Aerio-oncologie.org

Longer-Term Outcomes of Letrozole Versus Placebo After 5Years of Tamoxifen in the NCIC CTG MA.17 Trial: AnalysesAdjusting for Treatment CrossoverHuan Jin, Dongsheng Tu, Naiqing Zhao, Lois E. Shepherd, and Paul E. Goss See accompanying editorial on page 684 and articles on pages 709 and 722; listen to the podcast by Dr. Mayer at www.jco.org/podcasts hai, China; Dongsheng Tu and Lois E.
Purpose
The interim analysis of the National Cancer Institute of Canada Clinical Trials Group MA.17 trial
showed that letrozole was significantly better than placebo in disease-free survival (DFS) for postmenopausal women with hormone receptor–positive breast cancer following about 5 years of tamoxifen therapy. When patients were unblinded, those on placebo were offered letrozole.
Longer-term efficacy of letrozole, especially survival, was of particular interest because the median follow-up of the first interim analysis was only 2.5 years. Efficacy was difficult to assess because more than 60% of placebo patients crossed over to letrozole after being unblinded.
Patients and Methods
Two statistical approaches were used to adjust for the potential effects of treatment crossover: one was based on the inverse probability of censoring weighted (IPCW) Cox model and the other on a Cox model with time-dependent covariates.
Clinical Trials repository link available on With a median follow-up of 64 months, the hazard ratios (HRs) of letrozole and placebo from the IPCW analyses were HR of 0.52 (95% CI, 0.45 to 0.61; P Ͻ .001) for DFS, HR of 0.51 (95% CI, 0.42 to 0.61; P Ͻ .001) for distant disease-free survival (DDFS), and HR of 0.61 (95% CI, 0.52 to 0.71; P Ͻ .001) for overall survival (OS). The results from the analyses based on the Cox model with Research Institute, Queen’s University, time-dependent covariates were similar for letrozole and placebo: HR of 0.58 (95% CI, 0.47 to K7L 3N6; e-mail: [email protected].
0.72; P Ͻ .001) for DFS, HR of 0.68 (95% CI, 0.52 to 0.88; P ϭ .004) for DDFS, and HR of 0.76(95% CI, 0.60 to 0.96; P ϭ .02) for OS.
Conclusion
Exploratory analyses based on longer follow-up and adjusting for treatment crossover suggest that
extended adjuvant letrozole was superior to placebo in DFS, DDFS, and OS.
J Clin Oncol 30:718-721. 2011 by American Society of Clinical Oncology the O’Brien-Fleming stopping boundary for DFS INTRODUCTION
and a trend toward an overall survival (OS) advan- MA.17 was a double-blind, placebo-controlled trial tage, the independent data and safety monitoring conducted by the National Cancer Institute of Can- committee recommended stopping the trial and un- ada Clinical Trials Group (NCIC CTG) that evalu- blinding the study participants. Patients randomly ated the use of letrozole in the extended adjuvant assigned to placebo were offered the option of re- setting following 5 years of treatment with tamox- ceiving letrozole for a period of 5 years after they ifen in postmenopausal women with hormone were unblinded. Those on letrozole were provided receptor–positive early-stage breast cancer. Between the balance of their 5-year treatment and were ob- 1998 and 2002, 5,187 women were enrolled onto the study. The first protocol-specified interim anal- The median follow-up of all patients at the ysis was conducted in August 2003 after 40% of time of unblinding was 30 months. Longer-term the events needed for final analysis were observed.
effects of letrozole, especially its benefit in terms It showed that letrozole significantly improved of OS, was one of the major questions left unan- disease-free survival (DFS) compared with pla- swered by stopping the trial at the first interim cebo.1,2 Taking into consideration the crossing of analysis. Although follow-up of all patients 2011 by American Society of Clinical Oncology Information downloaded from jco.ascopubs.org and provided by at INSERM on May 4, 2013 from 193.54.110.33 Copyright 2012 American Society of Clinical Oncology. All rights reserved.
