J Clin Oncol 23: by American Society of Clinical Oncology INTRODUCTION

Transcription

1 VOLUME 23 NUMBER 30 OCTOBER JOURNAL OF CLINICAL ONCOLOGY O R I G I N A L R E P O R T Retrospective Analysis of Time to Recurrence in the ATAC Trial According to Hormone Receptor Status: An Hypothesis-Generating Study Mitch Dowsett, Jack Cuzick, Chris Wale, Tony Howell, Joan Houghton, and Michael Baum From the Royal Marsden Hospital; Cancer Research UK; University College London, London; Christie Hospital, Manchester, UK. Submitted February 3, 2005; accepted August 4, Authors disclosures of potential conflicts of interest are found at the end of this article. Address reprint requests to Mitch Dowsett, Royal Marsden Hospital London, SW3 6JJ United Kingdom; mitch.dowsett@icr.ac.uk by American Society of Clinical Oncology X/05/ /$20.00 DOI: /JCO A B S T R A C T Purpose Arimidex, tamoxifen alone, or in combination (ATAC) trial of anastrozole (Arimidex) versus tamoxifen or a combination of the two in 9,366 postmenopausal patients with primary breast cancer found a significant improvement in disease-free survival and time to recurrence (TTR) for anastrozole compared with tamoxifen, that was restricted to patients with hormone receptor-positive (ie, estrogen receptor positive [ER ] and/or progesterone receptorpositive [PgR ]) disease, the target population for these therapies. We retrospectively tested the hypothesis that this benefit might differ according to PgR status. Patients and Methods TTR was compared between the three treatment groups for subgroups defined by ER and PgR status using Cox s proportional hazards model, with and without adjustment for baseline variables. Results The unadjusted hazard ratio (HR) for anastrozole versus tamoxifen for TTR was 0.74 (95% CI, 0.64 to 0.87) for women with either ER or PgR tumors. In the ER /PgR subgroup (n 3,834) the HR was 0.84 (95% CI, 0.69 to 1.02) compared with 0.43 (95% CI, 0.31 to 0.61) in the ER /PgR-negative (PgR ) subgroup (n 880). In the adjusted model the HRs were 0.83 and 0.45, respectively. Conclusion Time to recurrence was longer for anastrozole- than tamoxifen-treated patients in both ER /PgR and ER /PgR subgroups, but the benefit was substantially greater in the PgR subgroup. As this was an exploratory analysis, this effect should be considered as hypothesis generating and assessed prospectively in other trials comparing the adjuvant use of an aromatase inhibitor with tamoxifen. J Clin Oncol 23: by American Society of Clinical Oncology INTRODUCTION For the past 20 years, tamoxifen has been the mainstay of adjuvant therapy in the treatment of hormone receptor-positive breast cancer. Overview analyses have revealed that for patients with estrogen receptor positive (ER ) disease, 5 years of adjuvant treatment with tamoxifen reduced recurrence and mortality rate by 47% and 26%, respectively. 1 In postmenopausal patients the potent third-generation aromatase inhibitor, anastrozole, has been found to be more effective than tamoxifen in the treatment of advanced disease, as preoperative therapy, as initial adjuvant therapy in early breast cancer 2-5 and also when given after 2 to 3 or 5 years of adjuvant tamoxifen. 6,7 Similar results have been seen for other aromatase inhibitors in various settings. 3,8,9 Thus, aromatase inhibitors promise to deliver greater disease control for hormone-dependent 7512

2 Time to Recurrence in the ATAC Trial breast cancer and also appear to be better tolerated than tamoxifen. The improved efficacy of aromatase inhibitors over tamoxifen appears, at least in part, to be due to the partial agonist activity of tamoxifen, which limits its antitumor efficacy. 10 Laboratory studies indicate that the degree to which tamoxifen acts as an agonist, and therefore the degree to which its effectiveness may be less than that with estrogen deprivation provided by the aromatase inhibitors, is determined by molecular characteristics that vary between cell lines and breast carcinomas. 11,12 Analysis of these molecular features might therefore reveal subpopulations that derive greater or lesser degrees of benefit from aromatase inhibitors relative to tamoxifen. Such data may allow better selection of therapy, improve the understanding of the molecular endocrinology of breast cancer and allow the rational development of combinations of novel targeted agents with these endocrine therapies. Expression of the progesterone receptor (PgR), a key molecular marker in breast cancer, is generally found to be highly estrogen dependent and its expression has been considered to be indicative of an intact ER mechanism. 