Cancer after Thyroidectomy: A Multi-InstitutionalExperience with 1,523 Patients

J Joshua Smith, MD, PhD, Xi Chen, PhD, David F Schneider, MD, James T Broome, MD, FACS,Rebecca S Sippel, MD, FACS, Herbert Chen, MD, FACS, Carmen C Solorzano, MD, FACS

BACKGROUND: The incidence of thyroid cancer in patients treated operatively for thyroid disease has beenhistorically low (<5%). Previous series have not specifically addressed cancer rates in botheuthyroid and hyperthyroid patients. This study examined cancer frequency in patientsreferred for removal of benign thyroid disease in a multi-institutional series.

STUDY DESIGN: A total of 2,551 patients underwent thyroidectomy at 3 high-volume institutions. Indeter-minate/malignant fine-needle aspiration diagnosis was excluded (n ¼ 1,028). Cancer caseswere compared among 1,523 patients with Graves’ disease (n ¼ 264), nodular goiter(n ¼ 1,095), and toxic nodular goiter (n ¼ 164). Fisher’s exact test, chi-square test, Wilcoxonrank sum, Kruskal-Wallis nonparametric t-tests, and multivariable logistic regression wereused.

RESULTS: Overall, 238 (15.6%) cancers were recorded: Graves’ disease (6.1%), nodular goiter (17.5%),and toxic nodular goiter (18.3%). Cancer rates were significantly different among thesegroups (p < 0.01) and significantly higher in nodular goiter and toxic nodular goiter vsGraves’ disease (p < 0.01); no significant differences in cancer rates were noted among insti-tutions. Overall, 275 patients had thyroiditis (18%). There was a significant association withyounger age, male sex, nodular thyroids, and cancer (p < 0.05). Presence of thyroiditis orperformance of preoperative fine-needle aspiration was not associated with cancer. Meancancer size was 1.1 cm (46% >0.5 cm; 39% >1 cm). Most patients underwent totalthyroidectomy (80%).

CONCLUSIONS: These data confirm higher than expected incidental thyroid cancer rates (15.6%) in the largestmulti-institutional surgical series to date. Nodular thyroids, males, and young patients weremore likely to harbor incidental carcinoma. These data support consideration of initial totalthyroidectomy as the preferred approach for patients referred to the surgeon with bilateralnodular disease. (J Am Coll Surg 2013;216:571e579. � 2013 by the American College ofSurgeons)

In the United States, where iodine deficiency is not a majorfactor, multinodular thyroid disease, chronic autoimmune(Hashimoto) thyroiditis, and Graves’ disease are commoncauses of goiter. Patients are generally referred to the

Disclosure Information: Nothing to disclose.

Presented at the Southern Surgical Association 124th Annual Meeting,Palm Beach, FL, December 2012.

Received December 7, 2012; Accepted December 11, 2012.From the Department of Surgery (Smith), Division of Surgical Oncologyand Endocrine Surgery (Broome, Solorzano), Vanderbilt-Ingram CancerCenter (X Chen, Broome, Solorzano), and Department of Biostatistics(X Chen), Vanderbilt University Medical Center, Nashville, TN, Depart-ment of Surgery, Division of General Surgery, University of WisconsinSchool of Medicine and Public Health, Madison, WI (Schneider, Sippel,H Chen), and Division of Endocrine Surgery, University of Miami MillerSchool of Medicine, Miami, FL (Solorzano).Correspondence address: Carmen C Solorzano, MD, FACS, Department ofSurgery,VanderbiltUniversityMedicalCenter, 597PrestonBldg, 2220PierceAve, Nashville, TN 37232-6860. email: carmen.solorzano@vanderbilt.edu

571ª 2013 by the American College of Surgeons

Published by Elsevier Inc.

surgeon for the management of cancerous or suspiciousthyroid nodules, obstructive nontoxic or toxic nodulargoiters, and toxic diffuse goiters. The risk of thyroid cancerwithin a nodular goiter being evaluated by ultrasound andfine-needle aspiration (FNA) is thought to be between 3%and 5%,1 yet recent studies suggest that this approach failsto truly exclude cancer at this rate in patients referred to thesurgeon for presumably benign thyroid disease.2-5

