Zo.utexas.edu

⅐ N U M B E R 2 5 ⅐ S E P T E M B E R 1 2 0 0 7 Prognostic and Predictive Value of Centrally Reviewed Expression of Estrogen and Progesterone Receptors in a Laboratory Medicine, European Instituteof Oncology, University of Milan, Milan,Italy; International Breast Cancer Study Randomized Trial Comparing Letrozole and Tamoxifen Group (IBCSG) Statistical Center, Dana-Farber Cancer Institute, Harvard School Adjuvant Therapy for Postmenopausal Early Breast Cancer: of Public Health, Boston, MA; Depart-ment of Pathological Anatomy, Univer-sity of Bari, Bari, Italy; Division of Giuseppe Viale, Meredith M. Regan, Eugenio Maiorano, Mauro G. Mastropasqua, Patrizia Dell’Orto, European Institute of Oncology, Milan,Italy; Division of Pathology and Labora- Birgitte Bruun Rasmussen, Johnny Raffoul, Patrick Neven, Zsolt Orosz, Stephen Braye, Christian O Beat Thu¨rlimann, Richard D. Gelber, Monica Castiglione-Gertsch, Karen N. Price, Aron Goldhirsch, Barry A. Gusterson, and Alan S. Coates Pathology, Nordsjaellands Hospital, Hill-eroed, Denmark; Service d’Anatomie etde Cytologie Pathologiques, Centre Hospitalier-site-Belfort-Montbéliard,Montbéliard, France; Department ofGyn Oncol and Multidisciplinary Breast To evaluate locally versus centrally assessed estrogen (ER) and progesterone (PgR) receptor status and the impact of PgR on letrozole adjuvant therapy compared with tamoxifen in National Institute of Oncology, Budap-est, Hungary; Australian New Zealand postmenopausal women with early breast cancer.
Breast Cancer Trials Group, Universityof Newcastle and Anatomical Pathol- Patients and Methods
Breast International Group (BIG) 1-98 randomly assigned 8,010 patients to four arms comparing letrozole and tamoxifen with sequences of each agent. The Central Pathology Office received Heights, NSW, Australia; Kantonsspital,St Gallen, Swiss Group for Clinical material for 6,549 patients (82%), of which 79% were assessable (6,291 patients). Prognostic and predictive value of both local and central hormone receptor expression on disease-free survival (DFS) were evaluated among 3,650 assessable patients assigned to the monotherapy arms.
land, Kantonsspital, St Gallen, Switzer-land, Swiss Group for Clinical Cancer Prognostic value and the treatment effect were estimated for centrally assessed ER and PgR expression levels using the Subpopulation Treatment Effect Pattern Plot.
Dana-Farber Cancer Institute, FrontierScience and Technology Research Foundation, Harvard School of PublicHealth, Boston, MA; IBCSG Coordinat- Central review confirmed 97% of tumors as hormone receptor–positive (ER and/or PgR Ն10%).
Of 105 tumors locally ER-negative, 73 were found to have more than 10% positive cells, and eight had 1% to 9%. Of 6,100 tumors locally ER positive, 66 were found to have no staining, and 54 had and Technology Research Foundation,Boston, MA; European Institute of only 1% to 9%. Discordance was more marked for PgR than ER. Patients with tumors reclassified centrally as ER-negative, or as hormone receptor–negative, had poor DFS. Centrally assessed ER tute of Southern Switzerland, Bell-inzona, Switzerland; Division of Cancer and PgR showed prognostic value. Among patients with centrally assessed ER-expressing tumors, letrozole showed better DFS than tamoxifen, irrespective of PgR expression level.
Western Infirmary, University of Glas-gow, UK; International Breast Cancer Conclusion
Central review changed the assessment of receptor status in a substantial proportion of patients, and should be performed whenever possible in similar trials. PgR expression did not affect the relative efficacy of letrozole over tamoxifen.
print at www.jco.org on August 6, 2007.
J Clin Oncol 25:3846-3852. 2007 by American Society of Clinical Oncology Presented in part at the San AntonioBreast Cancer Symposium, San Anto-nio, TX, December 8-11, 2005.
to anastrozole, in a joint analysis of the Austrian INTRODUCTION
(ABCSG Trial 8) and German (ARNO95) trials,4 tions are found at the end of thisarticle.