13 This is consistent with its absence in almost all estrogen receptornegative (ER ) tumors. In a recent retrospective study of the clinical outcomes of nearly 16,000 patients, it was found that PgR status had modest prognostic significance among systemically untreated patients. However, it was independently associated with both disease-free and overall survival in tamoxifen-treated patients, with patients having PgRnegative (PgR ) disease doing significantly worse than those with PgR-positive (PgR ) disease. 14 In contrast, the overview analysis of randomized trials of adjuvant tamoxifen treatment found that the recurrence of ER disease was reduced to a similar degree irrespective of whether tumors were PgR or PgR (with reductions in recurrence rates of 37% and 32%, respectively). 1 To date there is no information on the relationship of PgR status to the benefit derived from aromatase inhibitors in the adjuvant setting. We have therefore retrospectively assessed the effect of PgR status on time to recurrence (TTR) in the arimidex, tamoxifen, alone and in combination (ATAC) trial. The ATAC trial compared adjuvant treatment with anastrozole (arimidex) alone versus tamoxifen alone versus a combination of both for 5 years in postmenopausal women with early-stage, operable breast cancer. 2,4,5 The results of this study were published in abstract form after a median follow-up of 47 months. 15 The results reported here are after a median follow-up of 68 months. PATIENTS AND METHODS Patients The ATAC trial was a (1:1:1) randomized, double-blind trial of the adjuvant use of 1 mg of anastrozole versus 20 mg of tamoxifen versus a combination of the two for 5 years in postmenopausal women with early invasive breast cancer. Eligible patients were postmenopausal women with histologically proven, operable, invasive breast cancer who had completed primary surgery and radiotherapy or chemotherapy (when given). Detailed aspects of the study design, its conduct and overall results have been published elsewhere. 2,4,5 Median follow-up at the time of this analysis was 68 months. 5 After the first analysis, with a median 33 months follow-up, the combination arm was discontinued and patients were offered the opportunity to switch to anastrozole or tamoxifen. 5 The ATAC trial was initiated in Eligibility for entry was on the basis of patients being suitable for hormonal therapy rather than having ER and/or PgR disease. Data were collected on case record forms as ER, ER, or ER unknown (ERuk) and PgR, PgR, or PgR unknown (PgRuk) according to local criteria. Where no information was recorded for one or other receptor, this was treated as unknown status for the respective receptor. A total of 20% of patients were initially reported as receptor unknown, but before the first major analysis of the ATAC trial tumor blocks were retrieved from approximately 60% of these (ie, 12% overall). Immunohistochemical analyses were performed on these retrieved blocks at one of two central laboratories and the Allred score 16 was used to quantify the receptor staining with the samples recorded as positive if the score was greater than 2. In our previous reports 2,4,5 we aggregated the ER/PgR subgroups as hormone receptor-positive (ER and/or PgR including patients where the result for one receptor was unknown, n 7,839), hormone receptor negative (ER /PgR or ER / PgRuk, n 768) or unknown (all other groups, n 759). No records were made of the methodology or cut-off points used to define positivity in individual centers. The primary trial end points were disease-free survival, safety, and tolerability. Secondary trial end points included TTR, overall survival, and incidence of contralateral breast cancer. Since this was a study of the impact of the biology of tumors on treatment outcome, TTR was chosen as the most informative end point and defined as the time from randomization until the reappearance of breast cancer tumors (including contralateral tumors, but excluding patients who had died from non breast cancer causes before recurrence) with a further analysis excluding contralateral