

45. Haluska P: MEDI-573 as a novel approach to IGF-1R and IR-A signal-ing inhibition by blocking IGF ligands: phase I PK/PD, safety data, anddisease linkage studies in breast cancer. J Clin Oncol. 2011;29 (suppl 27;abstr 271).

46. Finn RS, Dering J, Conklin D, et al. PD 0332991, a selective cyclin Dkinase 4/6 inhibitor, preferentially inhibits proliferation of luminal es-trogen receptor-positive human breast cancer cell lines in vitro. BreastCancer Res. 2009;11:R77.

47. Dean JL, McClendon AK, Hickey TE, et al. Therapeutic response toCDK4/6 inhibition in breast cancer defıned by ex vivo analyses of hu-man tumors. Cell Cycle. 2012;11:2756-2761.

48. Finn RS: Results of a randomized phase 2 study of PD 0332991, a cyclin-dependent kinase (CDK) 4/6 inhibitor, in combinationwith letrozole vsletrozole alone for fırst-line treatment of ER�/HER2- advanced breastcancer (BC) [SABC S1-6]. Cancer Res. 2012;72(suppl):24.

49. Dees EC, Carey LA. Improving endocrine therapy for breast cancer: it’snot that simple. J Clin Oncol. 2013;31:171-173.

50. Baselga J, Semiglazov V, van Dam P, et al. Phase II randomized study ofneoadjuvant everolimus plus letrozole compared with placebo plusletrozole in patients with estrogen receptor-positive breast cancer. J ClinOncol. 2009;27:2630-2637.

51. Bild AH, Parker JS, Gustafson AM, et al. An integration of complemen-tary strategies for gene-expression analysis to reveal novel therapeuticopportunities for breast cancer. Breast Cancer Res. 2009;11:R55.

52. Loi S, Sotiriou C, Haibe-Kains B, et al. Gene expression profıling iden-tifıes activated growth factor signaling in poor prognosis (Luminal-B)estrogen receptor positive breast cancer. BMC Med Genomics. 2009;2:37.



Endocrine Resistance: What Do We Know?Todd W. Miller, PhD

OVERVIEW

Adjuvant therapy with antiestrogens targeting estrogen receptor � (ER) signaling prevents disease recurrence in many patients withearly-stage ER� breast cancer. However, a significant number of cases exhibit de novo or acquired endocrine resistance. While otherclinical subtypes of breast cancer (HER2�, triple-negative) have disproportionately higher rates of mortality, ER� breast cancer isresponsible for at least as many deaths because it is the most common subtype. Therefore, identifying mechanisms that drive endocrineresistance is a high clinical priority. A large body of experimental evidence indicates that oncogenic signaling pathways underlieendocrine resistance, including growth factor receptor tyrosine kinases (HER2, epidermal growth factor receptor [EGFR], fibroblastgrowth factor receptor 1/2 [FGFR], insulin-like growth factor-1 receptor [IGF-1R]/ insulin receptor [InsR]), PI3K/AKT/ mTOR, MAPK/ERK,Src, CDK4/CDK6, and ER itself. Combined targeting of ER and such pathways may be the most effective means to combat antiestrogenresistance, and clinical trials testing such strategies show promising results. Herein, we discuss pathways associated with endocrineresistance, biomarkers that may be useful to predict response to targeted agents, and avenues for further exploration to identifystrategies for the treatment of patients with endocrine-resistant disease.

More than 75% of breast cancers express estrogen re-ceptor � (ER) and/or progesterone receptor (PR, an

ER-regulated gene product). These cancers comprise the so-called “ER�” subtype, and the number of ER� breast can-cers is projected to increase.1 Such patients are typicallytreated with antiestrogen therapies that inhibit ER function.There are 3 main types of anti-estrogens: (1) selective ERmodulators (SERMs) such as tamoxifen (which has mixedagonist/antagonist activity); (2) selective ER downregulators(SERDs) such as fulvestrant; (3) agents that reduce circulat-ing estrogen levels such as aromatase inhibitors (AIs; letro-zole, anastrozole, exemestane). Such endocrine therapies aresome of the most effective targeted anticancer agents in his-tory. Most patients who succumb to ER� breast cancer weretreated at some point with an antiestrogen,2-4 so their diseaseis likely endocrine resistant. It therefore remains a clinicalpriority to understand mechanisms driving endocrine-resistance in ER� breast cancer. Herein, the discussion ofendocrine resistance refers to cancers that grow in the pres-ence of an antiestrogen; cancers that recur or progress aftercessation of antiestrogen therapy may not be endocrine-resistant, and such patients may benefıt from continued en-docrine therapy.4ER� breast cancers can exhibit de novo or acquired endo-

