作者: Xiaoli Ma , Gwénaël Le Teuff , Benjamin Lacas , Ming Sound Tsao , Stephen Graziano
DOI: 10.1016/J.JTHO.2016.02.002
关键词: Randomized controlled trial 、 Bioinformatics 、 Confidence interval 、 Hazard ratio 、 Oncology 、 Medicine 、 Clinical trial 、 Internal medicine 、 Adjuvant therapy 、 Cancer 、 Lung cancer 、 Proportional hazards model
摘要: Abstract Introduction Tumor protein p53 gene ( TP53 ) mutations are common in stage I through III non–small cell lung cancer, but clinical trials have shown inconsistent results regarding their relationship to the effects of adjuvant therapy. The objective is clarify putative prognostic and predictive effects. Methods A pooled analysis (exons 5–8) was conducted four randomized (the International Adjuvant Lung Cancer Trial, J BRonchus 10, Leukemia Group B-9633, Navelbine Trialist Association trial) platinum-based chemotherapy (ACT) versus observation (OBS). Hazard ratios (HRs) 95% confidence intervals (CIs) mutant versus wild-type (WT) for overall survival (OS) disease-free (DFS) were estimated using a multivariable Cox model stratified on trial adjusted sex, age, clinicopathological variables. Predictive value evaluated with an interaction between treatment . Results total 1209 patients (median follow-up 5.5 years) included. There 573 deaths (47%) 653 DFS events (54%). Mutations (434 [36%]) had no effect (OBS HR OS = 0.99, CI: 0.77–1.28, p = 0.95; 0.78–1.25, 0.92) marginally benefit from ACT (test interaction: OS, 0.06; DFS, 0.11). Patients WT tendency toward better outcomes than did those OBS group (HR 0.77, 0.62–0.95, 0.02; 0.75, 0.62–0.92, 0.005). In arm, deleterious observed 1.40, 1.10–1.78, 0.006; 1.31, 1.04–1.64, 0.02). Conclusions mutation ACT. who received ACT, tended be associated shorter wild-type