cancers as these are unlikely to be directly influenced by the interactions of therapy and the biology of the original disease. Statistical Analyses The data were retrospectively analyzed for TTR separately for each of the PgR subgroups by use of Cox s proportional hazards model. The comparisons conducted were between tamoxifen and anastrozole and between tamoxifen and the combination as these were the scheduled comparisons for the main trial. Separate analyses were conducted for these two comparisons using separate models. Additional analyses were performed with adjustment for baseline characteristics for the ER /PgR and ER /PgR subgroups. The variables included were: nodal status (negative or unknown; one to three nodes positive; greater than three nodes positive), tumor size ( 2 cm; 2 cm), tumor grade (well differentiated, moderately differentiated, poorly differentiated, or undifferentiated), and previous adjuvant chemotherapy (no; yes). The cut-off points for the adjustment factors were decided on before the adjustment analyses. All significant baseline variables were then included in multivariate models. The validity of the proportional hazard assumption was also examined. Analysis was conducted in STATA (version 8.0). All P values are two sided and

3 Dowsett et al 95% CIs are given. Due to subgroup analyses, retrospective evaluation, and multiple comparisons, the data presented should be regarded as exploratory and hypothesis generating. The P values and CI have not been adjusted for multiple comparisons, and should only be considered as descriptive measures of the strength of association. RESULTS The numbers and proportions in each receptor subgroup of the total of 9,366 patients are shown in Table 1. There were 8,004 (85%) patients in whom the status of both receptors was known, leaving 1,362 (15%) in whom at least one of the two receptors remained unknown. The subgroups of ER / PgRuk (n 518) and ERuk/PgRuk (n 743) are shown for completeness, but data from the other three small groups in which the status of one of the receptors was unknown (total n 101) are not shown. The proportions of patients in the four subgroups in which both ER and PgR were known were as follows: ER /PgR 71%; ER /PgR 17%; ER /PgR 3%; ER /PgR 9%. Outcome data for the combination arm are shown throughout without censoring at the time of any switch to anastrozole or tamoxifen. The hazard ratios (HRs) for breast cancer events for anastrozole versus tamoxifen and combination versus tamoxifen are shown for the different receptor groups in Table 1. A test of interaction between the three categories of PgR status (positive, negative, unknown) and treatment for ER disease was highly significant (P.0004). The P values should be interpreted with caution because of the retrospective nature of the analysis with the generation of multiple subgroups. In the ER /PgR subgroup, the HR was close to unity for both of the between-treatment comparisons, as would be expected (with both treatments being equally ineffective). For the ER /PgR subgroup the HR was 0.84 (95% CI, 0.69 to 1.02), which is an observed 16% reduction in breast cancer events on anastrozole compared with tamoxifen. This compares with the HR of 0.74 (95% CI, 0.64 to 0.87) found for the whole hormone receptor positive study group. 5 For the ER /PgR group the HR was lower at 0.43 (95% CI, 0.31 to 0.61), which was an observed 57% reduction in breast cancer events on anastrozole compared with tamoxifen. TTR is shown as Kaplan-Meier curves in Figure 1 (A) for the ER /PgR subgroup and (B) for the ER /PgR subgroup. These plots confirm that the TTR for the anastrozole-treated group was better than either the tamoxifen or combination-treated arm in both of these subgroups, while the outcome for these last two treatment arms was similar. The separation between the anastrozole curve and the tamoxifen and combination curves is greater for the ER /PgR subgroup compared with the ER /PgR subgroup. This appears to result from patients on tamoxifen and the anastrozole/ tamoxifen combination faring substantially worse in the ER / PgR subgroup than in the ER /PgR subgroup. In contrast, the outcome for patients treated with anastrozole was only marginally worse for patients in the ER /PgR subgroup compared with the ER /PgR subgroup, with 11% and 10% of patients, respectively, having a recurrence