crine resistance. Common clinical scenarios of endocrine re-sistance include: (1) de novo resistance where a patientpresents with metastatic disease resistant to all hormonaltherapies, or recurs soon after starting adjuvant hormonal

therapy and does not respond to further endocrine manipu-lation; (2) de novo resistance to some but not other hormonaltherapies; (3) acquired resistance after initial response toendocrine therapy, followed by temporary response to addi-tional hormonal therapies until the cancer becomes refrac-tory to all endocrine agents; (4) progression after initialresponse to a hormonal agent, and then transient response tothe same agent introduced years later. Preclinical evidencesuggests that the cellular andmolecularmechanisms govern-ing de novo versus acquired resistance may be the same, andthat mechanisms of resistance to different classes of anties-trogens are similar. However, such fındingsmay be biased byour ER�models. Instances of cancers that respond to the AIexemestane following progression on a nonsteroidal AI(letrozole or anastrozole), or that respond to fulvestrant fol-lowing progression on an AI, support the existence of agent-specifıc and class-specifıc types of endocrine resistance.

GROWTH FACTOR RECEPTOR SIGNALING PATHWAYSPROMOTE ENDOCRINE RESISTANCEThe only clinically available marker of antiestrogen resis-tance isHER2overexpression.5 The fındings thatHER2over-expression promotes the agonistic effects of tamoxifen, andthat AI therapy is more effective than tamoxifen in patientswith ER�/HER2� disease prompted a switch to non-SERMtherapies for these patients.6,7 However, less than 10% ofER� breast cancers overexpress HER2, and HER2 is viewed

From the Department of Pharmacology and Toxicology and Norris Cotton Cancer Center, Geisel School of Medicine at Dartmouth, Lebanon, NH.


Corresponding author: Todd W. Miller, PhD, Dartmouth-Hitchcock Medical Center, One Medical Center Dr., HB-7936, Lebanon, NH 03756; email: [email protected]


ENDOCRINE RESISTANCE: WHAT DO WE KNOW?


as clinically more important than ER because HER2� breastcancer is generally more aggressive. Thus, such patients areoften treated with HER2-directed (trastuzumab) and endo-crine therapies, either sequentially or in combination. Retro-spective clinical data suggest that endocrine therapy isbenefıcial in combinationwith anti-HER2 therapy8, but theseobservations await confırmation in a prospective study.Preclinical studies have implicated additional growth fac-

tor receptor signaling pathways in endocrine resistance.EGFR, IGF-1R, InsR, Ron, and FGFR1 activation promoteanti-estrogen resistance inmodel systems.9-13 Such receptorsconverge on the PI3K/AKT/mTOR and MEK/ERK path-ways, which have also been implicated in antiestrogen resis-tance.7,10,13,14 Components of these pathways are oftengenomically altered in human cancers.15,16 Whether mecha-nisms of resistance to tamoxifen, fulvestrant, and AIs arecommon remains to be determined, but activation of thePI3K or MEK pathways confers resistance to all forms of en-docrine therapy. Many of these fındings are supported bycorrelative, retrospective clinical evidence. For example, pa-tients with FGFR1-overexpressing ER� tumors had shorterdistant recurrence-free survival following adjuvant treat-ment with tamoxifen compared with patients with FGFR1-normal tumors.11 Patients with ER� tumors exhibiting a(phospho)protein signature of PI3K hyperactivation exhib-ited shorter recurrence-free survival following adjuvant en-docrine therapy compared with patients with PI3K-lowtumors.13 Similarly, patients with ER� tumors exhibiting agene expression signature of IGF-1R activation had a worseprognosis compared with patients with IGF-1R-inactive tu-mors.17 Such fındings have led to ongoing trials testing novelagents targeting these signaling pathways in combinationwith antiestrogens.

