Background: While existing data suggest a detriment of immune checkpoint inhibitors (ICI) in other targetable mutations in non-small cell lung cancer (NSCLC), limited retrospective analyses suggest patients with Kirsten rat sarcoma oncogene (KRAS)-mutated NSCLC benefit from ICI in the front-line (1L). To better define this benefit, pooled data from 12 registrational clinical trials investigating 1L ICI with or without chemotherapy (chemo) in patients with documented KRAS status (mutant or wildtype) was evaluated for efficacy of ICI+chemo, ICI alone, and chemo alone. Methods: Pooled data was evaluated for objective response rate (ORR) and overall survival (OS) by KRAS status (mutant, G12C, or wildtype). ORR and 95% confidence intervals (CI) were estimated using Clopper-Pearson method; median OS was estimated using Kaplan-Meier methods. Subgroup analyses were performed using Cox model stratified by KRAS status and PD-L1 status (Positive (combined positive score (CPS) ≥1), Negative (CPS<1), High (CPS≥50), Low (CPS<50)). Results:KRAS mutational status was reported in 1430 patients (61% wildtype, 39% mutated). KRAS G12C was reported in 11% of patients with a KRAS mutation (157/555). Demographics were similar between KRAS mutated, G12C, and wildtype patients. Amongst all patients, 60% were male, 89% white, 60% positive PD-L1, 67% former or current smokers. Table 1 shows outcomes of chemo+ICI, ICI alone, and chemo alone in each population. Conclusions: This retrospective, pooled analysis suggests that patients with KRAS-mutated NSCLC benefit from 1L chemo-ICI similarly to those with KRAS wild-type NSCLC, and should receive combination therapy upfront. Patients with KRAS-mutated NSCLC derived the greatest benefit from the combination of chemo-ICI as compared to ICI or chemo alone. The small number of patients with documented KRAS G12C mutation limits interpretation of the data for this subgroup. Clinical trials investigating targeted therapies for KRAS-mutated NSCLC in the 1L should include a chemo-ICI comparator arm.

NE = Not estimable