An Italian study from 2006 noted a 10% cancer rate in998 patients referred for surgical management ofunequivocal benign toxic (TNG) and nontoxic multinod-ular thyroid (MNG) disease.2 All patients had undergonepreoperative ultrasonography and FNA of suspicious ordominant thyroid nodules, with exclusion of such casesfrom the analysis. A more recent study (2012) reporteda 21.7% cancer rate in 1,161 patients undergoingthyroidectomy for MNG,5 and yet another report from

ISSN 1072-7515/12/$36.00

http://dx.doi.org/10.1016/j.jamcollsurg.2012.12.022

Abbreviations and Acronyms

FNA ¼ fine-needle aspirationMNG ¼ nontoxic multinodular goiterOR ¼ odds ratioTNG ¼ toxic nodular goiterTSH ¼ thyroid-stimulating hormoneVUMC ¼ Vanderbilt University Medical Center

572 Smith et al Incidental Cancer Rate after Thyroidectomy J Am Coll Surg

the United States found this incidence to be slightly lower(16%).4 In addition, the reported studies are from singleinstitutions and have not consistently addressed cancerrates in both euthyroid and hyperthyroid patients referredto the surgeon for the management of presumed benignthyroid disease. Thyroid cancer rates have been shownto be rising for tumors of all sizes.6 An updated knowl-edge of the cancer rate in presumed benign thyroidprocesses will help inform the risk and benefit discussionof surgical treatment with patients.The purpose of this study was to examine the incidence

of unsuspected (incidental) cancer in patients referred tothe surgeon for removal of benign thyroid disease ina multi-institutional setting.

METHODSA total of 2,551 patients at 3 institutions (2000 to 2011)were referred for surgical treatment of thyroid disease(Fig. 1). We excluded 1,028 patients who had a knownmalignant or indeterminate diagnosis by FNA. Thisincluded the cytology diagnoses of malignant, suspiciousfor malignancy, follicular lesion or atypia of undeter-mined significance, and follicular neoplasms. The remain-ing 1,523 patients were evaluated for the presence ofthyroid cancer on final pathology. These cancers are

Figure 1. Study design of patients referred to the surgeon for treatmVanderbilt University Medical Center (VUMC), University of Wisconsin (UWfor surgical treatment of thyroid disease. We excluded 1,028 patientaspiration (FNA) biopsy. MNG, nontoxic multinodular goiter; TMNG ¼ to

referred to as incidental cancers in the remainder of thisarticle. All patients underwent preoperative thyroid ultra-sound. Preoperative FNA of dominant/suspicious thyroidnodules was performed at the discretion of the surgeon orreferring physician. Diagnosis of Graves’ disease wasestablished by the referring endocrinologist using clinicalsigns and symptoms, thyroid scan, thyroid hormone andthyroid-stimulating hormone (TSH) assays, and/ormeasurement of thyrotropin receptor antibodies. A uni-or multinodular toxic goiter was diagnosed in patientspresenting with a nodular thyroid and suppressed TSHwith or without overt elevation of T3 and/or T4.Radio-iodine uptake scans were obtained in the majorityof toxic patients at the discretion of the referring physi-cian or treating surgeon. A thyroid was considerednontoxic and multinodular when it had 2 or morenodules and the patient had normal or elevated TSH.Thyroiditis was diagnosed biochemically by the presenceof thyroid peroxidase antibodies and clinical and patho-logic evidence of the disease.Clinical information was extracted in a retrospective

fashion from prospectively maintained databases. Historyof external beam radiation or family history of thyroidcancer was not available. Institutional Review Boardapproval was obtained at each site. The 3 institutionsincluded the Departments of Surgery from VanderbiltUniversity Medical Center (VUMC, n ¼ 691; 2007 to2011), the University of Wisconsin School of Medicineand Public Health (n ¼ 551; 2000 to 2010), andthe University of Miami Miller School of Medicine(n ¼ 281; 2003 to 2008).Fisher’s exact test, chi-square test, and Wilcoxon rank

sum and Kruskal-Wallis nonparametric t-tests wereused. Additionally, we applied a multivariable logistic

ent of benign thyroid disease. Three institutional databases from), and University of Miami (UM) were evaluated for patients referred

s with known malignant or indeterminate diagnosis by fine-needlexic multinodular goiter.