ER PROMOTES ENDOCRINE RESISTANCE:CROSSTALK BETWEEN ER AND GROWTH FACTORRECEPTOR PATHWAYSThe best-characterized mechanism of ER signaling involvesestrogen-induced dimerization and phosphorylation of ERto promote transcriptional activity. However, growth factorreceptor signaling pathways can modulate ER activation.The PI3K effector AKT, the TORC1 effector p70S6K, andERK can phosphorylate ER to promote estrogen-induced,tamoxifen-induced, and ligand-independent ER transcrip-tional activity.18-20 PI3K, Ras, and ERK can also promote ac-tivation of ER cofactors (Fig. 1).21-26 In turn, ER drivesexpression of genes encoding growth factor receptor path-way components. Neoadjuvant estrogen deprivation therapywith an AI reduced AKT and mTOR activation in ER� tu-mors in patients, and such reductions correlated with im-proved clinical response and outcome, suggesting thatestrogen-induced signaling activates these pathways.27,28While clinical fındings imply that antiestrogens should

suppress growth factor receptor signaling, preclinical evi-dence shows that such crosstalk is more complex. ForcedPI3K pathway activation (by knockdown of PTEN, or over-expression HER2, IGF-1R, or activated AKT1) confers resis-tance to tamoxifen, fulvestrant, and estrogen deprivation inER� breast cancer cell lines. Such resistance is typically re-versible by inhibition of PI3K.7,18,29 ER� MCF-7 breast can-cer xenografts with acquired resistance to tamoxifen exhibitincreased expression of IGF-1R, HER2, and EGFR.30 AI-resistant MCF-7/aromatase cells and xenografts exhibit in-creased levels of HER2 and the EGFR ligand amphiregulin,and lower levels of ER.31,32 Long-term estrogen-deprived,ER� breast cancer cell lines (which model AI resistance) ex-hibit hyperactivation of IGF-1R/InsR/PI3K/AKT/mTORsignaling.13MCF-7 cells and xenografts with acquired fulves-trant resistance shown similar changes.10,33 At fırst glance,such fındings seem to conflict with clinical observations.27,28However, cells may escape estrogen/ER dependence via up-regulation of growth factor receptor pathways by an alterna-tive mechanism(s) (i.e., independent of ER). The majority ofwell-characterized breast cancer cell lines are ER-, so gener-ation of more ER� models may help capture the heteroge-neity of ER� breast cancers. Patient-derived ER� breastcancer xenografts may also present clinically-relevant mod-els to study endocrine resistance.While patient-derived ER�tumors do not graft well in immunodefıcient mice, a recentlyderived immunodefıcient knock-in mouse expressing hu-man prolactin (hPRL) has an ER� tumor graft success rate of43.1% and is expected to increase the capacity to generatesuch models.34Preclinical and clinical evidence implicate ER itself in

endocrine resistance. Most breast cancers that progresson antiestrogen therapy retain ER. Following progressionon an AI, patients are often switched to fulvestrant, whicheffectively inhibits and partially downregulates ER.35,36 Ap-proximately 30% of patients who progress on an AI respondto second-line fulvestrant.37,38 High-dose fulvestrant may

KEY POINTS

� Over 75% of breast cancers express estrogen receptor �(ER). Patients with such tumors are treated withantiestrogens to block ER signaling. Many patients exhibitantiestrogen-resistant disease, making ER� breast cancera significant contributor to breast cancer mortality.

� Growth factor receptor signaling pathways that activatePI3K/AKT/mTOR promote antiestrogen resistance.

� Cell-cycle deregulation that circumvents the requirementfor estrogen receptor � (ER) confers antiestrogenresistance. This may occur in the form of CDK4/CDK6hyperactivation or Rb loss.

� ER itself can drive endocrine-resistant cell proliferation,and ER downregulators may be superior to estrogen-deprivation therapy with aromatase inhibitors (AIs).

� Drugs targeting the PI3K/AKT/mTOR and CDK4/CDK6pathways are being tested in ongoing trials in patients withER� breast cancer.

TODD W. MILLER


provide a longer time-to-progression than estrogen depriva-tion with the AI anastrozole as fırst-line treatment for ad-vanced breast cancer.39 In ER� cell lines with acquiredresistance to estrogen deprivation, ER remains transcription-ally active, and treatment with fulvestrant or knock-down ofER expression inhibits cell growth.40 These data suggest thatER may remain active under estrogen-depleted (AI-treated)conditions, and that further inhibition/downregulation ofER (with fulvestrant) may be superior to AI therapy.