Several large adjuvant trials have examined the use and in Boccardo et al.5 Extended adjuvant therapy of aromatase inhibitors in postmenopausal women with letrozole after 5 years of tamoxifen was re- with early breast cancer either in place of or follow- ported from the MA.17 trial.6 Each of these trials has mia Patologica e Medicina di Laborato-rio, Istituto Europeo di Oncologia, Via ing tamoxifen. The Arimidex Tamoxifen Alone or been updated,7-12 and each has shown a superiority Ripamonti, 435, 20141 Milano, Italy;e-mail: [email protected].
in Combination (ATAC) trial studied initial anas- of the aromatase inhibitor over 5 years of tamoxifen trozole1 while the Breast International Group (BIG) treatment, or a placebo (MA.17). Estrogen receptor 1-98 trial used letrozole.2 Switching to exemestane (ER) and progesterone receptor (PgR) content in following 2 to 3 years of tamoxifen was studied in the the primary tumor of patients with early-stage Intergroup Exemestane Study (IES),3 and switching invasive breast cancer are powerful predictors of Information downloaded from jco.ascopubs.org and provided by at Univ of Texas Libraries on January 23, 2011 from Copyright 2007 American Society of Clinical Oncology. All rights reserved.
PgR and Aromatase Inhibitors in Early Breast Cancer
response to adjuvant endocrine therapies.13 It is recommended yses of patient outcomes are therefore limited to the patients who were ran- that endocrine receptors be measured on all primary breast cancer domly assigned to the monotherapy arms.
specimens,14 and endocrine responsiveness is the primary basis for Local Pathology Assessment
selection of adjuvant systemic therapy.15 Before random assignment, steroid hormone receptor concentra- In a hypothesis-generating analysis of the influence of receptor tions/expression in the primary tumor were determined by local patholo- levels as measured at participating institutions in the ATAC study, gists. Hormone receptor status was determined by extraction assays in 7% Dowsett et al16 found that the superiority of anastrozole over tamox- of patients and classified as positive for Ն10 fmol/mg cytosol protein;otherwise an immunohistochemical (IHC) assay was used and classified as ifen was more marked in patients with tumors expressing ER but not positive for a proportion score of Ն 3 (ie, Ն10% staining cells and ignoring PgR, and suggested that this finding should be evaluated in similar trials. No such effect has been reported from the IES9 or BIG 1-982,8 Central Pathology Review
trials and no significant difference was seen in the ABCSG/ARNO Retrospective tissue collection was carried out in accordance with insti- studies4 based on institutional assessment of PgR status. Central re- tutional guidelines and national laws. Funding was provided to participating view of tumors for ER and PgR on a subset of 30% of the ATAC institutions by the trial’s pharmaceutical partner, Novartis, to cover the asso- patients is in progress.17 The present study reports the results of central ciated costs. The International Breast Cancer Study Group (IBCSG) Central pathological review of ER and PgR status in BIG 1-98, and particularly Pathology Laboratory in Milan, Italy, received material for 6,549 patients (82% examines the impact of PgR expression on the relative efficacy of of the enrolled population). The material was reviewed for histopathologic features and expression of tumor markers without knowledge of patients’treatment assignment or outcome. Assessable data were obtained from 6,291(79%) patients’ specimens, and of the 258 nonassessable sections: 37% had PATIENTS AND METHODS
sections detached; 34% had no tumor, 25% had ductal carcinoma in situ orlobular carcinoma in situ only; 2% had lymph node only; 1% had insufficientmaterial for review. The trial enrolled patients from 25 countries, and submis- The design and conduct of the study have been described previously.2 BIG sion of material varied widely: one country with 125 patients sent no material; 1-98 is a randomized, phase III, four-arm double-blind trial comparing mono- others submitted material for 26% to 100% of entered patients. For patients therapy with letrozole or with tamoxifen for 5 years, or sequential administra- randomly assigned to the monotherapy arms, material was submitted for tion of tamoxifen for 2 years followed by letrozole for 3 years, or the reverse 3,807 (77%) and was assessable for 3,650 (74%).