Longer-Term Outcomes of Letrozole in MA.17
continued after being unblinded, assessment of long-term effects of trial was unblinded in October 2003 on the recommendation of the Data letrozole is difficult because of selective crossover of patients. An Safety Monitoring Committee after a protocol-specified interim analysis dem- analysis of intent-to-treat (ITT) data3 performed on a database that onstrated superiority of letrozole over placebo in DFS. All patients randomlyassigned to the study were informed of the results from the interim analysis included data after patients were unblinded, which had a median and of their treatment allocation, and those receiving placebo were offered follow-up of 64 months from random assignment, showed that the letrozole for a planned period of 5 years. Follow-up of all patients continued difference in DFS was still highly significant between the two treat- after unblinding, and a database that included data after unblinding was locked ment arms: the hazard ratio (HR) of letrozole compared with placebo on July 28, 2006, for assessment of long-term effects of letrozole on clinical was 0.68 (95% CI, 0.55 to 0.83; P Ͻ .001), but no significant difference outcomes. This database was used for all analyses presented herein.
was found in distant disease-free survival (DDFS; HR, 0.80; 95% CI, Statistical Considerations
0.63 to 1.03; P ϭ .08) and OS (HR, 0.98; 95% CI, 0.78 to 1.22; P ϭ .85).
The clinical end points for this analysis are as defined in the original Results from this ITT analysis are difficult to interpret because almost MA.17 trial. Specifically, DFS, the primary end point, was defined as the time two thirds of the patients originally randomly assigned to receive from random assignment to the time of recurrence of the primary disease (in placebo chose to take letrozole after unblinding of their treatment breast, chest wall, nodal, or metastatic sites) or to the development of new assignment. Two additional analyses were performed: one using a contralateral breast cancer. Secondary end points included DDFS defined as landmark approach, which excluded patients randomly assigned to the time from random assignment to the time of recurrence in metastatic sites placebo who had had events before crossing over to active treatment and OS defined as time from random assignment to death from any cause.
For each time to an event end point, analyses based on two approaches and another based on the Cox model with treatment before and after that adjust for treatment crossover were performed. The first approach was switching as a time-dependent covariate. These two analyses com- based on an IPCW Cox regression model.11 This approach accounts for pared patients randomly assigned to placebo who had crossed over to selective treatment crossover by first censoring the original time to an event for letrozole after unblinding with those who remained on placebo, and women randomly assigned to placebo but who crossed over to letrozole after the analyses used data from the database after unblinding.4 The latter unblinding at the time of crossover and then recreates their new time to an analyses indicated that patients randomly assigned to placebo who event by weighting the time to an event of the women in the placebo groupwho had similar demographic and disease characteristics but who remained crossed over to letrozole, even after a substantial period of time since on placebo. The baseline factors that were significantly associated with the discontinuation of prior adjuvant tamoxifen, had improved DFS and probability of treatment crossover4 and also the specific end point at the 0.1 DDFS compared with those who did not cross over. However, the level were used to calculate the weights. A pooled logistic regression model,12 question of longer-term efficacy of adjuvant letrozole, defined as if no weighted on the basis of estimated conditional probabilities of having re- patient on placebo crossed over to letrozole,5 could not be answered by mained on placebo for women who crossed over, was used to estimate the HR of letrozole to placebo as if there was no treatment crossover. This approach is There have been several approaches proposed in the statistical The second approach was proposed by Shao et al10 (referred to as the literature that could be used to estimate longer-term clinical outcomes SCC approach on the basis of the first letters of the last names of three authors) of an experimental treatment in randomized clinical trials in the pres- and was based on a Cox model with a time-dependent treatment covariate: 1 ence of substantial treatment crossover. Korn and Freidlin6 have com- for women randomly assigned to letrozole and 0 for women randomly as- mented on some of the existing approaches. Recently, an approach signed to placebo until the time when they crossed over to letrozole and 1 using an inverse probability of censoring weighted (IPCW) Cox pro- afterward. Additional time-dependent covariates defined as the quadratic portional hazard model was used to adjust for selective crossover in functions of times when treatment was switched were also included in the assessing longer-term clinical efficacy of letrozole versus tamoxifen in model. Time of crossover was defined as infinity if women never switched.