Vol. 216, No. 4, April 2013 Smith et al Incidental Cancer Rate after Thyroidectomy 573

regression model to test the association of age, sex, preop-erative FNA, lymphocytic thyroiditis, and nodularity(ie, TNG or MNG) with the development of incidentalthyroid cancer as the outcomes variable. Odds ratioswere then estimated based on an exponentiated regressioncoefficient for the predictor variables.

RESULTSOverall rate of cancer in 1,523 patients treated forpresumed benign thyroid disease in a multi-institutionalsetting was 15.6% (n ¼ 238). Specifically for each insti-tution, we noted the following cancer rates: 15.9%(VUMC); 14.2% (University of Wisconsin School ofMedicine and Public Health); and 17.4% (University ofMiami Miller School of Medicine). There was no statis-tically significant difference in cancer rates among theinstitutions (p ¼ 0.48). Mean cancer size was 1.1 cm(range 0.1 to 9 cm) with 46% and 39% of cancers�0.5 cm and 1 cm, respectively. We did note a significantdifference in the mean incidental cancer size for theMiami group compared with the VUMC or Wisconsingroups (Fig. 2; 1.44 cm vs 0.93 cm or 0.99 cm, respec-tively; p ¼ 0.02). There were no differences in sizes ofincidental cancers according to clinical presentation(Graves’ disease vs TNG vs MNG; p ¼ 0.21). Themajority of cancers were papillary thyroid carcinoma(n ¼ 175 [73.5%]) with the follicular variant of papillary(n ¼ 39 [16.4%]), 11 follicular cancers (4.6%), and13 “other” pathologies (5.5%, eg, Hurthle cell,lymphoma, squamous cell) constituting the remainder

Figure 2. Mean incidental thyroid cancer size per institution. Meanincidental thyroid cancer size was significantly different betweeninstitutions. Vanderbilt University Medical Center (VUMC), 0.93 cmvs University of Wisconsin (Wisconsin), 0.99 cm vs University ofMiami (Miami), 1.44 cm (p ¼ 0.02).

of the cases. Most patients underwent total or near totalthyroidectomy (80%).More granular details about the incidental cancers were

available from one institution (VUMC). The majority ofcancers were stage I (84%), 10% were stage II, and 6%were stage III or higher. Multifocal tumors were presentin 30%. Radio-iodine ablation was performed in 32%of patients. Mean follow-up was 26 months and atlast follow-up there were no detected recurrences ormortalities.Demographic and clinical factors were evaluated to

determine any significant associations with cancer. Medianage at presentation for surgical treatment was 49 years and84.3% of patients were female. Overall, 275 patients hadlymphocytic thyroiditis (18%). Six hundred and sixty-onepatients underwent preoperative FNA (43%). In the over-all univariate analysis, no significant cancer association wasnoted with age, performance of preoperative FNA orlymphocytic thyroiditis (Table 1). A significant associationwith male sex and cancer was noted (Table 1). A higherproportion of patients with nodular disease was observedin the group of patients with cancer (Table 1; 93% vs81%; p < 0.0001).On multivariate analysis (Table 2), the association with

male sex and cancer remained significant (odds ratio[OR] ¼ 0.587; 95% CI, 0.411e0.849; p ¼ 0.004) andyoung age became significant (OR ¼ 0.988; 95% CI,0.978e0.999; p ¼ 0.02). These data suggest that foreach 1-year increase in age, we expect to see an approxi-mately 1.2% decrease in the odds of cancer developing.