sequence. Eligible postmenopausal women had early breast cancer assessed asER- and/or PgR positive. Between March 1998 and March 2000, patients were Assessment of ER and PgR and Quality Assurance
randomly assigned to one of the monotherapy arms, and from April 1999 to ER and PgR were determined by immunohistochemistry as previously May 2003, to all four arms. The ethics committees and required health author- described.19 To ensure the intraobserver and interobserver reliability of the ities of each participating center approved the study protocol, and all patients central assessment, 5% of the centrally evaluated tumors were blindly reas- gave written informed consent. The primary efficacy analysis among 8,010 sessed by the same pathologist and 10% by a different pathologist (E.M., patients2 was updated as specified by protocol, and reported among the 4,922 M.G.M.). If the recorded percentage of immunostained cells differed by more patients who were randomly assigned to the monotherapy arms only.8 Anal- than 10%, then collegial reevaluation at the multiheaded microscope was Fig 1. Disease-free survival (DFS) hazard
efficacy of letrozole versus tamoxifen. The box size is inversely proportional to the SEof the HR; the extending horizontal linesindicate the 95% CI. ER, estrogen recep- Information downloaded from jco.ascopubs.org and provided by at Univ of Texas Libraries on January 23, 2011 from Copyright 2007 American Society of Clinical Oncology. All rights reserved.
Viale et al
Table 1. Central Assessment of Estrogen and Progesterone Receptor Status by Immunohistochemistry in 6,291 of 8,010 Randomly Assigned Patients
performed by three pathologists (E.M., M.G.M., G.V.); this occurred for 0.6% Comparison of Patients With and Without Material for
Statistical Methods
Central Review
ER and PgR expression were recorded as the percentage of staining cells, There were no substantial differences in the distributions of tu- and were classified using two threshold values by dichotomizing IHC mor size or grade, or nodal status among patients with and without expression as present (Ն 1% staining cells) or absent (0%), and alterna- assessable material, except that the assessable cohort included fewer tively as positive (Ն 10% staining cells) or negative (Ͻ 10%). Concordancedenoted the proportion of tumors with the same classification by local and patients with unknown disease factors and fewer locally assessed ER- negative tumors. Patients with assessable material were more often The primary trial end point was disease-free survival (DFS), which was treated with less-than-mastectomy and radiation therapy, and fewer defined as the time from random assignment to the earliest time of invasive had prior chemotherapy. Comparisons were consistent whether look- recurrence; a new invasive breast cancer in the contralateral breast; any second ing at the population of all randomly assigned patients, or the mono- (nonbreast) malignancy; or death from any cause.2 The distribution of DFS and 3- and 4-year DFS were estimated using the Kaplan-Meier method.20 Cox The assessable cohort was markedly different from the nonas- proportional hazards regression21 (stratified for randomization option, ie, two sessable cohort with regard to the duration of follow-up, DFS, or four treatment arms, and chemotherapy use) was used to estimate hazard and relative treatment effects. Among the 4,922 patients who were ratios (HRs) and 95% CIs, and assess interactions of the treatment effect assigned to monotherapy, an 18% reduction in risk of an event according to subgroups defined by ER and PgR status. The nonparametric (HR ϭ 0.82; 95% CI, 0.71 to 0.95; P ϭ .007) was observed for patients STEPP (Subpopulation Treatment Effect Pattern Plot) methodology22 wasused to investigate trends in treatment effect differences across the continuum assigned letrozole as compared with tamoxifen at a median follow-up time of 51 months,8 with estimated 86% 4-year DFS. The assessablecohort of 3,650 patients had shorter follow-up time as compared with Role of Coordinating Group, Trial Steering Committee, and
the 1,272 not assessable (48 v 71 months), had better outcome than Funding Source
those not assessable (87% v 84%, 4-year DFS), and the treatment The IBCSG was responsible for study design and coordination, data effect was more pronounced, with a hazard ratio favoring letrozole collection and management, tissue management and central pathology assess- (HR ϭ 0.74; 95% CI, 0.62 to 0.88) compared with nonassessable ment, data analysis, and reporting (including the decision to publish). Novar- patients (HR ϭ 1.04; 95% CI, 0.81 to 1.34; Fig 1). Centers that began tis (Basel, Switzerland), the manufacturer of letrozole, provided financial recruitment earlier tended not to submit material, thus shortening the support for pathology material collection and imposed no restrictions on theinvestigators with respect to trial data. The manuscript was prepared by the authors, who had full access to the data and who made final decisions on Local Hormone Receptor Assessment
content, while the Steering Committee (including a minority membership of Of entered patients, 98% had ER-positive tumors as determined Novartis employees) reviewed the manuscript and offered changes.