These additional covariates measure crossover effect caused by the selective the adjuvant setting based on data from the Breast International nature of the treatment crossover. Baseline factors used for the IPCW analysis Group BIG 1-98 early-stage breast cancer adjuvant trial.7,8 This ap- were also included in the model as fixed covariates. HRs of letrozole to placebo, proach was discussed in some detail in a recent editorial in the Journal as if there was no treatment crossover, were derived from the coefficient of the of Clinical Oncology.9 In this study, we determined the longer-term clinical efficacy of letrozole in MA.17 by using the IPCW method. Inaddition, sensitivity analyses based on an approach that models thetreatment crossover effect in a Cox model with latent hazard rate10 were also performed to verify the robustness of our results from thisIPCW approach.
Patient Population
Characteristics of the women randomly assigned to placebo who switched to letrozole after the trial was unblinded can be found in Goss PATIENTS AND METHODS
et al.4 Briefly, among 2,587 patients who were originally randomlyassigned to receive placebo, 204 (7.9%) experienced recurrence or Study Design
death before the date of unblinding, 1,579 (61.0%) were confirmed to The details of the study design for MA.17 have been published previ- have crossed over to letrozole, and 804 (31.1%) elected no further ously.2 In brief, MA.17 was a phase III, randomized, double-blind, placebo- treatment after unblinding and were considered as having remained controlled clinical trial designed to investigate the efficacy of letrozole in on placebo. The median time from random assignment to treatment postmenopausal women with hormone receptor–positive primary breast can- crossover was 2.7 years (range, 1.1 to 7.0 years) for women who cer who were disease-free and within 3 months of completing approximately 5 crossed over from placebo to letrozole. A higher proportion of women years (range, 4.5 to 6 years) of adjuvant tamoxifen. Patients were randomlyassigned to letrozole (2.5 mg orally daily) or placebo for a planned 5 years. In who crossed over to letrozole were white compared with women who all, 5,187 women were enrolled but 17 patients were excluded from analysis did not cross over (91.9% v 90.0%; P ϭ .03) and young (median, 60.7 v because of noncompliance with good clinical practice guidelines. The MA.17 64.5 years; P Ͻ .001). Other characteristics more common in those 2011 by American Society of Clinical Oncology Information downloaded from jco.ascopubs.org and provided by at INSERM on May 4, 2013 from 193.54.110.33 Copyright 2012 American Society of Clinical Oncology. All rights reserved.
who crossed over to letrozole included an Eastern Cooperative Oncol- CI, 0.52 to 0.71; P Ͻ .001) for OS. By using the SCC approach for ogy Group (ECOG) performance status of 0 (92.3% v 86.8%; analysis, the HRs for letrozole versus placebo were 0.58 (95% CI, 0.47 P Ͻ .001), a longer period between initial diagnosis and random to 0.72; P Ͻ .001) for DFS, 0.68 (95% CI, 0.52 to 0.88; P ϭ .004) for assignment (median 64.7 v 63.7 months; P Ͻ .001), positive axillary DDFS, and 0.77 (95% CI, 0.61 to 0.97; P ϭ .03) for OS.
nodes (51.0% v 44.9%; P Ͻ .001), positive tumor hormone receptorstatus at diagnosis (98.3% v 95.5%; P ϭ .001), prior adjuvant chemo-therapy (49.2% v 37.1%; P Ͻ .001), and axillary node dissection DISCUSSION
(96.7% v 93.4%; P Ͻ .001). Among these characteristics, ethnicity(white v nonwhite), performance status (0 v 1 or 2), time from initial Although stopping at the first interim analysis at a median follow-up diagnosis to random assignment (Ͻ 5 v Ն 5 years), prior chemother- of 30 months in MA.17 provided clear-cut proof of the strong effect apy (yes v no), and pathologic N stage (0 v others) were associated with extended letrozole had on improving DFS and DDFS, the effect on OS DFS. Performance status (0 v 1 or 2), prior chemotherapy (yes v no), is still uncertain because the study was only powered to detect the and pathologic N stage (0 v others) were associated with DDFS. Age difference in the primary end point, which was DFS. The longer (Ͻ 70 v Ն 70 years), performance status (0 v 1 or 2), time from initial follow-up data with a median duration of 64 months in the database diagnosis to random assignment (Ͻ 5 v Ն 5 years), pathologic N stage after unblinding are useful for explaining this effect; however, results (0 v others), and prior chemotherapy (yes v no) were associated with from a direct and an ITT comparison of women randomly assigned to OS. These characteristics were used as covariates in the IPCW and letrozole and placebo are difficult to interpret because more than 60% SCC analyses, respectively, for each end point.
