PURPOSE To compare and characterize overall survival (OS) differences between clinical trial data from the KEYNOTE-010 trial and real-world data (RWD) from the Flatiron Health database in patients with programmed death ligand 1 (PD-L1)–expressing advanced non–small-cell lung cancer (NSCLC) who received second-line pembrolizumab monotherapy. METHODS Clinical trial data were from the randomized phase II/III KEYNOTE-010 trial that enrolled patients from August 28, 2013, to February 27, 2015. At data cutoff for KEYNOTE-010, the median survival follow-up time for pembrolizumab patients was 11.2 months. RWD were from Flatiron Health advanced NSCLC database and included patients who initiated second-line pembrolizumab from January 26, 2015, to February 28, 2019. At data cutoff for Flatiron, the median survival follow-up time for pembrolizumab-treated patients was 6.1 months. Clinical trial data from KEYNOTE-010 and RWD from Flatiron were analyzed without adjustment, with propensity adjustment, and filtered per the main KEYNOTE-010 eligibility criteria (EC) of histologically/cytologically confirmed PD-L1–positive NSCLC, Eastern Cooperative Oncology Group performance status of 0/1, no prior therapy with docetaxel for NSCLC, and laboratory values indicative of adequate organ function in addition to prior line of therapy requirements. RESULTS Among 243 patients from KEYNOTE-010 and 782 from Flatiron, median age was 63 v 68 years, and 64% v 54% were male, respectively. OS data from KEYNOTE-010 and Flatiron were similar without any adjustment (n = 782; hazard ratio [HR], 0.96; 95% CI, 0.80 to 1.15) and after both filtering and propensity adjustment (n = 221; HR, 0.99; 95% CI, 0.73 to 1.34). CONCLUSION Without any adjustment, as well as after applying similar EC and appropriate statistical methods, RWD demonstrated similar effectiveness for pembrolizumab in second-line NSCLC as observed in randomized clinical trials.
Prognostic model of survival outcomes in non-small cell lung cancer patients initiated on afatinib: pooled analysis of clinical trial data
