Comparison of outcomes with pembrolizumab monotherapy (P) versus combination with chemotherapy (P+C) in advanced non-small cell lung cancer (NSCLC).

Abstract

e21579 Background: P and P+C are standard-of-care (SOC) treatment options for advanced NSCLC. However, they have not yet been directly compared in clinical trials. Methods: We conducted a retrospective cohort study of patients with advanced NSCLC who initiated treatment with SOC P±C at our center from 2/11/16 to 10/15/19 (data cutoff 1/15/20). Patient demographic, clinicopathologic, therapeutic and outcomes data were extracted. All radiographic scans were independently evaluated by a thoracic radiologist using iRECIST. Survival time was defined from the start of P±C. Kaplan-Meier and Cox proportional hazards model were utilized. Results: Of 103 patients with median follow up of 17.7 months, 74 (71.8%) had received P, while 29 (28.2%) had received P+C. In PD-L1 tumor proportion score (TPS) unselected population, there were no significant differences in age, sex, smoking status, driver mutation, tumor mutational burden (TMB), line of therapy, ECOG performance status (PS) or immune-related adverse events (irAE) between P and P+C groups. 71.6% in P vs 13.8% in P+C had PD-L1 TPS ≥50% (p < 0.001). There were no significant differences between the two groups in objective response rate (ORR), disease control rate (DCR), unadjusted progression-free survival (PFS) or unadjusted overall survival (OS) (Table). Multivariable adjustment for confounding factors between P+C vs P revealed no differences in OS [hazard ratio (HR) for death, 1.53, 95% CI 0.55 – 4.25] or PFS [HR for progression/death, 1.75, 95% CI 0.63 – 4.91]. Further stratification into PD-L1 TPS ≥50% and < 50% showed no significant differences between P+C vs. P in adjusted OS [HR for death, TPS < 50%- 1.54 (95% CI 0.59 – 4.03); TPS ≥50%- 0.71 (95% CI 0.11 – 4.52)] or PFS [HR for progression/death, TPS < 50%- 1.58 (95% CI 0.72 – 3.48); TPS ≥50%- 0.64 (95% CI 0.06 – 6.93)]. ECOG PS and development of irAE influenced OS in all groups, while TMB was relevant in PD-L1 ≥50% only. Conclusions: Our study shows no significant differences in outcomes with P vs P+C in advanced NSCLC in a real-world setting, albeit with limitations of single-center design, limited sample size, different line settings and lack of disease burden stratification. Ongoing phase III trials comparing front line P vs P+C will definitively address the long-term clinical benefits -if any- of combining cytotoxic chemotherapy with anti-PD-1 drugs. [Table: see text]