Clinical and genomic predictors of response and toxicity to sotorasib in a real-world cohort of patients with advanced KRAS G12C-mutant non-small cell lung cancer.

Abstract

9083 Background: With the accelerated approval of the KRAS G12C inhibitor sotorasib for patients with advanced KRAS G12C-mutant non-small cell lung cancer (NSCLC), there is a new need to explore efficacy and identify clinical and genomic factors associated with activity and toxicity among patients treated in routine clinical practice. Methods: Patients(pts) treated with sotorasib outside of the clinical trial setting at three institutions (MSK, Columbia, NYU) were examined to identify factors associated with real-world PFS (rwPFS), overall survival (OS), real-world (rw) response, and clinically significant toxicity (CTCAE v5.0 grade 3 or higher treatment-related adverse events [G3+ TRAE]). TP53, STK11, and KEAP1 status were assessed by targeted NGS. Response and rwPFS were determined by investigator assessment of radiology reports. rwPFS and OS were estimated by Kaplan-Meier methods; log-rank tests were used for univariate group comparisons and Mantel-Haenszel method was used to determine hazard ratios (HRs). Fisher’s exact test was used to determine associations between categorical variables. Results: 105 pts with advanced KRAS G12C-mutant NSCLC treated with sotorasib were identified. The response rate for patients treated was 28%, with median rwPFS of 5.3 months, and median OS of 12.6 months. KEAP1 co-mutations were associated with shorter rwPFS and OS (rwPFS HR 3.19, P = 0.004; OS HR 4.10, P = 0.003); no significant differences in rwPFS or OS were observed among pts with or without TP53 co-mutations (rwPFS HR 1.10, P = 0.731; OS HR 1.19, P = 0.631) or STK11 co-mutations (rwPFS HR 1.66, P = 0.098; OS HR 1.73, P = 0.168). Dual STK11/ KEAP1 co-mutation status was associated with shorter OS compared to STK11 wild-type (WT)/ KEAP1 mutant (MUT) status (OS HR 4.04, P = 0.033) and STK11 MUT/ KEAP1 WT status (OS HR 5.41, P = 0.012). 16/105 (15%) pts experienced G3+ TRAEs. 15/16 (94%) patients with G3+ TRAE had previously been treated with immune checkpoint inhibitor (ICI). Among pts with prior ICI exposure (n = 86), last exposure within 12 weeks of sotorasib initiation was associated with G3+ sotorasib TRAEs ( P < 0.001) and TRAE-related sotorasib discontinuation ( P = 0.014); there was no association between prior irAE during ICI therapy and G3+ sotorasib TRAEs ( P = 0.757). In total, 15/53 (28%) pts with last ICI exposure within 12 weeks of sotorasib initiation experienced G3+ TRAEs, most commonly hepatotoxicity. No cases of G3+ TRAEs were observed in 33 pts with last ICI exposure > 12 weeks prior to sotorasib treatment. Conclusions: Among patients treated with sotorasib in routine clinical practice, KEAP1 co-mutations were associated with poor outcomes. Severe toxicity was almost exclusively associated with recent ICI exposure. These observations may help further guide sequencing of sotorasib and may help inform the next generation of KRAS G12C clinical trials.