locally, and 93% had steroid hormone receptors assessed locally using Table 2. Numbers of Patients As Classified by Local and Central
Table 3. Numbers of Patients As Classified by Local and Central
Assessment of Progesterone Receptor Status NOTE. Of 6,291 patients, the status of 86 is unknown either by local (n ϭ 3) NOTE. One thousand fifty-four of 6,291 patients’ status unknown by local or central (n ϭ 83) assessment (not tabulated).
(n ϭ 952), central (n ϭ 94), or both (n ϭ 8) assessments not tabulated.
Information downloaded from jco.ascopubs.org and provided by at Univ of Texas Libraries on January 23, 2011 from Copyright 2007 American Society of Clinical Oncology. All rights reserved.
PgR and Aromatase Inhibitors in Early Breast Cancer
immunohistochemistry. Seven of 25 participating countries used ex- Predictive and Prognostic Value of Hormone
traction assays for at least 10% of tumors. ER-positivity rate was Receptor Status
associated with assay type and was lower among tumors assayed with To investigate the predictive and prognostic value of hormone extraction versus IHC assays (95.6% v 98% ER positive).
receptor status, analyses were limited to the 3,650 patients (74%) with A substantial number of patients (14%) had tumors without a assessable material of the 4,922 patients assigned to monotherapy with local assessment of PgR. Five countries assessed PgR in fewer than 80% letrozole or tamoxifen for whom results with updated follow-up were of patients. Among those assessed locally, 76% were PgR positive.
PgR-positivity rate was associated with assay type, and (opposite ofER) was higher among tumors assayed with extraction versus IHC Comparing Local and Central Status
assays (86% v 75% PgR positive).
Central assessment was clearly superior to local assessment in Central Hormone Receptor Review
terms of prognostic value. The 87 patients whose tumors were classi- Median ER staining was 90% (interquartile range, 85% to 99%).
fied locally as ER positive and reclassified as ER negative by central ER was assessed as absent in 90 patients (1.4%) and as low level (1% to review (using a 10% cutoff) had poor outcome on average, with 9% staining) in a further 62 patients (1.0%). Median PgR staining was estimated 69% 3-year DFS, which was similar to patients whose tu- 70% (IQR, 10% to 90%). PgR was assessed as absent in 687 patients mors were classified locally and centrally as ER negative (62% 3-year (10.9%) and as low level (1% to 9% staining) in a further 641 (10.2%).
DFS; Fig 2A). The 47 patients whose tumors were classified locally The large majority of patients had tumors expressing both receptors as ER negative and reclassified as ER positive by central review had (76.6% at a 10% cutoff; 86.8% at a 1% cutoff; Table 1).
good outcome, with estimated 91% 3-year DFS, which was similar topatients whose tumors were classified locally and centrally as ER Comparison of Local Assessment and Central Review
Of 105 patients classified locally as ER negative, central review The relative influence of central and local PgR status (regardless found 73 to have at least 10% staining and a further eight to have of ER status) on prognosis is less clear (Fig 2B). The 247 patients whose staining of 1% to 9%. Conversely, of the 6,100 patients classified tumors were classified locally as PgR positive and reclassified as PgR locally as ER positive, central review found ER to be absent in 66(1.1%) and low in 54 (0.9%; Table 2). For PgR, the discordance was negative by central review had an estimated 86% 3-year DFS, lower more marked: of 1,223 assessed locally as PgR negative, central review than patients whose tumors were classified locally and centrally as PgR reclassified 544 (44.5%) as having PgR of at least 10% and 308 (25.2%) negative (90% 3-year DFS) and to patients whose tumors were reclas- with 1% to 9% staining. Of the 4,014 assessed locally as PgR positive, sified centrally as PgR positive (91% 3-year DFS). Patients whose central review found 183 (4.6%) with absent PgR and a further 247 tumors were not evaluated locally and classified centrally as PgR neg- (6.2%) with 1% to 9% staining (Table 3). Local assessment classified 7,999 of 8,010 entered patients (99.9%) as hormone receptor–positive Combining the two receptors to define an overall assessment of (ie, ER positive and/or PgR positive). Central review of 6,291 tumors hormone receptor status (either ER positive and/or PgR positive v classified 6,101 (97.0%) as positive at a 10% cutoff and 6,161 (97.9%) both negative), the 94 patients whose tumors were classified locally as as positive at a 1% cutoff (Table 4). Concordance varied by country, hormone receptor–positive and reclassified as hormone receptor– with nine countries having a concordance of Յ 80%. Countries and negative by central review had poor outcome on average, with esti- cooperative groups that provided a lower proportion of material for mated 65% 3-year DFS, as compared with 91% 3-year DFS among central review tended also to have lower concordance.
