Vol. 3, 1565-1569, September 1997 Clinical Cancer Research 1565

3 The abbreviations used are: LABC, locally advanced breast carcinoma;

NACT, neoadjuvant chemotherapy; ThM, Tumor-Node-Metastasis.

Relative Value of Physical Examination, Mammography, and Breast

Sonography in Evaluating the Size of the Primary Tumor and

Regional Lymph Node Metastases in Women Receiving

Neoadjuvant Chemotherapy for Locally Advanced

Breast Carcinoma1

Juan Herrada, Revathy B. Iyer,

E. Neely Atkinson, Nour Sneige,

Aman U. Buzdar, and Gabriel N. Hortobagyi2

Departments of Breast Medical Oncology [J. H., A. U. B., G. N. H.],

Diagnostic Radiology [R. B. I.], Biomathematics [E. N. A.], and

Pathology [N. S.], The University of Texas M. D. Anderson Cancer

Center, Houston, Texas 77030

ABSTRACT

The purpose of this study was to correlate physicalexamination and sonographic and mammographic measure-ments of breast tumors and regional lymph nodes withpathological findings and to evaluate the effect of neoadju-

vant chemotherapy on clinical Tumor-Node-Metastasisstage by noninvasive methods. This was a retrospective anal-

ysis of 100 patients with locally advanced breast cancerregistered and treated in prospective trials of neoadjuvant

chemotherapy. All patients received four cycles of a doxo-rubicin-containing regimen and had nomnvasive evaluationof the primary tumor and regional lymph nodes before andafter neoadjuvant chemotherapy by physical examination,sonography, and mammography and underwent breast sur-

gery and axillary dissection within 5 weeks after com-pletion

of neoadjuvant chemotherapy. The correlations betweenclinical and pathological measurements were determined by

Spearman rank correlation analysis. A proportional oddsmodel was used to examine predictive values.

Eighty-three patients had both a clinically detectableprimary tumor and lymph node metastases. Sixty-four pa-tients had a decrease in Tumor-Node-Metastasis stage afterchemotherapy. For 54% of patients, there was concordancein clinical response between the primary tumor and lymph

node compartment; for the rest, results were discordant.Physical examination correlated best with pathological find-

Received 1/10/97; revised 5/15/97; accepted 5/19/97.

The costs of publication of this article were defrayed in part by the

payment of page charges. This article must therefore be hereby marked

advertisement in accordance with 18 U.S.C. Section 1734 solely to

indicate this fact.

I This work was supported in part by Grant CA16672 and by the Nylene

Eckles Professorship in Breast Cancer Research.

2 To whom requests for reprints should be addressed, at Department ofBreast Medical Oncology, Box 56, The University of Texas M. D.Anderson Cancer Center, 15 15 Holcombe Boulevard, Houston, TX77030. Phone: (713)792-2817; Fax: (713)794-4385.

ings in the measurement of the primary tumor (P 0.0003),whereas sonography was the most accurate predictor of size

for axillary lymph nodes (P = 0.0005). The combination ofphysical examination and mammography worked best forassessment of the primary tumor (P 0.003), whereas

combining physical examination with sonography gave op-timal evaluation of regional lymph nodes (P 0.0001).

In conclusion, physical examination is the best nonin-vasive predictor of the real size of locally advanced primary

breast cancer, whereas sonography correlates better with

the real dimensions of axillary lymph nodes. The combina-tion of physical examination with either mammography or

sonography significantly improves the accuracy of noninva-sive assessment of tumor dimensions.

INTRODUCTION

The degree of response to preoperative or neoadjuvant chem-

otherapy is an important indicator of prognosis for patients with

LABC3 (1, 2). Accurate staging is of paramount importance to

determine the extent of disease before and after NACT and thus to

ascertain changes in tumor dimensions. Pathological staging of

surgical specimens (total or partial mastectomy) and lymph node

dissection provides the most accurate information about currently

accepted prognostic indicators. Two of these indicators, tumor size

and, more importantly, axillary lymph node status, are used as a

guide to select optimal adjuvant treatments in breast cancer patients

following NACT and local therapy (3). Clinical assessment of the

primary breast tumor and regional lymph nodes has been done by

physical examination, breast sonography, and mammography.