Table 1. Demographic and Clinical Factors Associatedwith the Presence or Absence of Incidental Thyroid Cancer in1,523 Patients

Benign(n ¼ 1,285)

Cancer(n ¼ 238) p Value

Mean age, y 49.1 48.6 0.49

Sex, n (%) 0.03

Female 1,095 (85) 189 (79)

Male 190 (15) 49 (21)

Preoperative FNA, n (%) 0.14

Yes 547 (43) 114 (48)

No 738 (57) 124 (52)

Thyroiditis,* n (%) 0.46

Yes 228 (18) 47 (20)

No 1,056 (82) 191 (80)

Nodular disease,y n (%) <0.0001

Yes 1,037 (81) 222 (93)

No 248 (19) 16 (7)

*One patient with missing data for thyroiditis.yNodular disease defined as patients with toxic nodular goiter and nontoxicmultinodular goiter.FNA, fine-needle aspiration.

Table 2. Results of a Multivariable Logistic RegressionModel to Test the Association of Age, Sex, PreoperativeFine-Needle Aspiration, Lymphocytic Thyroiditis, and NodularThyroid Disease with Incidental Thyroid Cancer (N ¼ 1,523)

Odds ratio 95% CI p Value

Age 0.988 0.978e0.999 0.02

Sex 0.587 0.411e0.849 0.004

Preoperative FNA 1.01 0.752e1.349 0.96

Lymphocytic thyroiditis 1.23 0.849e1.744 0.27

Nodular disease 4.1 2.373e7.431 <0.00001

FNA, fine-needle aspiration.

Table 3. Demographic and Clinical Factors Associatedwith the Presence or Absence of Incidental Thyroid Canceron Final Pathology in Patients with Multinodular Goiter, ToxicNodular Goiter, and Graves’ Disease

Benign Cancer p Value

Multinodular goiter, n 903 192

Mean age, y 51.8 49.3 0.02

Sex, n (%) 0.06

Females 775 (86) 154 (80)

Males 128 (14) 38 (20)

Preoperative FNA, n (%) >0.99

Yes 496 (55) 106 (55)

No 407 (45) 86 (45)

Thyroiditis,* n (%) 0.40

Yes 154 (17) 38 (20)

No 748 (83) 154 (80)

Toxic nodular goiter, n 134 30

Mean age, y 48.5 50 0.65

Sex, n (%) 0.03

Females 119 (89) 22 (73)

Males 15 (11) 8 (27)

Preoperative FNA, n (%) 0.37

Yes 36 (27) 11 (37)

No 98 (73) 19 (63)

Thyroiditis, n (%) 0.77

Yes 20 (15) 3 (10)

No 114 (85) 27 (90)

Graves’ disease, n 248 16

Mean age, y 38.6 40.3 0.86

Sex, n (%) >0.99

Females 202 (81) 13 (81)

Males 46 (19) 3 (19)

Preoperative FNA, n (%) >0.99

Yes 18 (7) 1 (6)

No 230 (93) 15 (94)

Thyroiditis, n (%) 0.21

Yes 54 (22) 6 (38)

No 194 (78) 10 (62)

*One patient missing with data for thyroiditis in nontoxic multinodulargoiter.FNA, fine-needle aspiration.

574 Smith et al Incidental Cancer Rate after Thyroidectomy J Am Coll Surg

Additionally, the odds of cancer developing in females areapproximately 41% lower than the odds for males. Lastly,nodular thyroid disease was significantly associated withincidental thyroid cancer (OR ¼ 4.1; 95% CI,2.373e7.431; p < 0.0001), indicating that the oddsfor developing cancer are approximately 307% higherin nodular disease compared with that observed in theGraves’ disease patients.

Subset analysis of multinodular goiter, toxicnodular goiter, and Graves’ disease patients

Cancer rates according to clinical presentation were asfollows: Graves’ disease (6.1%), MNG (17.5%), andTNG (18.3%). Cancer rates were significantly differentamong these groups (p < 0.01) and significantly higherin MNG and TNG vs Graves’ disease (p < 0.01).When the clinical indication for operation was MNG,patients with cancer were younger than those withoutcancer (Table 3; 49.3 vs 51.8 years; p ¼ 0.02). We noteda higher proportion of males in the group of MNGpatients with cancer; however, this did not reach statis-tical significance (Table 3; 20% vs 14%; p ¼ 0.06). Inaddition, there was no significant association with eitherpreoperative FNA or presence of lymphocytic thyroiditisand cancer (Table 3). The characterization of the patientswith TNG was recently described by our group in a sepa-rate publication3 and showed a significantly higherproportion of males in patients with cancer (Table 2;27% vs 11%; p ¼ 0.03). Lastly, for the patients under-going thyroidectomy for Graves’ disease, there was nosignificant association with age, sex, preoperative FNA,or thyroiditis and cancer (Table 3).