patients whose tumors were classified locally and centrally as hormone Table 4. Numbers of Patients As Classified by Local and Central Assessment of Combined Receptor Category
Information downloaded from jco.ascopubs.org and provided by at Univ of Texas Libraries on January 23, 2011 from Copyright 2007 American Society of Clinical Oncology. All rights reserved.
Viale et al
Fig 3. Disease-free survival (DFS) according to centrally assessed (A) estrogen
receptor (ER) expression, and (B) progesterone receptor (PgR) status, classifiedas absent (0%), 1% to 9%, and Ն 10% staining cells. Numbers of patients at risk level (P ϭ .12 for interaction); the treatment effect favoring letrozolewas HR ϭ 0.72 (95% CI, 0.60 to 0.86) among patients with ER- expressing tumors as compared with HR ϭ 1.32 (95% CI, 0.63 to 2.78) among those with ER-absent tumors (Fig 1).
Fig 2. Disease-free survival (DFS) according to central and local classification of
Centrally Assessed PgR Expression
(A) estrogen receptor (ER) status, (B) progesterone receptor (PgR) status, and (C) Because of the clearly poor outcome of patients with tumors not overall hormone receptor (HR) status. Numbers of patients and events andestimates of 3-year DFS and SEs are summarized.
expressing ER and the few patients classified as expressing PgR but notER, analyses to investigate the predictive value of PgR were limited tothe 3,533 patients who were confirmed to have ER-expressing tumorsby central review, of whom 3,508 patients had assessable PgR expres- receptor–positive (Fig 2C). Conclusions were similar if using the cen- sion. PgR expression was associated with DFS (P Ͻ .0001; Fig 3B).
tral classification with a 1% cutoff (data not shown).
There was benefit, in terms of DFS, of letrozole over tamoxifen regard- Centrally Assessed ER Expression
less of PgR expression level: for patients whose tumors were PgR- DFS was statistically significantly different according to ER ex- absent, HR ϭ 0.84 (95% CI, 0.54 to 1.31); for PgR-expressing tumors, pression (P Ͻ .0001; Fig 3A) among the 3,596 patients with assessable HR ϭ 0.70 (95% CI, 0.57 to 0.85). There was no statistical evidence of ER expression by central assessment. There was not clear statistical heterogeneity in the treatment effect according to level of PgR expres- evidence of differential treatment effect according to ER expression sion (P ϭ .47 for interaction; Fig 1), illustrated in the STEPP analysis of Information downloaded from jco.ascopubs.org and provided by at Univ of Texas Libraries on January 23, 2011 from Copyright 2007 American Society of Clinical Oncology. All rights reserved.
