Brief Report: Not Created Equal: Survival Differences by KRAS Mutation Subtype in NSCLC Treated With Immunotherapy

Abstract

Introduction: The predictive and prognostic implications of different KRAS mutation (KRASm) subtypes in metastatic NSCLC have not been clearly defined. We used a nationwide observational database to investigate whether KRASm subtypes differ in their association with survival in metastatic NSCLC treated with immune checkpoint inhibitor (ICI)-based therapy, across programmed death-ligand 1 (PD-L1) levels. Methods: Patients with advanced nonsquamous NSCLC who initiated first-line ICI-based therapy from 2016 to 2021 and had known PD-L1 expression and comprehensive genomic profiling including KRAS, STK11, KEAP1, and TP53 were included. Within PD-L1 expression subgroups (<1%, 1%–49%, ≥50%), Cox multivariable regression was used to evaluate the association between KRASm subtypes (G12C, G12V, G12D, other KRASm) and overall survival, estimated using Kaplan-Meier methodology. Results: Among the 1539 patients, 819 patients were KRAS wild type (KRASwt) and 720 were KRASm (296 KRAS G12C, 143 KRAS G12V, 97 KRAS G12D, 184 other KRASm). In the 50% or higher PD-L1 subgroup, patients with KRAS G12V had worse survival (median overall survival [mOS] = 8.2 mo) compared with KRASwt (mOS = 13.3 mo) and other KRAS subgroups (mOS ranging from 13.4 to 19.9 mo). On adjusted Cox multivariable regression in the 50% or higher PD-L1 subgroup, the hazard ratio for death for KRAS G12V ranged from 1.53 to 1.78 compared with KRASwt and other KRASm subtypes (all p < 0.05). Conclusions: Although patients with 50% or higher PD-L1 with KRAS G12C, G12D, and other subtypes exhibited similar survival to KRASwt, KRAS G12V was associated with significantly worse survival than KRASwt and other KRASm subtypes. All KRASm should not be regarded as uniform predictors of ICI responsiveness, even with high PD-L1 expression; KRAS G12V tumors may have worse outcomes with ICI-based therapy and benefit from treatment intensification.