SABCS 2016. S1-03. First results from the multicenter phase III DATA study comparing 3 versus 6 years of anastrozole after 2-3 years of tamoxifen in postmenopausal women with hormone receptor-positive early breast cancer
Tjan-Heijnen VC, Van Hellemond IE, Peer PG, Swinkels AC, Smorenburg CH, Van der Sangen M, Kroep JR, De Graaf H, Honkoop AH, Erdkamp F, Van den Berkmortel FW, Kitzen JJ, De Boer M, De Roos WK, Linn SC, Imholz AL, Seynaeve C.
Results: Patients were randomized from July 2006 till August 2009. Three years after randomization 1663 patients had no DFS events, with an equal distribution between the treatment arms. The patient and tumor characteristics were well balanced. The median age at randomization was 57 years (P5 = 45 years, P95 = 76 years), the median primary tumor size was 21 mm (P5 = 10 mm, P95 = 50 mm), 67% of the patients had node-positive disease, and in 2% the tumor was HER2-positive (14% unknown); 64% of the patients had received adjuvant chemotherapy and <1% adjuvant trastuzumab. The median adapted follow-up was 4.1 years (P5 = 2.9 years, P95 = 5.8 years). No unexpected safety issues were seen. The 5-year ADFS was 79% in the 3-year and 83% in the 6-year anastrozole treatment group, yielding a HR for ADFS-event of 0.78 (95% CI 0.61 to 1.00; p = 0.0528). In patients with node-positive disease (N = 1117), the HR for ADFS-event was 0.71 (95% CI 0.53 to 0.96; p=0.0232), in N0 disease (N=546) 1.01 (95% CI 0.62 to 1.63; p=0.9817) and in patients with both ER and PR positive breast cancer (N = 1264) 0.68 (95% CI 0.51 to 0.90; p=0.0072). The 5-year adapted overall survival was not different between the treatment groups.
Conclusion: These findings do not yet support the use of extended adjuvant AI prescription after 5 years of sequential endocrine therapy for postmenopausal patients with hormone receptor-positive breast cancer, but suggest benefit for a selected group of patients. Continued follow-up is needed to assess long-term efficacy and safety.
CARE Faculty Perspective: The DATA study is similar in design to the NCIC MA.17 trial of an AI (anastrazole) after a defined duration of tamoxifen (2-3 years). The difference being MA.17 had 4.5 – 6 years of tamoxifen followed by 5 years of letrozole vs placebo. Remembering that after a median of 2.4 years of study treatment the study (MA.17) was un-blinded in favor of letrozole. There has lingered a debate whether a full 5 years of letrozole was sufficiently warranted relative to 2.5 years. The DATA study is also attempting to answer a clinically relevant question for those patients treated with the switch strategy (TAM-AI), whether extending the duration of an AI beyond the 5 years of total hormonal therapy provides additional efficacy. The study was powered (80%) to detect an increase in 3 year aDFS from 90% to 94% (HR 0.60). The study did NOT meet its primary endpoint of 3 year aDFS. Furthermore, there is no difference in OS. However, it can be argued that the follow up is perhaps too short in a luminal breast cancer population. One should be cautious to over-interpret subgroup analyses, for the forest plots are only univariate analyses. At present, it is NOT standard of care to extend an AI beyond the 5 years of total hormonal therapy. However, until a predictive assay is validated, we may consider in individual cases of significant ‘high risk’ disease to have a discussion regarding the pros and cons and the limitations of the DATA study.