Some reports (4-6) have suggested that physical examination and

mammography are complementary in the assessment of primary

tumor response; whereas other reports have concluded that sono-

graphic measurements correlate best with pathological findings (7,

8). Fewer data are available on the clinical assessment of lymph

node metastases (9); therefore, the preoperative evaluation of their

response to neoadjuvant chemotherapy remains imprecise. This

limits information about the differential impact of NACT on the

primary breast tumor and nodal metastatic disease. We, therefore,

conducted a study of clinical assessment methods in women with

LABC.

The aims of our study were to: (a) correlate physical

examination and sonographic and mammographic measure-

Research. on October 29, 2020. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

Table 1 Patient characteristics and clinical disease stage beforeNACT

Total number of patients 100

Median age in years 47(Range) (29-69)

Primary tumor status (T)T0T1T2T3

Regional lymph node status (N)N0

N,

N2

N3

TNM stageIla (n = 3)

lib (n = 14)

lila (n = 39)

IlIb (n = 44)

2

8

20

4030

115236

3

95

22

7

7

2

3

21

19

1566 Evaluating the Size of Primary Tumor and Lymph Nodes

ments of tumors and regional lymph nodes after NACT with

pathological findings (the gold standard); (b) evaluate the effect

of NACT on clinical TNM stage by comparing baseline and

post-chemotherapy measurements for each of the three clinical

assessment modalities; and (c) assess noninvasively whether

neoadjuvant chemotherapy affects primary tumors and lymph

node metastases differently.

PATIENTS AND METHODS

Study Design. This was a retrospective analysis of pa-

tients registered in prospective trials of NACT for LABC con-

ducted by the Department of Breast Medical Oncology at The

University of Texas M. D. Anderson Cancer Center. The study

population consisted of 100 consecutive women with a histo-

logical diagnosis of LABC who met the following criteria: (a)

received three or four cycles of a doxorubicin-containing com-

bination NACT regimen described in detail previously (10, 1 1);

(b) had clinical (noninvasive) evaluation of the primary tumor

and regional lymph nodes before and after NACT by physical

examination, sonography, and mammography; and (c) under-

went breast surgery and axillary dissection within 5 weeks after

completion of NACT, with the results noted in a detailed path-

ological report. All patients were diagnosed and treated between

1985 and 1994. The median patient age was 47 years (range,

29-69 years). All noninvasive tests considered for this analysis

were performed at The University of Texas M. D. Anderson

Cancer Center. Mammography techniques used during the

course of the study included both film screen and xeromammo-

grams. All images were obtained on dedicated mammography

units: Siemens Mammomat (Stockholm, Sweden) for xeromam-

mography and film screen; and General Electric DMR seno-

graph (GE Medical Systems, Waakesha, WI) for film screen. A

dedicated film processor with extended processing was used. All

positioning was performed by registered mammography tech-

nologists. The sonographic examinations were performed on

three different units with high resolution 7-megahertz transduc-

ers. These units included Accuson (Accuson Corp., Mountain-

view, CA), ATL (Advanced Technology Laboratories, Inc.,

Bothell, WA), and Aloka (Corometrics Medical Systems, Inc.,

Wallingford, CT). Before March 1993, mammography was per-

formed with dedicated mammography units and the xeromam-

mography technique. Subsequent mammography was done us-

ing the low-dose film screen technique. All surgical procedures

were also performed at M. D. Anderson, and the surgical spec-

imens were examined and reported on by our breast patholo-

gists. Information on disease stage and other pretreatment pa-

tient characteristics are summarized in Tables 1 and 2. Most

patients had large, bulky primary tumors, and more than one-

third had bulky, matted axillary lymph nodes. Close to 90% of

the patients had clinically positive axillary lymph nodes.

