Abstract

Several randomized trials investigating the benefit of adjuvant chemotherapy after surgery in non-small cell lung cancer (NSCLC) have provided conflicting results. With over 7000 patients included, we analyzed results of 13 reports over the past 10 years in which patients received either platinum-containing chemotherapy or not. The major endpoint was to assess the magnitude of the benefit of adjuvant chemotherapy in terms of the absolute benefit. All phase III randomized trials and meta-analyses published as peer-reviewed papers or as abstracts from 1994 to 2007 were eligible. A literature-based meta-analysis was performed; event-based overall- and disease-free survival (OS/DFS) and Relative Risks (RRs) with 95% confidence intervals (CIs) were derived. Magnitudes of benefit were evaluated with: absolute benefit and the number of patients treated for one patient to benefit. Seven sub-populations were examined. Combined effect estimation was computed with fixed- and random-effect models; a heterogeneity test was also applied. Twelve trials plus an individual patient meta-analysis (7334 patients) were gathered; the trials were designed to determine if cisplatin- or carboplatin-based chemotherapy improves survival over surgery. When data were pooled and plotted, significant differences in favor of chemotherapy were seen in OS in all seven sub-population, with a relative benefit of 7-12% and an absolute benefit ranging from 2.5% to 4.1%. A more significant trend for chemotherapy was found in DFS. No significant heterogeneity was observed for all outcomes and sub-populations. The absolute benefit of adjuvant chemotherapy remains essentially the same regardless of how data are screened. While significant differences are clearly found in this analysis, the small magnitude of benefit seen with this large population, especially when considering the number of patients needed for one to benefit, raises important issues when weighing risks and benefits of treatment for individual patients.