DISCUSSIONThis is the first multi-institutional study to demonstratea higher than expected rate of incidental thyroid cancerin patients referred to the surgeon for presumed benigndisease. In this study, the overall cancer incidence was15.6% and there was no statistically significant differencein cancer rates among institutions. This cancer rate was at

least 1.4 times higher than recently reported in the liter-ature (Table 4).2,7,8 Incidental cancer rates according toclinical indication for thyroidectomy were as follows:6.1% for Graves’ disease, 17.5% for MNG, and 18.3%for TNG. Cancer rates were significantly different amongthese groups and significantly higher in MNG and TNGvs Graves’ patients.A recent study by Luo and colleagues4 evaluated 838

patients with MNG and found younger age, male sex,and smaller nodules were predictors of incidental thyroid

Table 4. Selected Surgical Series Reporting Incidental Thyroid Cancer Rates in Patients undergoing Thyroidectomy forBenign Thyroid Disease

First author, year

Current studyCerci, 20078 Bradly, 20097 Luo, 20124 Miccoli, 20062 Lasithiotakis, 20125

Treatment period 2001e2005 2000e2008 1994e2009 2002e2003 1990e2008 2000e2011

Single or multiple institution Single Single Single Single Single Multiple

N 294 678 838 998 1,161 1,523

Ultrasound evaluation Yes Yes Yes Yes Yes Yes

Excluded suspected/cancers on FNA No Yes Yes Yes No Yes

Overall cancer rate, n (%) 29/294 (10) 81/678 (12) 137/838 (16) 104/998 (10.4) 252/1,161 (21.7)15% if FNAþ excluded

238/1,523(15.6)

Size of cancer, cm

Mean 0.96 2.8 NA 1.4 NA 1.1

Median NA NA NA 0.7 0.5 0.6

Range 0.2e3 0.2e7 NA 0.1e7.5 NA 0.1e9

% >1 cm 27 45 41 35 39

Cancer rate according to diagnosis, %

TNG 9 NA NA 8.2 NA 18.3

MNG 11 10 16 12.9 21.7 17.5

Graves’ disease NA 8 NA 8.7 NA 6.1

Hashimoto 28

FNA, fine-needle aspiration; MNG, nontoxic multinodular goiter; NA, data not available; TNG, toxic nodular goiter.

Vol. 216, No. 4, April 2013 Smith et al Incidental Cancer Rate after Thyroidectomy 575

malignancy in presumably benign thyroid disease. Incontrast, Miccoli and colleagues2 reported no significantassociation with age, sex, or thyroid volume and the pres-ence of incidental cancer, instead they noted that cancerwas more frequent in euthyroid patients than in toxicgoiter patients. In addition, a large study of 1,161patients in an endemic MNG area in Crete, Greece founda significant association between younger age and pres-ence of incidental thyroid cancer, but failed to showmale sex as a risk factor for cancer.5 These studies,however, are single-institutional experiences and fromvariable patient populations (European vs North Amer-ican). It is plausible that the difference in patient popula-tion (eg, environmental exposures, diet, and presentationto physicians) affects the rate of malignancy among thosepatients who do present for surgery. Certainly endemicgoiter derives from a different biological milieu and there-fore cannot be expected to represent a nonendemicpatient population.An important question is why the incidental cancer

rates found in this study are higher than most reportedsurgical series studying similar subjects7-9 (Table 4; 7%to 16%). One possible explanation is that pathologists atdifferent institutions are performing more or less aggres-sive evaluations of the thyroid tissue and therefore findingvariable rates of incidental thyroid cancer.7,10 This multi-institutional study found similar rates of incidental cancers

among the 3 participating institutions, arguing against aninstitution- or pathologist-specific bias toward moreaggressive discovery of incidental cancers. A more plau-sible explanation is that the overall incidence of thyroidcancer is increasing and that our tri-institutional datamerely reflect that trend during the past 10 years. Reviewof the Surveillance Epidemiology and End Results data-base indicates that papillary thyroid carcinoma rates areincreasing.6 The observed increases in thyroid cancer rateshave been reported for both small and >1-cm tumors andmight be due in part to enhanced detection.11,12