Methods. The two longest perpendicular dimensions of

the primary tumor and clinically positive regional lymph nodes

were recorded in centimeters and multiplied to determine the

surface area in square centimeters. The primary breast tumor

and lymph node metastasis measurements were obtained by

physical examination, sonography, and mammography per-

formed before the first cycle and after the fourth cycle of NACT

(usually less than a week before the surgical resection). Patients

(T.,N0)

(T2N1)(T3N0)

(T3N1)

(T3N2)

(T,N2)(T1N,)(T0N2)(T4N0)(T4N,)(T4N7)

(T4N3)

with LABC were examined jointly by surgical, medical, and

radiation oncologists to determine the extent of disease and plan

the optimal combination of therapies. This group always in-

cluded the treating medical oncologist. The baseline measure-

ments and the preoperative measurements were the result of a

consensus after a joint physical examination. The measurements

from the physical examination were obtained from a review of

each patient’s chart. At the time of the review, if more than one

set of measurements for the same mass was recorded in the

medical record, the largest dimensions were considered for the

analysis. If more than one tumor was present in the same breast,

only the largest tumor mass was considered for this review.

The sonographic and mammographic measurements were

obtained from a retrospective review of sonograms and mam-

mograms by a single radiologist (R. B. I.), thus avoiding varia-

tions in interpretation of films (12, 13). Areas of microcalcifi-

cations on mammograms were considered nonmeasurable

tumor.

Pathological measurements were obtained from the surgi-

cal pathology report. Measurements were obtained prospec-

tively at the time of the initial interpretation by one of two

dedicated breast pathologists. If more than one tumor was pres-

ent in the same breast (or surgical pathology specimen), only the

largest tumor was considered. For measurement of nodal me-

tastases, the size of the largest lymph node involved by tumor

was considered.

The correlations between clinical and pathological assess-

ments were determined by Spearman rank correlation analysis.

A proportional odds model was used to examine predictive

values. The “Logistic” procedure from the S.A.S. program was

used to establish the concordance between the observed cate-

gory and the category predicted by the regression model.

Research. on October 29, 2020. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

Clinical Cancer Research 1567

Table 2 Disease stage of 83” patients after NACT (befo re surgery) as assessed clinically and pathologically

No. of patients (%) according to

Physical examination Sonography Mammography Pathology

Primary tumor status (T)

T0 20(24%) 11 (13%) 21 (25%) 8(10%)T, 13(16%) 26(31%) 17(20%) 35 (42%)T, 33 (40%) 41 (49%) 38 (46%) 30 (36%)T3 11(13%) 5(6%) 6(7%) 8(10%)

T41’ 6(7%) 0(0%) 1 (1%) 2(2%)Regional lymph node status (N)

N0 56 (67%) 43 (52%) 43 (52%) 25 (30%)

N, 21 (25%) 39(47%) 39(47%) 56 (68%)

N, 6(7%) 1(1%) 1(1%) 2(2%)

ThM stage

0 12(14%) 6(7%) 12(14%) 5(6%)

I 11 (13%) 14(17%) 9(11%) 13(16%)

ha 37 (45%) 39 (47%) 37 (45%) 32 (39%)lIb 11 (13%) 20(24%) 19(23%) 24 (29%)lIla 5 (6%) 4 (5%) 5 (6%) 7 (8%)

IIIb 7(8%) 0(0%) 1 (1%) 2(2%)

(1 Seventeen of the initial 100 women were excluded because they had no clinically measurable disease or their films were not available forreview.

1’ Only a measurable residual breast mass with associated skin edema was considered T4.

Table 3 Spearman’s rank correlation coefficie nt (r) of clinical (noninvasive) measurements obtained after NACT (before surgery) compared

with pathological measurements

Noninvasive method of measurement

Correlation with pathology

Tumor size Lymph node size

Physical examination

Sonography

Mammography

r = 0.726 P < 0.0001 r 0.318 P 0.003r = 0.600 P < 0.0001 r = 0.514 P = 0.0001

r 0.649 P < 0.0169 r 0.430 P 0.4251

RESULTS

Thirty-six of the 100 patients evaluated had no change in

clinical TNM stage after NACT, whereas 64 had a decrease of

one or more stages. Of these 64 patients, 2% had a three-stage

decrease (downstaging) in clinical TNM stage, I 2% had a

two-stage decrease, and 86% had a one-stage decrease.