PgR and Aromatase Inhibitors in Early Breast Cancer
DFS by treatment, which shows consistent separation between treat- dealing with protein immunohistochemical expression in clinical tri- ment groups across the continuum of PgR expression levels (Fig 4).
als. Discordant results between local and central laboratories havebeen recently analyzed with specific reference to another antigen(HER-2) in the setting of adjuvant therapy in breast cancer.24 This DISCUSSION
study reported concordance between local and central assessmentin only 81.6% of the patients when the same US Food and Drug This study involves the central review of thousands of pathological Administration–approved test kit was employed, and only 75% specimens from patients entered in the BIG 1-98 trial across 25 coun- when different immunohistochemical procedures were used lo- tries. Yet, the cohort studied, based on assessable material from 74% of cally. The authors strongly suggest quality assurance programs to randomized patients, was not entirely typical of the trial population as a whole. Patients whose material was unavailable for review were Our data provide additional support for the use of a cutoff of 1% enrolled in centers which began recruiting patients early, had poorer staining cells for both ER and PgR as indicating a better prognosis and outcome overall, and showed less or no benefit from the use of letro- at least some degree of endocrine responsiveness. This is in accord zole instead of tamoxifen. Countries and cooperative groups that with our previous experience (Viale et al, submitted for publication) provided a lower proportion of specimens for central review tended and similar to Collins et al26 and Harvey et al,27 who recommended an also to have lower concordance of local and central assessment. We Allred total score18 of 3 or above (which may equate to 1% weak thus question whether the nonassessable material would also show low concordance with central assessment. The inclusion of patients Our study shows that the endocrine treatment effect on DFS of whose tumors were not hormone receptor–positive would result in postmenopausal patients with breast cancer is primarily influenced by poorer overall outcome and a diluted treatment effect.
ER status. The role of PgR could be determined only in patients with Nevertheless, the value of central review is clear. Approximately ER-expressing tumors, since few with tumors not expressing ER en- 1% of tumors assessed locally as ER positive were reclassified centrally tered the trial. Of these, a better outcome was seen among patients as ER negative, and, though fewer in number, approximately 1% of with PgR-positive (Ն10%) tumors—a finding consistent with the tumors classified as ER negative locally were reassessed as ER positive report of Bardou et al.13 Patients treated with letrozole manifested following central review. Reclassification of receptor status occurred better outcome than those treated with tamoxifen regardless of their in both directions, and the central review was validated by its superior PgR status, and there was no statistical evidence of heterogeneity in the prognostic discrimination. Also, a substantial number of patients as- treatment effect whether PgR was considered as a categorical variable sessed locally as PgR negative were reassessed centrally as PgR positive, or as a continuum in the STEPP analysis.
and a smaller number changed in the reverse direction. False-positive Using real-time polymerase chain reaction assessment of ER and results by local testing constitute an interesting issue that has not been PgR status, Baehner et al28 found that the level of PgR was prognostic widely studied, but obviously led to the use of an ineffective therapy for in untreated patients but not predictive of tamoxifen benefit, a finding parallel with our observation that PgR expression level did not predict These findings highlight the need for accurate and standardized the additional value of letrozole on ER-expressing tumors.
assessment of any biologic parameter used to select antitumor treat- Our data do not confirm the hypothesis that aromatase inhibi- ments. Several previous studies have underscored the need for internal tors may offer a particular advantage over tamoxifen in patients whose and external quality control in the immunohistochemical analysis of tumors express ER but not PgR raised by Dowsett et al16 based on local protein expression.23 The authors highlighted the lack of any assess- laboratory PgR assessment in the ATAC trial. Our results conform ment of observer errors in the interpretation of results in most studies more closely with the findings of no effect of PgR on relative efficacy ofaromatase inhibitor and tamoxifen in the IES9 and the lack of signifi-cant difference seen in the ARNO/ABCSG trials.4 AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS
OF INTEREST
Although all authors completed the disclosure declaration, the followingauthor(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information aboutASCO’s conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section inInformation for Contributors. Employment or Leadership Position: None Consultant or Advisory
Role: Aron Goldhirsch, Novartis (C) Stock Ownership: Beat
Fig 4. STEPP (Subpopulation Treatment Effect Pattern Plot) analysis of 4-year
Thu¨rlimann, Novartis Honoraria: Aron Goldhirsch, Novartis Research
disease-free survival (DFS) according to centrally assessed progesterone recep-tor (PgR) expression among patients whose tumors were estrogen receptor Funding: Patrick Neven, Novartis; Aron Goldhirsch, Novartis Expert
expressing (Ն 1% ER) by central assessment.
Testimony: None Other Remuneration: None
Information downloaded from jco.ascopubs.org and provided by at Univ of Texas Libraries on January 23, 2011 from Copyright 2007 American Society of Clinical Oncology. All rights reserved.
