Association of KRAS variant subtypes with survival and recurrence in patients with surgically treated colorectal liver metastases.

e16245 Background: KRAS mutations are common in colorectal cancer and associated with poor outcomes after colorectal liver metastases (CRLM) resection. However, the relationship between specific KRAS subtypes and survival/recurrence in CRLM patients post-resection is not well-defined. Methods: We retrospectively analyzed 1389 CRLM patients undergoing curative-intent hepatectomy (2020–2022). KRAS variants were determined by targeted sequencing with Sanger confirmation. OS and PFS were analyzed by Kaplan–Meier/log-rank tests, and prognostic factors were evaluated using Cox regression. Results: KRAS mutations were detected in 28.0% (389/1389) of patients. KRAS mutation prevalence was higher in females (32.8% vs 25.8%; P = 0.008), CEA-positive (29.8% vs 24.2%; P = 0.030), CA19-9-positive (31.0% vs 25.6%; P = 0.030), and patients with multiple liver metastases (31.4% vs 23.9%; P = 0.002; Table 1). KRAS mutations were more frequent in right-sided colon primaries (34.9%) compared to left hemicolon (24.5%). Among KRAS-mutant patients (n = 389), codon 12 variants accounted for 62.7%, predominantly G12D (32.1%) and G12V (12.9%) (Supplementary Table 1). Univariate Cox analysis showed that KRAS mutations were associated with shorter PFS (HR = 1.48; P < 0.001) and OS (HR = 2.18; P < 0.001), with the largest hazard ratios for G12V (PFS: HR = 2.06; OS: HR = 3.61; both P < 0.001; Table 2). KRAS-mutant tumors had shorter mPFS and mOS compared to WT tumors (mOS: 43 months vs not reached; mPFS: 15.87 vs 23.93 months; both P < 0.001) (Figure 1A). Codon 12 mutations, particularly G12V, conferred the worst survival outcomes (Figure 1B and 1C). Baseline characteristics were similar between codon 12 and non-codon 12 KRAS-mutant groups, except codon 12 variants were more common in males (66.0% vs 55.3%; P = 0.049; Table 3). In the codon 12–mutant subgroup (n = 244), G12V tumors were more frequent in CEA-positive than CEA-negative patients (25.4% vs 9.3%; P = 0.003) and in those with multiple liver metastases (25.2% vs 13.9%; P = 0.036; Table 4). Conclusions: KRAS mutations were frequent and independently predicted worse OS and PFS after curative hepatectomy. Prognostic impact varied across KRAS subtypes, with G12V indicating a high-risk subgroup. Incorporating codon-specific KRAS information into risk stratification and perioperative decision-making may refine patient selection and trial design.