The first trial testing the efficacy of prolonging to a total duration of 10 years an aromatase inhibitor (AI) as adjuvant treatment of hormone receptor–positive breast cancer has been reported at the 2016 American Society of Clinical Oncology (ASCO) meeting and published in the New England Journal of Medicine.1 In the MA17R study, 1918 postmenopausal patients having previously received 4.5 to 6 years of an AI, preceded in most cases by tamoxifen, were randomized to 5 years of letrozole or placebo. The primary end point was disease-free survival (DFS) defined as the time from randomization to breast cancer recurrence or a new primary breast cancer, whichever came first. Death (either owing or not owing to breast cancer) was not included in the DFS definition, and patients who died without breast cancer recurrence were censored at the date of death. Coherently with study design and eligibility criteria, median age was quite high, 65 years. The high proportion of elderly patients and the low risk of recurrence led to the fact that at the time of analysis there were more deaths (100 in each arm) than breast cancer recurrences or new primaries (67 and 98 with letrozole and placebo, respectively). The primary analysis showed a 4% advantage in 5-year DFS for the letrozole arm (95% vs 91%) and an HR of 0.66 (95% CI, 0.48-0.91), with a significant P value of .01. There was no significant difference in overall survival. No other time-to-event analyses were presented at ASCO, and the message coming from the conference and the press coverage was that postmenopausal patients completing 5 years of aromatase inhibitors should or might be offered further 5 years of treatment, notwithstanding prudential comments of the discussants. However, the article reports a posthoc analysis including death within the DFS definition. Such analysis actually shows that the absolute DFS advantage at 5 years is 2% (from 88% to 90%), the HR is 0.80 (95% CI, 0.63-1.01) or 0.79 (95% CI, 0.63-1.00) if adjusted, and the P value is no longer significant, .06 or .05 in the unadjusted and adjusted analysis, respectively (Table).
Perrone F, Di Maio M, Del Mastro L. A Case Where Switching the End Points for Clinical Trial Interpretation Might Be the Right Choice. JAMA Oncol. 2017;3(6):735–736. doi:10.1001/jamaoncol.2016.4896