COMBINED TARGETING OF ER AND GROWTHFACTOR RECEPTOR PATHWAYS ABROGATESENDOCRINE RESISTANCEPI3K/AKT/mTOR signaling is required for growth ofendocrine-resistant breast cancer cell lines. The estrogen-independent growth of long-term estrogen-deprived cells isinhibited by the PI3K/mTOR inhibitor BEZ235, or theTORC1 inhibitor everolimus (Afınitor, now approved fortreatment of advanced ER� breast cancer in combinationwith the AI exemestane).13 Similarly, the pan-PI3K inhibitorbuparlisib (BKM120) inhibits estrogen-independent growthof MCF-7 xenografts in mice, and the PI3K/mTOR inhibitorwortmannin inhibits growth of letrozole-resistant MCF-7/aromatase xenografts.40-42 Estrogen stimulation blocks theapoptotic effects of BEZ235 in ER� cells, suggesting thatcombined blockade of ER and PI3K may be most effec-tive.43 Indeed, combined treatment with fulvestrant and bu-parlisib induced regression ofMCF-7 xenografts, and fulves-trant plus the IGF-1R/InsR inhibitor OSI-906 completely

inhibited tumor growth, while single-agent treatments onlyslowed growth.12,40

CELL CYCLE DEFECTS PROMOTE ANTIESTROGENRESISTANCEDysregulation of cell cycle checkpoints is common in cancer.Since ER regulates the expression of many genes involved incell cycle progression, and antiestrogens block such ER func-tions, antiestrogen resistance can arise by genetic alterationsthat circumvent the requirement for ER. A commonly dereg-ulated checkpoint involves the Cyclin-D/CDK4/CDK6/Rbpathway. Cyclin-D1/CDK4 and Cyclin-D3/CDK6 com-plexes phosphorylate Rb family proteins, thereby promotingactivation of E2F transcription factors to drive expression ofgenes encoding proteins required for cell cycle progression.Genes encoding Cyclin-D1, Cyclin-D3, CDK4, and CDK6are amplifıed, and the Rb tumor suppressor is lost or muta-tionally inactivated in many cancers. Rb loss confers tamox-ifen resistance in ER� models.44 Patients with ER� breastcancer exhibiting a gene expression signature of Rb loss hadshorter recurrence-free survival following adjuvant tamox-ifen.44 A tumor gene expression signature of E2F activationwas associated with higher residual tumor cell proliferationfollowing presurgical AI therapy.40 Therefore, activation ofthe CDK4/CDK6/E2F axis promotes endocrine resistance,and treatment with a CDK4/6 inhibitor or knockdown ofCDK4 expression abrogates endocrine-resistant cell prolifer-ation.40,45 For reasons that remain unclear, ER� breast can-cer cell lines are much more sensitive to CDK4/6 inhibition

FIG 1. Reciprocal crosstalk between ER� and growth factor receptor signaling pathways. RTKs and GPCRsactivate the PI3K (blue) and MEK signaling pathways. These signal transducers can then phosphorylate ER(green) and/or coactivators and corepressors to modulate ER transcriptional activity not necessarily dependenton ER ligands. In turn, ER transcribes genes encoding components of growth factor signaling pathways, thuscompleting a signaling cycle of RTKs to ER to RTKs. ER also complexes with RTKs and Src to rapidly inducenongenomic signaling. ER-interacting proteins are shown in color. Figure was modified from Miller andcolleagues14 and reprinted with permission (Copyright 2011, American Society of Clinical Oncology).



than ER- cell lines.46 In a recent phase II trial, the CDK4/6inhibitor PD-0332991 in combination with the AI letrozoleextended progression-free survival compared with letrozolealone as fırst-line treatment for metastatic ER�/HER2-breast cancer.47 This drug combination is being tested in aphase III study.

BIOMARKERS OF RESPONSE TO TARGETEDAGENTS (?)Following the identifıcation of growth factor receptor path-ways as causes of endocrine resistance, much effort was ex-pended to identify biomarkers predictive of sensitivity todrugs targeting the various nodes of these pathways in orderto identify patient subpopulations thatmay reap greater ben-efıt from treatment. Such biomarkers were typically identi-fıed by screening large panels of cell lines, and searching forgenetic or proteomic markers enriched in sensitive or resis-tant lines. The most mature targeted agent to combat endo-crine resistancemediated by growth factor receptor signalingis everolimus, but biomarkers associated with everolimussensitivity in breast cancer remain unclear.Mutations in PIK3CA, the gene encoding the p110� cata-

lytic subunit of PI3K, occur in 28-47%of ER� breast cancers,and are associated with sensitivity to PI3K and AKT inhibi-tors in cancer cell lines.13,16,48-50 Thus, PIK3CA was a logicalbiomarker to select patients for inclusion in early-phase clin-ical trials with PI3K/AKT inhibitors. However, this geneticbiomarker has not been evaluated in the context of PI3K in-hibition with endocrine therapy. Early clinical data showedthat among 8 patients with advanced ER� breast cancer whoexhibited a response (by [18F]FDG-PET) to buparlisib plusan AI, only two patients had PIK3CA-mutant tumors.51 Inanother study considering patients enrolled in phase I trialstesting PI3K/AKT/mTOR inhibitors, there was a higher re-sponse rate in tumors containing a PIK3CA mutation com-pared with those with wild-type PIK3CA (30% vs. 10%).52These data imply that (1) a signifıcant fraction of PIK3CA-wild-type tumors can respond to PI3K inhibition, and (2) abetter biomarker of response is needed.Genetic biomarkers are easier tomeasure andmore reliable