In the current series, even with ubiquitous use of ultra-sound and liberal use of preoperative FNA by the refer-ring endocrinologist or surgeon (43%), a considerablenumber of cancers were not detected preoperatively.The argument can be made that these cancers are not clin-ically significant, given that the incidence of thyroidmicrocarcinoma (<1 cm) at autopsy ranges from 6% to36%13,14 and the majority of these cancers will likely notlead to disease-related mortality.9,15 The authors, however,believe that a considerable portion (39% �1 cm) of theseincidental cancers are clinically relevant and, despite ourcurrent screening methods, they are being missed. Atthis time, it cannot be determined which cancers willlead to clinically significant disease, as even some micro-carcinomas present with lymph node or distant metas-tasis. Unfortunately, the long (10 to 15 years) follow-up

576 Smith et al Incidental Cancer Rate after Thyroidectomy J Am Coll Surg

needed to demonstrate disease-related mortality is notavailable.The current study also finds a significantly higher rate of

incidental cancers in the TNG and MNG patients whencompared with Graves’ disease patients (approximately18% vs 6%; OR ¼ 4.1). Prevalence of thyroid cancer inGraves’ disease is reported to range from 0.5% to 15%,with most large surgical series reporting rates around 3%to 5% and usually lower than TNG or MNG.2,16 Onthe other hand, the incidence of cancer in patients withnodular Graves’ disease has been reported to be 15%.17

In the current study, <2% of the Graves’ disease patientshad nodular disease. We hypothesize that the differencesin cancer rates could be due to changes in the underlyingthyroidal epithelial biology that leads to nodularity asopposed to the hyper- or hypothyroid state. In nodularthyroid disease such as MNG or TNG, impairment ofthyroid hormone synthesis and/or increased TSH levelsleading to hypertrophy and hyperplasia can trigger acqui-sition of genetic abnormalities favoring up-regulation ofthe TSH-signaling pathway.18 Unchecked constitutiveactivation of this pathway could prompt excessive prolifer-ation and de-differentiation in pockets of the gland, evenas colloidal involution progresses in other areas to favora fertile ground for progression to neoplasia.18 These theo-ries will need to be evaluated by analyzing the tissues ofpatients with or without cancer, such as those identifiedin this study for markers of proliferation and knowndrivers of thyroid carcinogenesis, but this is beyond thescope of the current study.Both Graves’ disease and nodular goiters occur much

more frequently in women; however, the current dataand those of others suggest that male sex4 and possiblyyounger age are associated with higher rates of incidentalthyroid cancer in nodular toxic or nontoxic goiter. Theseobservations have not held for patients with Graves’disease. Although the cause is unclear, it seems apparentthat more caution should be taken when one approachesthe young and/or males with goiters. Sex and young age,therefore, should be taken into account when makingindividual patient recommendations about the surgicalmanagement of their thyroid disease. Additional studiesmust explore these associations before making durablerecommendations about surgical decision making (eg,males with MNG or TNG should undergo totalthyroidectomy).In contrast to the current study’s findings, other

authors have noted an increased incidence of cancer inpatients with underlying thyroiditis.7,19-21 We did not en-counter this association in the overall analysis or whenanalyzed based on specific clinical presentation (MNG,TNG, or Graves’ disease). Certainly, the definition of

thyroiditis (ie, thyroid antibodies, clinical need forthyroxine supplementation, and pathologic evaluation)will affect the rate of malignancy among thyroiditispatients. It is interesting to note that both lymphocyticthyroiditis and Graves’ disease represent a spectrum ofautoimmune thyroid disease. If the immunogenic natureof Graves’ disease was indeed protective, then one wouldexpect this protective effect to be demonstrated forlymphocytic thyroiditis patients as well. The associationof thyroiditis and cancer has been debated in the litera-ture and is by no means resolved.Patients with Graves’ disease undergoing surgical