over the 68-month follow-up (Table 1). The HRs quoted above are unadjusted. HRs were also calculated after adjusting for baseline nodal status, tumor size, tumor grade, and the use of adjuvant chemotherapy before hormonal adjuvant therapy. As already shown, the HR in the unadjusted model for ER-positive disease was 0.84 for PgR and 0.43 for PgR tumors. In the adjusted model the HRs were 0.83 (95% CI, 0.68 to 1.00) and 0.45 (95% CI, 0.32 to 0.63), respectively. The data presented in Table 1 include contralateral tumor incidence in the recurrence rates. The data for the ER/ PgR and ER /PgR subgroups are shown after excluding contralateral breast cancer in Table 2. The HRs were 0.83 overall, 0.79 for the hormone receptor-positive Table 1. Number of Breast Cancer Events in the Three Treatment Groups According to Estrogen Receptor/Progesterone Receptor Status, and Hazard Ratios of Events in Each Subgroup for Anastrozole versus Tamoxifen and Tamoxifen versus Combination No. of Patients Breast Cancer Events A T C A v T C v T No. % No. % No. % HR 95% CI P HR 95% CI P ER /PgR 5, to to ER /PgR 1, to to ER /PgR to to ER /PgR to to ER /PgRuk to to ERuk/PgRuk to to Other 101 Totals 9, to to Abbreviations: A, anastrozole; T, tamoxifen; C, combination of; ER, estrogen receptor; PgR, progesterone receptor; uk, unknown JOURNAL OF CLINICAL ONCOLOGY

4 Time to Recurrence in the ATAC Trial Fig 1. Time to recurrence in the (A) ER-positive/PgR-positive and the (B) ER-positive/PgR-negative patient subgroups. group, 0.92 for the ER /PgR subgroup, and 0.42 for the ER /PgR subgroup (unadjusted). DISCUSSION The ATAC trial has reported a significant improvement in TTR for anastrozole versus tamoxifen in the hormone receptor positive trial population, with a HR after 68 months median follow-up of 0.74 (95% CI, 0.64 to 0.87; P.0002). 5 There was no significant difference between tamoxifen and the combination. The retrospective analyses conducted here revealed that in the ATAC trial anastrozole-treated patients with ER /PgR tumors had only half the recurrences over the follow-up period as patients treated with tamoxifen. This compares with the approximately 16% lower recurrence rate seen with anastrozole relative to tamoxifen in the patients with ER /PgR tumors. The 95% CI for benefit in the ER / PgR tumors includes unity but this is likely to have resulted from the reduced power associated with subgroup analysis. The validity of the result is supported by the similar outcome of tamoxifen- and combination-treated patients, in both the ER /PgR and ER /PgR subgroups. It should be noted, however, that there was no a priori hypothesis that the relative benefit would be greater for ER / PgR than ER /PgR tumors. Thus, while this result is consistent with other clinical and biologic observations, it should be considered as hypothesis generating and requires confirmation before it influences clinical decision making. Until this is the case the overall trial result should influence clinical practice rather than this retrospective subgroup analysis. PgR status should not be regarded as a reason for denying a patient anastrozole. Coombes et al 9 reported that after 2 to 3 years of adjuvant therapy with tamoxifen, and the aromatase inhibitor, exemestane, was equally effective in patients with ER /PgR and ER /PgR tumors (HRs 0.66 and 0.58, respectively). However, that trial was conducted in patients who had not relapsed during their first 2 to 3 years of adjuvant tamoxifen treatment, preventing the possibility of conducting the same analysis presented here. In particular it is notable that more than 15% of ER PgR patients on tamoxifen in the ATAC trial recurred throughout the first 3 years. Another adjuvant trial of 5 years tamoxifen versus a nonsteroidal aromatase inhibitor, in this case letrozole (the BIG 1-98 trial), has recently reported, but not yet published, improved disease-free survival with the aromatase inhibitor that was similar for both ER /PgR and ER /PgR patients. The reasons for this discordance with the ATAC trial are unclear at this stage but possibilities include a drugspecific effect, different methods for PgR testing (heterogenous in both studies), and recruitment from different patient populations (ATAC: 35% from UK and 24% from the United States; BIG 1-98: 4% from UK and none from the United States). Table 2. Number of Breast Cancer Events, Excluding Contralateral Tumors, in the Three Treatment Groups According to Estrogen Receptor/Progesterone Receptor Status, and Hazard Ratios of Events in Each Subgroup for Anastrozole Versus Tamoxifen and Tamoxifen Versus Combination Breast Cancer Events A T C A v T C v T No. of Patients No. % No. % No. % HR 95% CI P HR 95% CI P ER /PgR 5, to to ER /PgR 1, to to Total 9, to to Abbreviations: A, anastrozole; T, tamoxifen; C, combination of; ER, estrogen receptor; PgR, progesterone receptor; uk, unknown

