Landscape of KRASG12C, Associated Genomic Alterations, and Interrelation With Immuno-Oncology Biomarkers in KRAS-Mutated Cancers

Abstract

Purpose: Promising single-agent activity from sotorasib and adagrasib in KRAS^G12C-mutant tumors has provided clinical evidence of effective KRAS signaling inhibition. However, comprehensive analysis of KRAS-variant prevalence, genomic alterations, and the relationship between KRAS and immuno-oncology biomarkers is lacking.

Materials and methods: Retrospective analysis of deidentified records from 79,004 patients with various cancers who underwent next-generation sequencing was performed. Fisher's exact test evaluated the association between cancer subtypes and KRAS variants. Logistic regression assessed KRAS^G12C comutations with other oncogenes and the association between KRAS variants and immuno-oncology biomarkers.

Results: Of the 79,004 samples assessed, 13,758 (17.4%) harbored KRAS mutations, with 1,632 (11.9%) harboring KRAS^G12C and 12,126 (88.1%) harboring other KRAS variants (KRAS^non-G12C). Compared with KRAS^non-G12C across all tumor subtypes, KRAS^G12C was more prevalent in females (56% v 51%, false discovery rate-adjusted P value [FDR-P] = .0006), current or prior smokers (85% v 56%, FDR-P < .0001), and patients age > 60 years (73% v 63%, FDR-P ≤ .0001). The most frequent KRAS variants across all subtypes were G12D (29.5%), G12V (23.0%), G12C (11.9%), G13D (6.5%), and G12R (6.2%). KRAS^G12C was most prevalent in patients with non-small-cell lung cancer (9%), appendiceal (3.9%), colorectal (3.2%), tumor of unknown origin (1.6%), small bowel (1.43%), and pancreatic (1.3%) cancers. Compared with KRAS^non-G12C-mutated, KRAS^G12C-mutated tumors were significantly associated with tumor mutational burden-high status (17.9% v 8.4%, odds ratio [OR] = 2.38; FDR-P < .0001). KRAS^G12C-mutated tumors exhibited a distinct comutation profile from KRAS^non-G12C-mutated tumors, including higher comutations of STK11 (20.59% v 5.95%, OR = 4.10; FDR-P < .01) and KEAP1 (15.38% v 4.61%, OR = 3.76; FDR-P < .01).

Conclusion: This study presents the first large-scale, pan-cancer genomic characterization of KRAS^G12C. The KRAS^G12C mutation was more prevalent in females and older patients and appeared to be associated with smoking status. KRAS^G12C tumors exhibited a distinct comutation profile and were associated with tumor mutational burden-high status.

FIG 1.

Prevalence of KRAS variants by tumor subtype. Most common KRAS mutation variants observed in all KRAS-mutated tumors (n = 13,758) and subtypes. CRC, colorectal cancer; NSCLC, non–small-cell lung cancer; SBA, small bowel adenocarcinoma; TUO, tumor of unknown origin.

FIG 2.

Frequency of KRAS^G12C and KRAS^non-G12C by cancer subtypes. (A) Distribution of 1,632 patients with confirmed KRAS^G12C by tumor subtype. (B) Distribution of 12,126 patients with other confirmed KRAS variants (KRAS^non-G12C) by tumor subtype. Cancer subtype distribution was significantly different between G12C and non-G12C KRAS mutation groups (P < .0001). (C) Frequency of KRAS^G12C mutations in 14 cancer types. CRC, colorectal cancer; NSCLC, non–small-cell lung cancer; SBA, small bowel adenocarcinoma; TUO, tumor of unknown origin.

FIG 3.

Comparison of oncogenic comutations for KRAS^G12C and KRAS^non-G12C cohorts. Comparison of comutations identified in the KRAS^G12C- and KRAS^non-G12C-mutated cohorts. Comutations altered in more than 5% of patients with a confirmed KRAS mutation were included and are shown by subgroups: (A) all cancers, (B) NSCLC, and (C) CRC. Logistic regression was used to calculate the OR and 95% CIs (Data Supplement). FDR-P < .05 was considered statistically significant. Some comutations with KRAS^G12C in all cancers could be caused by coenrichment of the oncogene mutation and KRAS^G12C mutation in NSCLC, as described in the Results section. Careful interpretation of the analysis results using the Data Supplement is recommended. *Significant G12C status (KRAS^G12C and KRAS^non-G12C) effect on the mutation of the oncogene in the subgroup (all cancers or NSCLC or CRC) at FDR-P < .05. **Significant G12C status (KRAS^G12C and KRAS^non-G12C) × cancer subtype (NSCLC or CRC) interaction effect on the mutation of the oncogene at FDR-P < .05. CRC, colorectal cancer; FDR-P, false discovery rate-adjusted P value; NSCLC, non–small-cell lung cancer; OR, odds ratio.

FIG 4.

Evaluation of immune biomarkers by KRAS^G12C, KRAS^non-G12C, and KRAS WT. Comparison of TMB-high (defined as > 10 mut/Mb), high PD-L1 expression, and MSI-high cases across three KRAS cohorts (KRAS^G12C, KRAS^non-G12C, and KRAS-WT). The association between PD-L1 and KRAS^G12C in all cancers could be caused by coenrichment of PD-L1–positive and KRAS^G12C mutation in NSCLC, as described in the Results section: (A) all cancers, (B) NSCLC, and (C) CRC. FDR-P < .05 was considered statistically significant. ***FDR-P < .0001; **FDR-P ≥ .0001 and FDR-P < .01; *FDR-P ≥ .01 and FDR-P < .05. CRC, colorectal cancer; FDR-P, false discovery rate-adjusted P value; MSI, microsatellite instability; mut/Mb, mutations per megabase; NSCLC, non–small-cell lung cancer; PD-L1, programmed death ligand 1; TMB, tumor mutational burden; WT, wild-type.