of women randomly assigned to placebo elected to cross over andreceive letrozole after unblinding. In the analyses described here, we Outcomes
used two approaches that adjusted for crossover from placebo to As reported previously, at a median follow-up of 64 months, the active treatment by using efficacy data from the post unblinding data- adjusted HRs for letrozole versus placebo in our ITT analysis were 0.68 base. Results from both of these two approaches suggest that letrozole (95% CI, 0.56 to 0.83; P Ͻ .001) for DFS, 0.81 (95% CI, 0.63 to 1.04; was statistically significantly superior to placebo in all three end points, P ϭ .09) for DDFS, and 0.99 (95% CI, 0.79 to 1.24; P ϭ .83) for OS.3 including OS, which was an important secondary end point in the Adjusting for treatment crossover, both IPCW and SCC analyses MA.17 trial but was not found to be significant from either the first showed significant improvements for letrozole versus placebo for all interim or post unblinding ITT analyses. IPCW and SCC approaches three clinical end points (Fig 1). In the IPCW analyses, the HRs for showed that letrozole potentially reduces the risk of death by 35% and letrozole and placebo were 0.52 (95% CI, 0.45 to 0.61; P Ͻ .001) for DFS, 0.51 (95% CI, 0.42 to 0.61; P Ͻ .001) for DDFS, and 0.61 (95% We also performed an analysis that adjusted the characteristics associated with probability of crossover and outcome by includingthem as covariates in an ordinary Cox regression model with a treat-ment variable. The results are shown in Figure 1. We may find thatthese results are almost the same as those from the simple ITT analysis, which implies that simple covariate adjustment with a Cox model is not designed to adjust results to account for selective crossover. Unlike that analysis and another analysis that simply censors women at the time of treatment crossover, our analyses adjust for the effect of treat- ment crossover and provide a more reliable inference regarding the longer-term outcomes, including survival for extended endocrine therapy with letrozole. But, as pointed out by Korn and Freidlin,6 a major limitation of the statistical approaches used to adjust for treat- ment crossover is their requirement of some unverifiable assump- tions. The three main assumptions mentioned were (1) effect of treatment is the same no matter when the treatment is given, (2)absolute treatment benefit is never greater than the actual treatment times, and (3) all patients receive the same benefit from the treatment.
These assumptions may be reasonably satisfied by the data in this study. For example, Goss et al4 showed that women who started letrozole later still benefited from letrozole treatment. Another study13pointed out another limitation of the SCC approach: it specifically requires an assumption that treatment crossover is independent of prognosis. In our study, although all women randomly assigned toplacebo were offered the option of crossing over, those with comor- Fig 1. Hazard ratios (HRs) and 95% CIs for letrozole versus placebo for all
bidities or other contraindications may not be likely to accept this women randomly assigned in the National Cancer Institute of Canada ClinicalTrials Group MA.17 study. Inverse probability of censoring weighted (IPCW) option; as we have shown, the women who crossed over were younger analyses excluded 34 disease-free survival (DFS), 19 distant disease-free and had better performance status. The consistency of results from the survival (DDFS), and 21 overall survival (OS) events observed after the IPCW and SCC approaches suggests, however, a degree of robustness crossover. COX, Cox proportional hazard model; E, number of events; ITT,intent to treat; SCC, approach proposed by Shao, Chang, and Chow.12.
2011 by American Society of Clinical Oncology Information downloaded from jco.ascopubs.org and provided by at INSERM on May 4, 2013 from 193.54.110.33 Copyright 2012 American Society of Clinical Oncology. All rights reserved.