removal of their thyroid are best treated with totalthyroidectomy.22 The relatively low rate of incidentalthyroid cancer observed in Graves’ disease might notfactor into the current treatment algorithm for thesepatients. In contrast, patients with a toxic MNG haveboth their toxic state as well nodular disease/malignancyrisk to consider when deciding between radio-iodine abla-tion and total thyroidectomy for definitive therapy. Inthese patients, the presence or suspicion of cancer mightfavor total thyroidectomy.3,22 This updated informationabout the rate of malignancy should at least be providedto patients before decision making about the treatment oftheir TNG. Lastly, patients with TNG or MNG might beable to treat their goiter with less than total removal oftheir thyroid (ie, lobectomy) if disease is isolated toa single lobe. As noted here, use of ultrasound andFNA fell short in identifying 18% of patients with inci-dental thyroid cancer. In those patients with cancers �1cm, the patient might be subjected to a second procedure(completion thyroidectomy). Given this information,clinical presentation and symptoms, a total thyroidec-tomy in the hands of an experienced thyroid surgeonmight be the best treatment for these patients. Thesedata reiterate the importance of full use of FNA inpatients who might be offered a nonsurgical managementoption for their goiter in an attempt to minimize the rateof missed cancers.An obvious limitation of this study is its retrospective

nature. Information related to family history and expo-sure to external beam radiation was not available in allpatients. It is possible but unlikely that there werea disproportionate number of such patients in the groupwith incidental cancers. The study is a surgical series and,as such, might not accurately reflect (ie, overestimate) theprevalence of thyroid cancer in the studied diseases asa whole. Although all patients had a thyroid ultrasound,not all patients underwent FNA, and it is thereforepossible that the screening process might not have beenadequate and could have led to observed higher incidentalcancer rates. This is a multi-institutional series reflecting

Vol. 216, No. 4, April 2013 Smith et al Discussion 577

the management of thyroid disease at centers withhigh-volume surgeons who routinely perform totalthyroidectomy. The authors suggest caution when recom-mending total thyroidectomy based solely on the inci-dental cancer rate data presented in this study. Thenatural history of these incidental cancers is most likelyindolent, yet the complications of total thyroidectomycan be lifelong and substantial.

CONCLUSIONSThese data confirm higher than expected incidentalthyroid cancer rates in the largest multi-institutionalsurgical series to date (15.6%). The percentage of cancercases in MNG or TNG patients was significantly higherthan the Graves’ disease cancer rate. Males and youngpatients were more likely to harbor incidental carcinoma.These data support consideration of initial total thyroid-ectomy as the preferred approach for patients referred tothe surgeon with bilateral nodular disease.

Author Contributions

Study conception and design: Smith, X Chen, Broome,Solorzano

Acquisition of data: Smith, Schneider, Broome, Sippel,H Chen, Solorzano

Analysis and interpretation of data: Smith, X Chen,Broome, Solorzano

Drafting of manuscript: Smith, Broome, SolorzanoCritical revision: Smith, X Chen, Schneider, Broome,Sippel, H Chen, Solorzano

REFERENCES

1. Marqusee E, Benson CB, Frates MC, et al. Usefulness of ultra-sonography in the management of nodular thyroid disease.Ann Int Med 2000;133:696e700.

2. Miccoli P, Minuto MN, Galleri D, et al. Incidental thyroidcarcinoma in a large series of consecutive patients operatedon for benign thyroid disease. Anz J Surg 2006;76:123e126.

3. Smith JJ, Chen X, Schneider DF, et al. Toxic nodular goiterand cancer: a compelling case for thyroidectomy. Ann SurgOncol 2012 Oct 30. [Epub ahead of print].

4. Luo J, McManus C, Chen H, Sippel RS. Are there predictorsof malignancy in patients with multinodular goiter? J Surg Res2012;174:207e210.