5 Dowsett et al In addition to being retrospective, another disadvantage of this study was that the ER and PgR status was determined by a variety of assays, positivity was ascribed according to locally assigned cut-off points, and in most cases these values were not recorded. However, such diverse methodology is likely to weaken the likelihood of detecting an improved outcome in the ER /PgR group. Our data cannot indicate whether the difference in the relative efficacy of anastrozole and tamoxifen according to PgR status is a direct result of this status or is due to segregation of PgR with another biologic determinant or determinants. For example, it has recently been demonstrated that PgR levels in breast cancer cell lines are downregulated by the growth factors insulin-like growth factor (IGF-1), heregulin, and epidermal growth factor (EGF). 17 In breast carcinomas, there is an inverse relationship between HER2 and PgR expression, 18 and we have found in another set of samples that HER2 and/or EGF receptor were positive in 8.2% of ER /PgR (n 427) and 29.5% of ER /PgR (n 190) breast carcinomas (Dowsett et al, in preparation). These data are consistent with a recent preliminary report from Arpino et al. 19 It has been reported that the improved response to the aromatase inhibitor letrozole over tamoxifen in the preoperative setting was substantially greater in tumors that were HER2 and/or EGFR. 20 Furthermore, a similar improved response rate in this clinical setting has also been observed for anastrozole over tamoxifen in HER2 tumors. 21 Therefore, the relatively greater benefit of anastrozole over tamoxifen in ER /PgR tumors might be explained, at least in part, by a greater expression of type 1 growth factor receptors in this group. In conclusion, in the ATAC trial adjuvant treatment with anastrozole resulted in a significant increase in TTR compared with tamoxifen in patients with hormone-sensitive tumors and these efficacy benefits were accompanied by tolerability advantages. Retrospective subgroup analyses reported here showed that TTR was longer for anastrozole- than tamoxifentreated patients in both the ER /PgR and the ER /PgR subgroups of patients, but the differential benefit was greater in ER /PgR tumors. These data are exploratory, should be considered hypothesis generating, and should be confirmed prospectively in other trials comparing the adjuvant use of an aromatase inhibitor with tamoxifen. Additional markers such as HER2 are being assessed to determine if these provide a mechanistic explanation for this finding. Acknowledgment We thank the trial investigators, monitors, nurses, data managers, pathologists, other support staff members, and most importantly the patients participating in the trial. Authors Disclosures of Potential Conflicts of Interest Although all authors completed the disclosure declaration, the following authors or their immediate family members indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. For a detailed description of the disclosure categories, or for more information about ASCO s conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors. Authors Employment Leadership Consultant Stock Honoraria Research Funds Testimony Other Mitch Dowsett AstraZeneca (A); Novartis (B) AstraZeneca (A); Novartis (A); Pfizer AstraZeneca (C); Novartis (C) (A) Jack Cuzick AstraZeneca (B); Eli Lilly (A); Novartis (A) Tony Howell AstraZeneca (B); Eli Lilly (A); Novartis (A) AstraZeneca (B); Novartis (A); Pharmacia (A) Joan Houghton AstraZeneca (C) AstraZeneca (B) Michael Baum AstraZeneca (B) AstraZeneca (B) Dollar Amount Codes (A) $10,000 (B) $10,000-99,999 (C) $100,000 (N/R) Not Required 7516 JOURNAL OF CLINICAL ONCOLOGY