The effects of NACT on the primary breast tumor and

regional lymph node compartments could be evaluated in 83

patients who had clinically evident primary breast tumors and

lymph node metastases at the time of diagnosis of LABC.

Forty-one % of these patients had concordant clinical downstag-

ing of both the primary tumor and regional lymph node metas-

tases; 13% had no appreciable decrease in either compartment;

in 25%, the tumor decreased while the nodal metastases did not;

and in another 2 1 %, the opposite occurred. The difference

between tumor decrease alone (25%) and lymph node decrease

alone (21%) was not significant (P = 0.63).

Measurements done after NACT by each of the three

noninvasive methods were correlated separately with the path-

ological measurements. We found that physical examination

showed the best correlation (correlation coefficient, r - 0.726)

with pathology for the primary tumor, and sonography provided

the best correlation for the lymph nodes (r = 0.5 14; Table 3).

To examine the value of each of the three noninvasive

measurements after NACT in predicting pathological findings,

we established four size categories based on the estimated

surface areas of the mass (in cm2): 0, >0 to 1, > I to 5, and >5.

The data were then fit using the proportional odds model. Let y�

be the categorized size of thej”� patient, e.g., if patientj’s tumor

as measured by pathology is between I and 5 cm (2), then y3 =

3, since this tumor falls into the third size category. Let

denote the probability that y� � i. Then the proportional odds

model assumes that � 1/(1 + exp(-r�)), where r� a + f3,I � � � � � I3pXjp’a is an intercept parameter, X�k is the value

of the k” independent variable for patient j, and �k �5 the slope

parameter for the k#{176}’independent variable. In our analysis, two

independent variables were used: the natural logarithm of the

tumor size as measured by mammography and by sonography.

The results (considering log data only) are summarized in Table

4. For the primary tumor, of the three noninvasive methods,

physical examination was the best predictor of the size of the

residual pathological mass (P 0.0003), followed closely by

sonography (P = 0.0005); mammography had the weakest

correlation with pathological findings (P = 0.0132). For lymph

node metastases, sonography was the best predictor of the

surface area of the largest lymph node (P 0.0005). Mammog-

raphy (P = 0.0132) and physical examination were much less

accurate (P = 0.0181). We also examined the predictive value

of two or three noninvasive methods used in combination. The

results of these analyses are summarized in Tables 5 and 6 and

Research. on October 29, 2020. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

1568 Evaluating the Size of Primary Tumor and Lymph Nodes

Table 4 Results of univariate analysis by proportional odds model topredict pathology tumor size using noninvasive measurements

pa

Noninvasive method of

measurement Tumor size Lymph node size

Physical examination 0.0003 0.0 132

Sonography 0.0005 0.0005Mammography 0.0132 0.0181

‘, Compared with pathological measurement.

give the effects of adding variables to the regression model. x2 for

the addition of sonography to physical examination in the model

for primary tumor size is obtained by subtracting the value for the

model containing both physical examination and sonography

(169.064) from the value for the model containing only physical

examination (174.366), yielding 5.302. This value is distributed as

x2 with one degree of freedom, from which the corresponding P

can be obtained. We started our analysis from the physical exam-

ination, because routine clinical practice dictates that this noninva-

sive method be performed first, and in virtually all patients, before

any further test (mammogram or sonogram) is considered. For the

primary tumor, a significant improvement in the predictive value

was noted with the combination of sonography and physical ex-

amination (P = 0.021) and with the combination of mammography

and physical examination (P 0.003). The addition of mammog-

raphy to the combination of physical examination and sonography

significantly further improved the predictive value (P 0.032), but

the addition of sonography to the combination of physical exami-

nation and mammography did not (P 0.377; Table 5). For the

nodal metastases, the combination of mammography and physical

examination improved predictive value over physical examination

alone (0.027), and the combination of breast sonography and phys-

ical examination resulted in an even greater improvement in the

predictive value (P = 0.0001). The addition of mammography to

the combination of physical examination and sonography did not

significantly further improve predictive value (P = 0.612), but

adding sonography to the combination of physical examination and

mammography did (P = 0.001 ; Table 6).

