Résumé
Abstract only 1094 Background: With the necessity of reducing randomized clinical trial (RCT) duration, cost and number of patients, surrogate endpoints of overall survival (OS) are increasingly being used in cancer RCTs. However, most of these endpoints currently lack of standardized definition enabling a comparison of RCT results. Some recommendations have been proposed for specific cancer sites but they do not rely on a formal consensus methodology. The objective of the Definition for the Assessment of Time-to-event Endpoints in CANcer trials (DATECAN) project is to provide guidelines to standardize definitions of time-to-event endpoints in RCTs for different cancer sites. Here, we present results for BREAST cancer. Methods: We relied on the modified Delphi consensus method, a validated formalized consensus process for the development of practice guidelines. International experts with various backgrounds and expertises were involved. First, the coordinating committee, a group of statisticians and epidemiologists involved in the design and conduct of RCTs, led a comprehensive literature review to identify time-to-event endpoints, events of interest and the existence of guidelines in adjuvant and metastatic settings. The steering committee, which included additional medical experts, validated the list and prepared the questionnaire sent for rating to an independent expert committee. Results: The consensus process involved 2 rounds of scoring (31 experts) and 1 in-person meeting (in parallel to ASCO'12). Each expert had to rate on a 1-9 scale if s/he agreed or not for including events (e.g. death from breast cancer) in the definition of time-to-event endpoints (e.g. progression-free survival). 150 events had to be scored for the 11 selected endpoints. Consensus was reached for 57% of the events after the 2 rounds of scoring. After the in-person meeting, consensus was reached for all the remaining events except one. Conclusions: The DATECAN guidelines should help standardizing definitions of commonly used endpoints. This process should (i) facilitate the comparison of RCTs and (ii) improve the quality of future RCTs by providing better estimation of sample size and treatment effect.