Résumé
The aim of the study was to evaluate the reporting of time-to event endpoints (TTEE) in operable non-small-cell lung cancer randomized clinical trials.
Eligible trials were randomized trials with pre-operative or perioperative or adjuvant chemotherapy for operable NSCLC. Articles were extracted from two Cochrane meta-analyses.
Thirty-four studies were included in the review. Among the 34 articles, a total of 62 TTEE were recorded. Overall survival (OS) was the most frequent TTEE used (32 terms, 51.6%). Other TTEE used were 16 disease-free survival (25%), 5 progression-free survival (8%), 3 time to recurrence (4.8%), 1 time to disease progression (1.6%), 1 recurrence free survival (1.6%), 1 event free survival (1.6%), 1 disease specific survival (1.6%), 1 disease free interval (1.6%), 1 cancer free survival (1.6%). In the Methods section, using the four key points to define TTEE we observed that the “starting point”, “events”, “information on censoring”, “assessment of events” were clearly defined for 43 (69.4%), 34 (54.8%), 6 (9.7%), 33 (53.2%) endpoints respectively. In the results section, using the five key points, we observed that the “Kaplan-Meier estimation”, “estimation of effect size”, “precision (confidence interval)”, “number of events”, “number of patients at risk”, “multivariate analysis” were clearly identified for 46 (74.2%), 31 (50%), 30 (48.4%), 37 (59.7%), 28 (45.2%), and 17 (27.4%) endpoints, respectively.
A majority of articles failed to provide a complete reporting of TTEE. Guidelines for TTEE reporting in operable NSCLC randomized clinical trials is warranted.