DISCUSSION

In this study, almost two-thirds (64%) of patients with

LABC treated with NACT achieved a decrease in TNM stage.

Primary breast tumors and nodal metastases appeared to be

equally affected by neoadjuvant chemotherapy. However, al-

though these results are in accordance with other clinical obser-

vations, they need to be interpreted cautiously. Our patient

population was selected to evaluate the relative accuracy of the

noninvasive methods used since 1985 to assess LABC. For the

purpose of our study, microcalcifications on mammograms were

considered nonmeasurable disease, and nonpalpable masses

were considered as T0, even if skin changes noted prior to

NACT had not resolved completely afterward.

Adjuvant chemotherapy, given after local control of the

primary tumor, has been demonstrated to improve disease-free

and overall survival rates in patients with breast cancer (14, 15)

The paradigm on which the use of adjuvant chemotherapy is

based is the treatment of micrometastases. Our data suggest that

Table 5 Summary of the effects of adding additional variables to theproportional odds model

The �2 values are the likelihood ratio statistics obtained when thevariables to the right of the + sign are added to the equation containing

the variables in parentheses. All x2 values have one degree of freedom.

x2 values are obtained by subtracting the -2 log likelihood value of themodel with the additional variable from the value for the modelscontaining only the variable(s) in parentheses.

Variables in equation X2

(None) + physical examination 39.529 <0.0001(Physical examination) + sonography 5.302 0.02 13

(Physical examination) + mammography 9.095 0.0026(Physical examination + sonography) 4.572 0.0325

+ mammography(Physical examination + mammography) 0.779 0.3774

+ sonography

the axillary metastatic cell clone, although expressing a more

aggressive phenotype, appears to be as sensitive to chemother-

apy as the primary tumor cells are.

For the assessment of the primary breast tumor, physical

examination measurements in our study showed the highest corre-

lation with and the highest predictive value for the pathological

findings. We, therefore, conclude that physical examination by

experienced examiners remains the best single noninvasive method

of assessing the size of the primary tumor in women with LABC.

However, other investigators have published differing conclusions.

In 1987, Fomage et a!. (7) described 31 breast cancer cases with

pathologically defined T1 (n = 23) and T2 (n = 8) tumors. The

maximal diameters of the tumors, obtained by physical examina-

tion (29 cases), mammography (20 cases), and sonography (31

cases), were correlated with the values at pathological examination.

That study showed that sonography had the highest correlation

coefficient. The lesions included in that study were, however,

generally much smaller than those in our study and might have

been more difficult to delimit on physical examination. In fact, the

largest tumor reported in the study by Fomage et a!. (7) was 3 cm

in longest diameter on pathological evaluation. Furthermore,

Fornage et a!. (7) included only unidimensional measurements,

whereas in our study, we used two-dimensional measurements

because surface area defines tumor extent more accurately. Also, it

should be stressed that the physical examination measurements in

our report usually represented the consensus of two or more expert

breast oncologists. In another study, Serrano-Migallbn et a!. (8)

studied 34 patients with breast masses (16 benign and 18 malig-

nant). The maximum diameters estimated by physical examination,

sonography, and mammography were compared in each case to the

pathological measurements. That 1993 study also concluded that

sonography had the best correlation with pathological findings.

However, differences in patient numbers and methodology may

account for the discrepancy with our results.

For the assessment of lymph node metastases, sonographic

measurements showed the highest correlation with and the high-

est predictive value for the pathological findings. Thus, we

conclude that sonography is the best single noninvasive method

of assessing the extent of nodal involvement based on size of

nodes.