Viale et al
Collection and assembly of data: Giuseppe Viale, Meredith M. Regan,
AUTHOR CONTRIBUTIONS
Mauro G. Mastropasqua, Patrizia Dell’Orto, Karen N. Price
Data analysis and interpretation: Giuseppe Viale, Meredith M. Regan,
Conception and design: Giuseppe Viale, Richard D. Gelber, Monica
Eugenio Maiorano, Mauro G. Mastropasqua, Alan S. Coates Castiglione-Gertsch, Aron Goldhirsch, Alan S. Coates Manuscript writing: Giuseppe Viale, Meredith M. Regan, Eugenio
Administrative support: Giuseppe Viale, Richard D. Gelber, Monica
Maiorano, Mauro G. Mastropasqua, Karen N. Price, Alan S. Coates Castiglione-Gertsch, Karen N. Price, Barry A. Gusterson Final approval of manuscript: Giuseppe Viale, Meredith M. Regan,
Provision of study materials or patients: Giuseppe Viale,
Eugenio Maiorano, Mauro G. Mastropasqua, Patrizia Dell’Orto, Birgitte Mauro G. Mastropasqua, Patrizia Dell’Orto, Birgitte Bruun Bruun Rasmussen, Johnny Raffoul, Patrick Neven, Zsolt Orosz, Stephen Rasmussen, Johnny Raffoul, Patrick Neven, Zsolt Orosz, ¨ hlschlegel, Beat Thu¨rlimann, Richard D. Gelber, Monica Castiglione-Gertsch, Karen N. Price, Aron Goldhirsch, Barry A.
9. Coombes RC, Kilburn LS, Snowdon CF, et al:
18. Allred DC, Harvey JM, Berardo M, et al:
REFERENCES
Survival and safety of exemestane versus tamoxifen Prognostic and predictive factors in breast cancer by after 2-3 years’ tamoxifen treatment (Intergroup immunohistochemical analysis. Mod Pathol 11:155- 1. Baum M, Budzar AU, Cuzick J, et al: Anastro-
Exemestane Study): A randomised controlled trial.
zole alone or in combination with tamoxifen versus 19. Regan MM, Viale G, Mastropasqua MG, et al:
tamoxifen alone for adjuvant treatment of post- 10. Goss PE, Ingle JN, Martino S, et al: random-
Re-evaluating adjuvant breast cancer trials: Assess- menopausal women with early breast cancer: First ized trial of letrozole following tamoxifen as ex- ing hormone receptor status by immunohistochem- results of the ATAC randomised trial. Lancet 359: tended adjuvant therapy in receptor-positive breast ical versus extraction assays. J Natl Cancer Inst cancer: Updated findings from NCIC CTG MA.17.
2. BIG 1-98 Collaborative Group: A comparison
20. Kaplan EL, Meier P: Nonparametric estima-
of letrozole and tamoxifen in postmenopausal 11. Howell A, Cuzick J, Baum M, et al: Results of
tion from incomplete observation. J Am Stat Assoc women with early breast cancer. N Engl J Med the ATAC (Arimidex, Tamoxifen, Alone or in Combi- nation) trial after completion of 5 years’ adjuvant 21. Cox DR: Regression models and life tables
3. Coombes RC, Hall E, Gibson LJ, et al: A
treatment for breast cancer. Lancet 365:60-62, 2005 (with discussion). J Royal Stat Soc B 34:187-220, randomized trial of exemestane after two to three 12. Jonat W, Gnant M, Boccardo F, et al: Effec-
years of tamoxifen therapy in postmenopausal tiveness of switching from adjuvant tamoxifen 22. Bonetti M, Gelber RD: A graphical method to
women with primary breast cancer. N Engl J Med to anastrozole in postmenopausal women with assess treatment-covariate interactions using the hormone-sensitive early-stage breast cancer: A Cox model on subsets of the data. Stat Med 19: 4. Jakesz R, Jonat W, Gnant M, et al: Switching
meta-analysis. Lancet Oncol 7:991-996, 2006 13. Bardou VJ, Arpino G, Elledge RM, et al: Pro-
23. Kirkegaard T, Edwards J, Tovey S, et al:
responsive early breast cancer to anastrozole after 2 gesterone receptor status significantly improves Observer variation in immunohistochemical analysis years’ adjuvant tamoxifen: Combined results of outcome prediction over estrogen receptor status of protein expression, time for a change? Histopa- ABCSG trial 8 and ARNO 95 trial. Lancet 366:455- alone for adjuvant endocrine therapy in two large breast cancer databases. J Clin Oncol 21:1973- 24. Perez EA, Suman VJ, Davidson NE, et al:
5. Boccardo F, Rubagotti A, Puntoni M, et al:
HER2 testing by local, central, and reference labo- Switching to anastrozole versus continued tamox- 14. Clinical practice guidelines for the use of
ratories in specimens from the North Central Cancer ifen treatment of early breast cancer: Preliminary tumor markers in breast and colorectal cancer: Treatment Group N9831 intergroup adjuvant trial.