in archived tissue than (phospho)protein biomarkers. Whileit has been considered that the degree of PI3K pathway acti-vation may predict sensitivity to pathway inhibition, the lev-els of phospho-AKT were not associated with sensitivity toAKT inhibition in cancer cell lines.50 Furthermore, PIK3CAmutations are not associatedwith PI3Kpathway activation incell lines as assessed by AKT phosphorylation.53

The most appropriate biomarker of sensitivity to PI3Kpathway inhibitors may be pharmacodynamic. If a tumor issensitive to a drug, the tumor should exhibit a rapid meta-bolic response. This concept is being pursued in ongoingclinical trials testing novel agents. Early clinical data suggestthat a metabolic response to buparlisib plus an AI as assessedby [18F]FDG-PET after two weeks of therapy is associatedwith longer time-on-study, and thus improved response, inpatients with metastatic disease.54 With such a pharmacody-namic biomarker, each patient serves as their own control,and drug response can be assessed within weeks of initiationof therapy. A similar strategy is being tested in the presurgi-cal/neoadjuvant setting, where changes in molecular mark-ers (e.g., phospho-S6 as a marker of TORC1 activation) aredetected by comparing pre- and post-treatment tumor tissues.In contrast to PI3K pathway biomarkers, cancer cell sensi-

tivity to CDK4/6 inhibitors may be more predictable usinggenetic biomarkers. Genomic loss of Rb renders a cell insen-sitive to CDK4/6 inhibition,55,56 so patients with Rb-defıcienttumors should not be treated with a CDK4/6 inhibitor. Am-plifıcation of the gene encodingCyclin-D1, or loss of the geneencoding p16INK4A, may be indicative of cells that have hy-peractivated CDK4/6 signaling. Whether these genetic le-sions are associated with greater sensitivity to CDK4/6inhibition is being tested in the trials with letrozole � PD-0332991.

CONCLUSIONOver twodecades of researchhas led to the implication of twomajor signaling axes in endocrine resistance: (1) growth fac-tor receptor/PI3K/AKT/mTOR, and (2) CDK4/CDK6/Rb/E2F. Drugs targeting these pathways are being testedclinically in combinationwith antiestrogens, andhave shownpromising results thus far. A looming concern is that inher-ent biases in the model systems used to identify mechanismsof endocrine resistance guided pathway identifıcation. Forexample, there is an over-representation of PIK3CA muta-tions among ER� breast cancer cell lines compared withER� tumors. IfPIK3CAmutations confer sensitivity to PI3Kinhibitors, most ER� breast cancer cell lines will likely re-spond. However, many ER� breast tumors are PIK3CA-wild-type, and available ER� models to study this genotypeare few. With the establishment of more ER� breast cancercell lines and patient-derived tumor xenografts, it is hopedthat we will establish a broader genotypic repertoire of ER�breast cancermodels, and begin to identify novel drug targetsfor endocrine-resistant cancers.



Employment or Leadership Position: None. Consultant or Advisory Role: Todd W. Miller, Pfizer. Stock Ownership: None. Honoraria: None. ResearchFunding: Todd W. Miller, Piramal Life Science. Expert Testimony: None. Other Remuneration: None.

TODD W. MILLER


References

1. Anderson WF, Katki HA, Rosenberg PS. Incidence of Breast Cancer inthe United States: Current and Future Trends. J Natl Cancer Inst. 2011;103:1397-1402.

2. Early Breast Cancer Trialists’ Collaborative Group. Effects of chemo-therapy and hormonal therapy for early breast cancer on recurrence and15-year survival: an overview of the randomised trials. Lancet. 2005;365:1687-1717.

3. Forbes JF, Cuzick J, Buzdar A, et al. Effect of anastrozole and tamoxifenas adjuvant treatment for early-stage breast cancer: 100-month analysisof the ATAC trial. Lancet Oncol. 2008;9:45-53.

