Estimating typical, best-case and worst-case life expectancy scenarios for patients starting chemotherapy for advanced non-small-cell lung cancer: a systematic review of contemporary randomized trials.

INTRODUCTION: We sought to estimate life expectancy scenarios for patients starting chemotherapy for advanced non-small-cell lung cancer (NSCLC).
METHODS: We searched for randomized first-line chemotherapy trials published from January 2000 to April 2008. We recorded median time to progression (TTP) and median overall survival (OS) and extracted the following percentiles (represented scenario) from each OS curve: 90th (worst-case), 75th (lower-typical), 25th (upper-typical) and 10th (best-case). For each OS curve we divided these percentiles (scenarios) in turn by the median to determine if a simple relationship existed between each scenario and the median.
RESULTS: From 60 trials (29,657 patients), the mean for median TTP was 4.8 months (interquartile range [IQR] 4.0-5.3), the mean for median OS was 9.2 months (IQR 8.1-10.1) and the mean ratio for median OS to median TTP was 2.0 (IQR 1.7-2.2). The mean (IQR) in months for each OS scenario was: worst-case, 2.4 (1.9-2.7); lower-typical, 4.8 (4.2-5.4); upper-typical, 16.3 (14.4-18.1); and best-case, 25 (21.0-28.0). The mean values (IQR) for each scenario divided by the median were: worst-case/median 0.26 (0.21-0.29); lower-typical/median 0.53 (0.5-0.57); upper-typical/median 1.81 (1.69-1.93) and best-case/median 2.84 (2.57-3.19). These values can be approximated by the simple multiples: 0.25, 0.5, 2 and 3. Independent predictors of longer OS were ECOG PS<2, adenocarcinoma, and longer TTP; all p-values<0.001.
CONCLUSION: Simple multiples of an OS curve's median provided accurate estimates of typical (half to double the median), best-case (triple the median), and worst-case (one quarter of the median) life expectancy scenarios for patients starting chemotherapy for advanced NSCLC.