Prognostic and predictive value of extended RAS mutation and mismatch repair status in stage III colorectal cancer

The prognostic and predictive value of KRAS gene mutations in stage III colorectal cancer is controversial because many recent clinical trials have not involved a surgery‐alone arm. Additionally, data on the significance of extended RAS (KRAS/NRAS) mutations in stage III cancer are not available. Hence, we undertook a combined analysis of two phase III randomized trials, in which the usefulness of adjuvant chemotherapy with tegafur–uracil (UFT) was evaluated, as compared with surgery alone. We determined the association of extended RAS and mismatch repair (MMR) status with the effectiveness of adjuvant chemotherapy. Mutations in KRAS exons 2, 3, and 4 and NRAS exons 2 and 3 were detected by direct DNA sequencing. Tumor MMR status was determined by immunohistochemistry. Total RAS mutations were detected in 134/304 (44%) patients. In patients with RAS mutations, a significant benefit was associated with adjuvant UFT in relapse‐free survival (RFS) (hazard ratio = 0.49; P = 0.02) and overall survival (hazard ratio = 0.51; P = 0.03). In contrast, among patients without RAS mutations, there was no difference in RFS or overall survival between the adjuvant UFT group and surgery‐alone group. We detected deficient DNA MMR in 23/304 (8%) patients. The MMR status was neither prognostic nor predictive for adjuvant chemotherapy. An interaction analysis showed that there was better RFS among patients treated with UFT with RAS mutations, but not for those without RAS mutations. Extended RAS (KRAS/NRAS) mutations are proposed as predictive indicators with respect to the efficacy of adjuvant UFT chemotherapy in patients with resected stage III colorectal cancer.