6 Time to Recurrence in the ATAC Trial REFERENCES 1. Early Breast Cancer Trialists Collaborative Group: Tamoxifen for early breast cancer: An overview of the randomised trials. Lancet 351: , The ATAC Trialists Group: Anastrozole alone or in combination with tamoxifen versus tamoxifen alone for adjuvant treatment of postmenopausal women with early breast cancer: First results of the ATAC randomised trial. Lancet 359: , Smith IE, Dowsett M: Aromatase inhibitors in breast cancer. N Engl J Med 348: , The ATAC Trialists Group: Anastrozole alone or in combination with tamoxifen versus tamoxifen alone for adjuvant treatment of postmenopausal women with early-stage breast cancer: Results of the ATAC (Arimidex, Tamoxifen, Alone or in Combination) trial efficacy and safety update analyses. Cancer 98: , The ATAC Trialists Group: Results of the ATAC (Arimidex, Tamoxifen, Alone or in Combination) trial after completion of 5 years adjuvant treatment for breast cancer. Lancet 365:60-62, Jakesz R, Jonat W, Gnant M, et al: Switching of postmenopausal women with endocrineresponsive early breast cancer to anastrozole after 2 years adjuvant tamoxifen: Combined results of ABCSG trial 8 and ARNO 95 trial. Lancet 366: , Jakesz R, Samonigg H, Greil R, et al: Extended adjuvant treatment with anastrozole: Results from the Austrian Breast and Colorectal Cancer Study Group Trial 6a (ABCSG-6a). Proc Am Soc Clin Oncol 23:10s, 2003 (abstr 527) 8. Goss PE, Ingle JN, Martino S, et al: A randomized trial of letrozole in postmenopausal women after five years of tamoxifen therapy for early-stage breast cancer. N Engl J Med 349: , Coombes RC, Hall E, Gibson LJ, et al: A randomized trial of exemestane after two to three years of tamoxifen therapy in postmenopausal women with primary breast cancer. N Engl J Med 350: , Dowsett M, Howell A: Breast cancer: Aromatase inhibitors take on tamoxifen. Nat Med 8: , Kurokawa H, Lenferink AE, Simpson JF, et al: Inhibition of HER2/neu (erbb-2) and mitogenactivated protein kinases enhances tamoxifen action against HER2-overexpressing, tamoxifenresistant breast cancer cells. Cancer Res 60: , Osborne CK, Bardou V, Hopp TA, et al: Role of the estrogen receptor coactivator AIB1 (SRC-3) and HER-2/neu in tamoxifen resistance in breast cancer. J Natl Cancer Inst 95: , Jensen EV, DesSombre ER: Steroid hormone binding and hormone receptors, in Holland JF, Frei Spaceiiiqq E, Bast Jr RC, et al (eds): Cancer Medicine (ed 4 th ). Baltimore, MD, Williams and Wilkins, 1996, pp Bardou VJ, Arpino G, Elledge RM, et al: Progesterone receptor status significantly improves outcome prediction over estrogen receptor status alone for adjuvant endocrine therapy in two large breast cancer databases. J Clin Oncol 21: , Dowsett M, on behalf of the ATAC Trialists Group: Analysis of time to recurrence in the ATAC (arimidex, tamoxifen, alone or in combination) trial according to estrogen receptor and progesterone receptor status. Breast Cancer Res Treat 82:S7, 2003 (suppl 1; abstr 4) 16. Harvey JM, Clark GM, Osborne CK, et al: Estrogen receptor status by immunohistochemistry is superior to the ligand-binding assay for predicting response to adjuvant endocrine therapy in breast cancer. J Clin Oncol 17: , Cui X, Zhang P, Deng W, et al: Insulin-like growth factor-i inhibits progesterone receptor expression in breast cancer cells via the phosphatidylinositol 3-kinase/Akt/mammalian target of rapamycin pathway: Progesterone receptor as a potential indicator of growth factor activity in breast cancer. Mol Endocrinol 17: , Konecny G, Pauletti G, Pegram M, et al: Quantitative association between HER-2/neu and steroid hormone receptors in hormone receptor-positive primary breast cancer. J Natl Cancer Inst 95: , Arpino G, Weiss H, Lee A, et al: Estrogen receptor positive (ER ), progesterone receptor negative (PgR-) breast cancer: New insights into molecular mechanisms and clinical implications. Breast Cancer Res Treat 88:S21, Ellis MJ, Coop A, Singh B, et al: Letrozole is more effective neoadjuvant endocrine therapy than tamoxifen for ErbB-1- and/or ErbB-2-positive, estrogen receptor-positive primary breast cancer: Evidence from a phase III randomized trial. J Clin Oncol 19: , Smith IE, Dowsett M, Ebbs SR, et al: Neoadjuvant treatment of postmenopausal breast cancer with anastrozole, tamoxifen, or both in combination: The Immediate Preoperative Anastrozole, Tamoxifen, or Combined with Tamoxifen (IMPACT) multicenter double-blind randomized trial. J Clin Oncol 23: ,