Longer-Term Outcomes of Letrozole in MA.17
In summary, the analyses presented here provide statistical evi- Declaration and the Disclosures of Potential Conflicts of Interest section in dence related to longer-term efficacy, especially on survival, of letro- zole in the treatment of hormone receptor–positive or hormone Employment or Leadership Position: None Consultant or Advisory
Role: Naiqing Zhao, Bayer Pharmaceuticals (C), Novartis (C); Paul E.
receptor– unknown breast cancer following 5 years of tamoxifen. Fol- Goss, Novartis (C) Stock Ownership: None Honoraria: Paul E. Goss,
lowing the suggestions of Korn and Freidlin,6 results presented here Novartis Research Funding: Naiqing Zhao, Bayer Pharmaceuticals,
should be considered as exploratory, and readers should be aware Novartis Expert Testimony: None Other Remuneration: None
there are strong assumptions behind these analyses.
AUTHOR CONTRIBUTIONS
AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS
OF INTEREST
Conception and design: All authors
Administrative support: Lois E. Shepherd
Although all authors completed the disclosure declaration, the following Provision of study materials or patients: Paul E. Goss
author(s) indicated a financial or other interest that is relevant to the subject Collection and assembly of data: Dongsheng Tu, Lois E. Shepherd,
matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those Data analysis and interpretation: Huan Jin, Dongsheng Tu, Naiqing
relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about Manuscript writing: All authors
ASCO’s conflict of interest policy, please refer to the Author Disclosure Final approval of manuscript: All authors
cancer who complete 5 years of tamoxifen. J Clin of survival in the Breast International Group BIG REFERENCES
1-98 trial by using the inverse probability of cen- 5. Sommer A, Zeger SL: On estimating efficacy
soring weighted method. J Clin Oncol 29:1093- 1. Goss PE, Ingle JN, Martino S, et al: A random-
from clinical trials. Stat Med 10:45-52, 1991 ized trial of letrozole in postmenopausal women 6. Korn EL, Freidlin B: Causal inference for de-
10. Shao J, Chang M, Chow SC: Statistical infer-
after five years of tamoxifen therapy for early-stage finitive clinical end points in a randomized clinical ence for cancer trials with treatment switching. Stat breast cancer. N Engl J Med 349:1793-1802, 2003 trial with intervening nonrandomized treatments.
2. Goss PE, Ingle JN, Martino S, et al: Random-
11. Robins JM, Finkelstein DM: Correcting for
ized trial of letrozole following tamoxifen as ex- 7. BIG 1-98 Collaborative Group, Mouridsen H,
noncompliance and dependent censoring in an AIDS tended adjuvant therapy in receptor-positive breast Giobbie-Hurder A, et al: Letrozole therapy alone or in Clinical Trial with inverse probability of censoring cancer: Updated findings from NCIC CTG MA.17.
sequence with tamoxifen in women with breast weighted (IPCW) log-rank tests. Biometrics 56:779- 3. Ingle JN, Tu D, Pater JL, et al: Intent-to-treat
8. Colleoni M, Giobbie-Hurder A, Regan MM, et
12. D’Agostino RB, Lee ML, Belanger AJ, et al:
analysis of the placebo-controlled trial of letrozole al: Analyses adjusting for selective crossover show Relation of pooled logistic regression to time depen- for extended adjuvant therapy in early breast cancer: improved overall survival with adjuvant letrozole dent Cox regression analysis: The Framingham NCIC CTG MA.17. Ann Oncol 19:877-882, 2008 compared with tamoxifen in the BIG 1-98 study.
4. Goss PE, Ingle JN, Pater JL, et al: Late
13. White IR: Estimating treatment effects in ran-
extended adjuvant treatment with letrozole im- 9. Finkelstein DM, Schoenfeld DA: Correcting
domized trials with treatment switching. Stat Med proves outcome in women with early-stage breast for discretionary treatment crossover in an analysis 2011 by American Society of Clinical Oncology Information downloaded from jco.ascopubs.org and provided by at INSERM on May 4, 2013 from 193.54.110.33 Copyright 2012 American Society of Clinical Oncology. All rights reserved.

Source: http://www.aerio-oncologie.org/wp-content/uploads/2013/06/JCO-2012-Jin-MA17.pdf