Time trends in overall survival (OS) gain in advanced non-small-cell lung cancer (NSCLC).

Abstract

e19144 Background: There is an ongoing debate about the adequacy of OS as primary endpoint (PEP) in phase III trials in various tumor types, including NSCLC (Soria Ann Oncol 2010). Moreover, median OS and post-progression survival are increasing in NSCLC (Hotta PLoS One 2011 and Hayashi Ann Oncol 2012). We assessed the frequency and time trends in OS gain in recent trials. Methods: We searched PubMed for phase III trials in advanced NSCLC published between January 1998 and December 2011 in 13 leading journals. The PEP was the endpoint stated explicitly, used for N calculation, or listed first. Gain in OS was ascertained when there were significant differences in any pairwise comparison on OS following the original authors’ definitions. Results: We retrieved 148 trials that enrolled a total of 70,716 patients in 325 trial arms. The median number of patients per arm was larger for the 2005-2011 period than for the 1998-2004 period (207 vs 150; P<0.01). OS was used as PEP alone in 99 trials, and as one of the co-primary endpoints in 3 trials (the PEP could not be ascertained for one trial). There was no trend regarding the use of OS as PEP in the two periods compared, either overall (P=0.44) or in first line (P=0.41). Gain in OS according to trial features is shown in the Table. Conclusions: Significant gains in OS appear to decrease along the years, despite the growth in size of NSCLC trials and the continued use of OS as PEP. [Table: see text]