4. Davies C, Pan H, Godwin J, et al: Long-term effects of continuing adju-vant tamoxifen to 10 years versus stopping at 5 years after diagnosis ofoestrogen receptor-positive breast cancer: ATLAS, a randomised trial.Lancet. Epub 2012 Dec 5.

5. Arpino G, Green SJ, Allred DC, et al. HER-2 amplifıcation, HER-1 ex-pression, and tamoxifen response in estrogen receptor-positive metastaticbreast cancer: a southwest oncology group study. Clin Cancer Res. 2004;10:5670-5676.

6. Ellis MJ, Tao Y, Young O, et al. Estrogen-independent proliferation ispresent in estrogen-receptor HER2-positive primary breast cancer afterneoadjuvant letrozole. J Clin Oncol. 2006;24:3019-3025.

7. Shou J, Massarweh S, Osborne CK, et al. Mechanisms of tamoxifen re-sistance: increased estrogen receptor-HER2/neu cross-talk in ER/HER2-positive breast cancer. J Natl Cancer Inst. 2004;96:926-935.

8. Hayashi N, Niikura N, Yamauchi H, et al. Adding hormonal therapy tochemotherapy and trastuzumab improves prognosis in patients withhormone receptor-positive and human epidermal growth factor recep-tor 2-positive primary breast cancer. Breast Cancer Res Treat. 2013;137:523-531.

9. McClaine RJ, Marshall AM, Wagh PK, et al. Ron receptor tyrosine ki-nase activation confers resistance to tamoxifen in breast cancer celllines. Neoplasia. 2010;12:650-658.

10. FrogneT, BenjaminsenRV, Sonne-HansenK, et al. Activation of ErbB3,EGFR and Erk is essential for growth of human breast cancer cell lineswith acquired resistance to fulvestrant. Breast Cancer Res Treat. 2009;114:263-275.

11. Turner N, Pearson A, Sharpe R, et al. FGFR1 amplifıcation drives endo-crine therapy resistance and is a therapeutic target in breast cancer.Can-cer Res. 2010;70:2085-2094.

12. FoxEM,Miller TW,Balko JM, et al. A kinome-wide screen identifıes theinsulin/IGF-I receptor pathway as a mechanism of escape from hor-mone dependence in breast cancer. Cancer Res. 2011;71:6773-6784.

13. Miller TW, Hennessy BT, Gonzalez-Angulo AM, et al. Hyperactivationof phosphatidylinositol-3 kinase promotes escape from hormone de-pendence in estrogen receptor-positive human breast cancer. J Clin In-vest. 2010;120:2406-2413.

14. Kurokawa H, Lenferink AE, Simpson JF, et al. Inhibition of HER2/neu(erbB-2) andmitogen-activated protein kinases enhances tamoxifen ac-tion against HER2-overexpressing, tamoxifen-resistant breast cancercells. Cancer Res. 2000;60:5887-5894.

15. Miller TW, Balko JM, Arteaga CL. Phosphatidylinositol 3-kinase andantiestrogen resistance in breast cancer. J Clin Oncol. 2011;29:4452-4461.

16. Miller TW, Rexer BN, Garrett JT, et al. Mutations in the phosphatidyl-inositol 3-kinase pathway: Role in tumor progression and therapeuticimplications in breast cancer. Breast Cancer Res. 2011;13:224.

17. Creighton CJ, Casa A, Lazard Z, et al. Insulin-like growth factor-I acti-vates gene transcription programs strongly associated with poor breastcancer prognosis. J Clin Oncol. 2008;26:4078-4085.

18. Campbell RA, Bhat-Nakshatri P, Patel NM, et al. Phosphatidylinositol3-kinase/AKT-mediated activation of estrogen receptor alpha: a newmodel for anti-estrogen resistance. J Biol Chem. 2001;276:9817-9824.

19. Yamnik RL, Digilova A, Davis DC, et al. S6 kinase 1 regulates estrogenreceptor alpha in control of breast cancer cell proliferation. J Biol Chem.2009;284:6361-63699.

20. Kato S, Endoh H, Masuhiro Y, et al. Activation of the estrogen receptorthrough phosphorylation by mitogen-activated protein kinase. Science.1995;270:1491-1494.

21. Coso OA, Chiariello M, Yu JC, et al. The small GTP-binding proteinsRac1 and Cdc42 regulate the activity of the JNK/SAPK signaling path-way. Cell. 1995;81:1137-1146.

22. Minden A, Lin A, Claret FX, et al. Selective activation of the JNK signal-ing cascade and c-Jun transcriptional activity by the small GTPases Racand Cdc42Hs. Cell. 1995;81:1147-1157.