5. Lasithiotakis K, Grisbolaki E, Koutsomanolis D, et al. Indica-tions for surgery and significance of unrecognized cancerin endemic multinodular goiter. World J Surg 2012;36:1286e1292.

6. Enewold L, Zhu KM, Ron E, et al. Rising thyroid cancer inci-dence in the United States by demographic and tumor charac-teristics, 1980�2005. Cancer Epidemiol Biomarker Prev2009;18:784e791.

7. Bradly DP, Reddy V, Prinz RA, Gattuso P. Incidental papil-lary carcinoma in patients treated surgically for benign thyroiddiseases. Surgery 2009;146:1099e1104.

8. Cerci C, Cerci SS, Eroglu E, et al. Thyroid cancer in toxic andnon-toxic multinodular goiter. J Postgrad Med 2007;53:157e160.

9. Ito Y, Higashiyama T, Takamura Y, et al. Prognosis of patientswith benign thyroid diseases accompanied by incidental papil-lary carcinoma undetectable on preoperative imaging tests.World J Surg 2007;31:1672e1676.

10. Bisi H, Fernandes VS, de Camargo RY, et al. The prevalenceof unsuspected thyroid pathology in 300 sequential autopsies,with special reference to the incidental carcinoma. Cancer1989;64:1888e1893.

11. Davies L, Welch HG. Increasing incidence of thyroid cancer inthe United States, 1973�2002. JAMA 2006;295:2164e2167.

12. Verkooijen HM, Fioretta G, Pache JC, et al. Diagnosticchanges as a reason for the increase in papillary thyroid cancerincidence in Geneva, Switzerland. Cancer Cause Control2003;14:13e17.

13. Harach HR, Franssila KO, Wasenius VM. Occult papillarycarcinoma of the thyroid. A "normal" finding in Finland. Asystematic autopsy study. Cancer 1985;56:531e538.

14. de Matos PS, Ferreira AP, Ward LS. Prevalence of papillarymicrocarcinoma of the thyroid in Brazilian autopsy andsurgical series. Endocr Pathol 2006;17:165e173.

15. Neuhold N, Schultheis A, Hermann M, et al. Incidental papil-lary microcarcinoma of the thyroiddfurther evidence of a verylow malignant potential: a retrospective clinicopathologicalstudy with up to 30 years of follow-up. Ann Surg Oncol2011;18:3430e3436.

16. Pazaitou-Panayiotou K, Michalakis K, Paschke R. Thyroidcancer in patients with hyperthyroidism. Horm Metab Res2012;44:255e262.

17. Kraimps JL, Bouin-Pineau MH, Mathonnet M, et al. Multi-centre study of thyroid nodules in patients with Graves’disease. Br J Surg 2000;87:1111e1113.

18. Kumar V, Abbas AK, Faust N, et al, eds. Robbins and CotranPathologic Basis of Disease, Professional Edition. 8th ed. Phil-adelphia: Saunders Elsevier; 2010.

19. Minuto MN, Berti P, Miccoli M, et al. Minimally invasivevideo-assisted thyroidectomy: an analysis of results and a revi-sion of indications. Surg Endosc 2012;26:818e822.

20. Repplinger D, Bargren A, Zhang YW, et al. Is Hashimoto’sthyroiditis a risk factor for papillary thyroid cancer? J SurgRes 2008;150:49e52.

21. Kebebew E, Treseler PA, Ituarte PH, Clark OH. Coexistingchronic lymphocytic thyroiditis and papillary thyroid cancerrevisited. World J Surg 2001;25:632e637.

22. Bahn RS, Burch HB, Cooper DS, et al. Hyperthyroidism andother causes of thyrotoxicosis: management guidelines of theAmerican Thyroid Association and American Association ofClinical Endocrinologists. Endocr Pract 2011;17:456e520.

Discussion

DR WILLIAM BARRY INABNET (New York, NY): Dr Smith andcolleagues report a large, multi-institutional study that assesses inci-dentally detected thyroid cancer in a large cohort of 1,523 patientswho underwent thyroidectomy for benign disease. The Miami,

Vanderbilt, and Wisconsin endocrine surgery teams continue theirimmense clinical research productivity, as this study clearly