results of the Italian Tamoxifen Anastrozole Trial.
Adopted on May 17, 1996 by the American Society of Clinical Oncology. J Clin Oncol 14:2843-2877, 25. Reference deleted by author.
6. Goss PE, Ingle JN, Martino S, et al: A random-
26. Collins LC, Botero ML, Schnitt SJ: Bimodal
ized trial of letrozole in postmenopausal women 15. Goldhirsch A, Glick JH, Gelber RD, et al:
frequency distribution of estrogen receptor immu- after five years of tamoxifen therapy for early-stage Meeting highlights: International expert consensus nohistochemical staining results in breast cancer: breast cancer. N Engl J Med 349:1793-1802, 2003 on the primary therapy of early breast cancer 2005.
An analysis of 825 cases. Am J Clin Pathol 123:16- 7. Boccardo F, Rubagotti A, Aldrighetti D, et al:
Switching to an aromatase inhibitor provides mortal- 16. Dowsett M, Cuzick J, Wale C, et al: retrospec-
27. Harvey JM, Clark GM, Osborne CK, et al:
ity benefit in early breast carcinoma: Pooled analysis tive analysis of time to recurrence in the ATAC trial Estrogen receptor status by immunohistochemistry of 2 consecutive trials. Cancer, Feb 12 [epub ahead is superior to the ligand-binding assay for predicting hypothesis-generating study. J Clin Oncol 23:7512- response to adjuvant endocrine therapy in breast 8. Coates AS, Keshaviah A, Thurlimann B, et al:
Five years of letrozole compared with tamoxifen as 17. Dowsett M, Allred DC, on behalf of the
28. Baehner FL, Habel LA, Quesenberry CP, et al:
initial adjuvant therapy for postmenopausal women TransATAC Investigators: Relationship between Quantitative RT-PCR analysis of ER and PR by with endocrine-responsive early breast cancer: Up- quantitative ER and PgR expression and HER2 sta- Oncotype DX™ indicates distinct and different as- date of study BIG 1-98. J Clin Oncol 25:486-492, tus with recurrence in the ATAC trial. Breast Cancer sociations with prognosis and prediction of tamox- Res Treat 100:S21, 2006 (abstr 48, suppl 1) ifen benefit. Breast Cancer Res Treat 100:S20, 2006 Acknowledgment
We thank the many pathologists who submitted tumor blocks and slides, Rosita Kammler, and Stefania Andrighetto. We thank the patients, physicians, nurses, and data managers who participated in this clinical trial; Novartis; and the IBCSG (funded by the Swedish Cancer Society, The Cancer Council Australia, ANZ-BCTG, FSTRF, the Swiss Group for Clinical Cancer Research [SAKK], US National Cancer Institute [CA-75362], and the Foundation for Clinical Cancer Research of Eastern Switzerland [OSKK]).
Appendix
The Appendix is included in the full-text version of this article, available online at www.jco.org. It is not included in the PDF version Information downloaded from jco.ascopubs.org and provided by at Univ of Texas Libraries on January 23, 2011 from Copyright 2007 American Society of Clinical Oncology. All rights reserved.

Source: http://www.zo.utexas.edu/courses/kalthoff/bio137/pdfs/11Spring/Viale%20et%20al.(2007).pdf