23. Logan SK, Falasca M, Hu P, et al. Phosphatidylinositol 3-kinase medi-ates epidermal growth factor-induced activation of the c-JunN-terminal kinase signaling pathway.Mol Cell Biol. 1997;17:5784-5790.

24. HibiM, Lin A, Smeal T, et al. Identifıcation of an oncoprotein- andUV-responsive protein kinase that binds andpotentiates the c-Jun activationdomain. Genes Dev. 1993;7:2135-2148.

25. Derijard B, Hibi M, Wu IH, et al. JNK1: a protein kinase stimulated byUV light andHa-Ras that binds and phosphorylates the c-Jun activationdomain. Cell. 1994;76:1025-1037.

26. Massarweh S, Schiff R. Unraveling the mechanisms of endocrine resis-tance in breast cancer: new therapeutic opportunities. Clin Cancer Res.2007;13:1950-1954.

27. Generali D, Fox SB, Brizzi MP, et al. Down-regulation of phosphatidyl-inositol 3�-kinase/AKT/molecular target of rapamycin metabolic path-way by primary letrozole-based therapy in human breast cancer. ClinCancer Res. 2008;14:2673-2680.

28. Baselga J, Semiglazov V, van Dam P, et al. Phase II randomized study ofneoadjuvant everolimus plus letrozole compared with placebo plusletrozole in patients with estrogen receptor-positive breast cancer. J ClinOncol. 2009;27:2630-2637.

29. Miller TW, Perez-Torres M, Narasanna A, et al. Loss of Phosphataseand Tensin homologue deleted on chromosome 10 engages ErbB3 andinsulin-like growth factor-I receptor signaling to promote antiestrogenresistance in breast cancer. Cancer Res. 2009;69:4192-4201.

30. Massarweh S, Osborne CK, Creighton CJ, et al. Tamoxifen resistance inbreast tumors is driven by growth factor receptor signaling with repres-sion of classic estrogen receptor genomic function.Cancer Res. 2008;68:826-833.

31. SabnisG, SchayowitzA,GoloubevaO, et al. Trastuzumab reverses letro-zole resistance and amplifıes the sensitivity of breast cancer cells to es-trogen. Cancer Res. 2009;69:1416-1428.

32. Wang X, Masri S, Phung S, et al. The role of amphiregulin inexemestane-resistant breast cancer cells: evidence of an autocrine loop.Cancer Res. 2008;68:2259-2265.

33. Massarweh S, Osborne CK, Jiang S, et al. Mechanisms of tumor regres-sion and resistance to estrogen deprivation and fulvestrant in amodel ofestrogen receptor-positive, HER-2/neu-positive breast cancer. CancerRes. 2006;66:8266-8273.

34. Rui H, Utama, FE, Yanac, AF, et al. Prolactin-humanized mice: an im-proved animal recipient for therapy-response-testing of patient-derivedbreast cancer xenotransplants [SABCS1-8].Cancer Res. 2012;72(suppl):24.

35. Robertson JFR, Dixon, JM, Sibbering, DM, et al. Tumor biomarkerchanges following pre-surgical treatment with 500 mg fulvestrant plus



anastrozole versus 500mg fulvestrant alone and 1mg anastrozole alone[SABC 24]. Cancer Res. 2009;69(suppl):24.

36. McCormack P, Sapunar F. Pharmacokinetic profıle of the fulvestrantloading dose regimen in postmenopausal women with hormonereceptor-positive advanced breast cancer. Clin Breast Cancer. 2008;8:347-351.

37. Perey L, Paridaens R, Hawle H, et al. Clinical benefıt of fulvestrant inpostmenopausal women with advanced breast cancer and primary oracquired resistance to aromatase inhibitors: fınal results of phase IISwiss Group for Clinical Cancer Research Trial (SAKK 21/00).AnnOn-col. 2007;18:64-99.

38. Ingle JN, Suman VJ, Rowland KM, et al. Fulvestrant in women withadvanced breast cancer after progression on prior aromatase inhibitortherapy: North Central Cancer Treatment Group Trial N0032. J ClinOncol. 2006;24:1052-6.

39. Robertson JF, Llombart-Cussac A, Rolski J, et al. Activity of fulvestrant500 mg versus anastrozole 1 mg as fırst-line treatment for advancedbreast cancer: results from the FIRST study. J Clin Oncol. 2009;27:4530-4535.

40. Miller TW, Balko JM, Fox EM, et al. ER�-dependent E2F transcriptioncan mediate resistance to estrogen deprivation in human breast cancer.Cancer Discov. 2011;1:338-351.