After NACT, the combination of physical examination and

mammography had the best predictive value for the true dimen-

Research. on October 29, 2020. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

is to maximize the use of breast-conserving therapy in patients with

large T2 or locally advanced breast cancer.

Pathological complete remission continues to be the best

prognostic factor for metastases-free survival. However, neither

imaging method evaluated in this study is sensitive enough to

identify with certainty the absence of residual malignant cells.

Therefore, we consider that careful monitoring of response to

preoperative systemic therapy with physical examination and

sonography is useful to determine the optimal approaches to

preoperative and operative treatment, whereas findings on pa-

thology are the major determinants to select postoperative sys-

temic therapy and prognosis.

Clinical Cancer Research 1569

Table 6 Summary of the effects of adding additional variables to theproportional odds model for lymph node metastases

The x2 values are the likelihood ratio statistics obtained when thevariables to the right of the + sign are added to the equation containing

the variables in parentheses. All x2values have one degree of freedom.

x2values are obtained by subtracting the -2 log likelihood value of themodel with the additional variable from the value for the modelscontaining only the variable(s) in parentheses.

Variables in equation X2

(None) + physical examination 6.229 0.0126(Physical examination) + sonography 14.895 0.0001(Physical examination) + mammography 4.893 0.0270(Physical examination + sonography) 0.256 0.6129

+ mammography(Physical examination + mammography) 10.258 0.0014

+ sonography

sions of the primary breast tumor, whereas the combination of

physical examination and sonography provided the best assess-

ment of the lymph node metastasis.

It is noteworthy that the tumors located in different com-

partments are best defined by different methods. The different

anatomical structures of the breast and the axillary area may

account for this. Lluch et a!. (9) evaluated 60 patients (33 with

pathologically defined axillary involvement) who underwent

neoadjuvant chemotherapy to study the role of physical exam-

ination, sonography, mammography, and computed tomography

in defining tumor and nodal size; results with these techniques

were compared to the pathologically determined tumor size in

the surgical specimens. Liuch et a!. (9) reported that computed

tomography was the best noninvasive method to detect axillary

involvement, whereas for the primary tumor, physical examina-

tion had the best correlation coefficient, as in our study.

The implication of our study is that for purposes of clinical

practice, both physical examination and sonography should be

routinely used in the evaluation of patients with LABC undergoing

treatment with NACT. Although serial mammography provides an

additional benefit in the assessment of the primary tumor size, it is

not useful in the evaluation of lymph node metastases. Therefore,

serial mammography should be used only in selected cases, such as

those for whom neither sonography nor physical examination iden-

tifies a well-defined, and therefore measurable, tumor. By restrict-

ing the use of serial mammography, unnecessary radiation expo-

sure and expense to the patient could be avoided. However, the role

of a preoperative mammogram after NACT to assess the presence

of nonmeasurable disease as manifested by microcalcifications was

not evaluated in this study. For women in whom breast-conserving

therapy is being considered after downstaging by NACT, careful

diagnostic film screen mammography with appropriate focal spot

magnification views is still recommended to define the extent of

microcalcifications as a predictor of ductal carcinoma extent.

The noninvasive determination of response to neoadjuvant

therapy is important to guide the choice of subsequent therapy. For

instance, if no substantial response to NACT is seen, a different,

non-cross-resistant chemotherapy regimen (or radiotherapy) might

be used preoperatively to enhance resectability or increase the

possibility of breast-conserving surgery. Although these were not

the goals of the present study, the results of this analysis are being

tested in a subsequent clinical trial, in which one of the objectives

REFERENCES

1. Feldman, L. D., Hortobagyi, G. N., Buzdar, A. U., Ames, F. C., andBlumenschein, G. R. Pathological assessment of response to induction

chemotherapy in breast cancer. Cancer Res., 46: 2578-2581, 1986.

2. Hortobagyi, G. N., Ames, F. C., Buzdar, A. U., Kau, S. W.,McNeese, M. D., Paulus, D., Hug, V., Holmes, F. A., Romsdahl, MM.,Fraschini, G., et a!. Management of stage III primary breast cancer withprimary chemotherapy, surgery, and radiation therapy. Cancer (Phila.),62: 2507-2516, 1988.

3. McCready, D. R., Hortobagyi, G. N., Kau, S. W., Smith, T. L.,Buzdar, A. U., and Balch, C. M. The prognostic significance of lymphnode metastases after preoperative chemotherapy for locally advanced

breast cancer. Arch. Surg., 124: 21-25, 1989.

4. Cocconi, G., di Blasio, B., Alberti, G., Bisagni, G., Botti, E., andPeracchia, G. Problems in evaluating response of primary breast cancerto systemic therapy. Breast Cancer Res. Treat., 4: 309-313, 1984.

5. Segel, M. C., Paulus, D. D., and Hortobagyi, G. N. Advancedprimary breast cancer: assessment at mammography of response toinduction chemotherapy. Radiology, 169: 49-54, 1988.

6. Dershaw, D. D., Drossman, S., Liberman, L., and Abramson, A. As-sessment of response to therapy of primary breast cancer by mammographyand physical examination. Cancer (Phila.), 75: 2093-2098, 1995.

7. Fornage, B. D., Toubas, 0., and Morel, M. Clinical, mammographic,

and sonographic determination of preoperative breast cancer size. Can-cer (Phila.), 60: 765-771, 1987.

8. Serrano-Migallbn, J. A., Sandoval-Guerrero, F. J., Miranda-Hemandez,H., and Martinez-Macias, R. Estudio comparativo entre clinica, mastografiay ultrasonido, pars determinar ci tamano de las lesiones mamarias. Rev.Inst. Nal. Cancerol. (Mex), 39: 1931-1936, 1993.

9. Lluch, A., Cervantes, A., Pardo, J. D., et al. Assessment of primarytumor and axillary involvement in breast cancer after neoadjuvant chemo-therapy. A comparison of clinical examination, mammography, ultra.sonog-raphy and CT-scan. Proc. Am. Soc. Clin. Oncol., 14: 135, 1995.

10. Swenerton, K. D., Legha, S. S., Smith, T.. Hortobagyi, G. N.,Gehan, E. A., Yap, H-Y., Gutterman, J. U., and Blumenschein, G. R.Prognostic factors in metastatic breast cancer treated with combinationchemotherapy. Cancer Res., 39: 1552-1562, 1979.

11. Hortobagyi, G. N., Smith, T. L.. Legha, S. S., Swenerton, K. D.,Gehan, E. A., Yap, H-Y., Buzdar, A. U., and Blumenschein, G. R.

Multivariate analysis of prognostic factors in metastatic breast cancer.J. Clin. Oncol., 1: 776-786, 1983.

12. Fornage, B. D. Ultrasound of the breast. Ultrasound Quart., II:

1-39, 1993.

13. Elmore, J. G., Wells, C. K., Lee, C. H., Howard, D. H., andFeinstein, A. R. Variability in radiologists’ interpretation of mammo-

grams. N. Engl. J. Med., 331: 1493-1499, 1994.

14. Early Breast Cancer Trialists’ Collaborative Group. Systemic treat-ment of early breast cancer by hormonal, cytotoxic, or immune therapy.Lancet, 339: 1-15, 1992.

15. Early Breast Cancer Trialists’ Collaborative Group. Systemic treat-ment of early breast cancer by hormonal, cytotoxic, or immune therapy.

Lancet. 339: 71-85, 1992.

Research. on October 29, 2020. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

1997;3:1565-1569. Clin Cancer Res   J Herrada, R B Iyer, E N Atkinson, et al.   carcinoma.neoadjuvant chemotherapy for locally advanced breast and regional lymph node metastases in women receivingbreast sonography in evaluating the size of the primary tumor Relative value of physical examination, mammography, and

  Updated version

  http://clincancerres.aacrjournals.org/content/3/9/1565

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://clincancerres.aacrjournals.org/content/3/9/1565To request permission to re-use all or part of this article, use this link

Research. on October 29, 2020. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from