41. Voliva CF, Pecchi, S, Burger, M, et al. Biological characterization ofNVP-BKM120, a novel inhibitor of phosphoinosotide 3-kinase in PhaseI/II clinical trials [AACR abstract 4498]. Cancer Res. 2010;70(suppl):8.

42. Sabnis G, Goloubeva O, Jelovac D, et al. Inhibition of the phosphatidyl-inositol 3-kinase/Akt pathway improves response of long-termestrogen-deprived breast cancer xenografts to antiestrogens. Clin Can-cer Res. 2007;13:2751-2757.

43. Crowder RJ, Phommaly C, Tao Y, et al. PIK3CA and PIK3CB inhibitionproduce synthetic lethality when combined with estrogen deprivationin estrogen receptor-positive breast cancer. Cancer Res. 2009;69:3955-3962.

44. Bosco EE, Wang Y, Xu H, et al. The retinoblastoma tumor suppressormodifıes the therapeutic response of breast cancer. J Clin Invest. 2007;117:218-228.

45. Thangavel C, Dean JL, Ertel A, et al. Therapeutically activating RB: re-establishing cell cycle control in endocrine therapy-resistant breast can-cer. Endocr Relat Cancer. 2011;18:333-345.

46. Finn RS, Dering J, Conklin D, et al. PD 0332991, a selective cyclin Dkinase 4/6 inhibitor, preferentially inhibits proliferation of luminal es-trogen receptor-positive human breast cancer cell lines in vitro. BreastCancer Res. 2009;11:R77.

47. Finn RS, Crown JP, Lang I, et al. Results of a Randomized Phase 2 Studyof PD 0332991, a Cyclin-Dependent Kinase (CDK) 4/6 Inhibitor, inCombination with Letrozole vs Letrozole Alone for First-Line Treat-ment of ER�/HER2-Advanced Breast Cancer (BC).Cancer Res. 72:91s,2012.

48. Maira SM, Pecchi S, Huang A, et al. Identifıcation and characterizationof NVP-BKM120, an orally available pan-class I PI3-kinase inhibitor.Mol Cancer Ther. 2012;11:317-328.

49. Tanaka H, Yoshida M, Tanimura H, et al. The selective class I PI3Kinhibitor CH5132799 targets human cancers harboring oncogenicPIK3CA mutations. Clin Cancer Res. 2011;17:3272-3281.

50. She QB, Chandarlapaty S, Ye Q, et al. Breast tumor cells with PI3Kmu-tation or HER2 amplifıcation are selectively addicted to Akt signaling.PLoS ONE. 2008;3:e3065.

51. Sangai T, Akcakanat A, Chen H, et al. Biomarkers of response to AktinhibitorMK-2206 in breast cancer.Clin Cancer Res. 2012;18:5816-5828.

52. Mayer IA, Balko JM, KubaMG, et al. SU2C phase Ib study of pan-PI3Kinhibitor BKM120 plus aromatase inhibitor letrozole in ER�/HER2-metastatic breast cancer [SABC abstract PD09-05]. Cancer Res. 2011;71(suppl):24.

53. Janku F, Wheler JJ, Westin SN, et al. PI3K/AKT/mTOR inhibitors inpatients with breast and gynecologic malignancies harboring PIK3CAmutations. J Clin Oncol. 2012;30:777-782.

54. Stemke-Hale K, Gonzalez-Angulo AM, Lluch A, et al. An integrativegenomic andproteomic analysis of PIK3CA,PTEN, andAKTmutationsin breast cancer Cancer Res. 2008;68:6084-6091.

55. Mayer IA, Abramson VG, Balko JM, et al. SU2C phase Ib study of pan-PI3K inhibitor BKM120with letrozole in ER�/HER2-metastatic breastcancer (MBC). J Clin Oncol. 2012;30(suppl: abstr 510).

56. Dean JL, Thangavel C, McClendon AK, et al: Therapeutic CDK4/6 in-hibition in breast cancer: key mechanisms of response and failure. On-cogene. Epub 2010 May 17.

57. Fry DW, Harvey PJ, Keller PR, et al. Specifıc inhibition of cyclin-dependent kinase 4/6 by PD 0332991 and associated antitumor activityin human tumor xenografts.Mol Cancer Ther. 2004;3:1427-1438.

TODD W. MILLER


· adjuvant chemotherapy: which patient? what regimen? natalie turner, mbbs, laura biganzoli, md,...

Documents