Abstract 3131: Novel multi-gene classifier for prediction of relapse-free survival after FU-based adjuvant chemotherapy for stage III colon cancer: A biomarker study of the ACTS-C trial, a randomized controlled study

Abstract

Abstract Background & Purpose: The ACTS-CC trial is a phase III study that have demonstrated the efficacy of S-1 as adjuvant chemotherapy for stage III colon cancer by evaluating its non-inferiority to UFT/LV. As an additional study, gene expression analysis of 5-FU metabolizing enzymes and folate metabolizing enzymes was performed prospectively to identify predictive and/or prognostic biomarkers. We analyzed the correlation between these gene expression levels and relapse free survival (RFS), and then constructed the scoring system to predict the efficacy of each regimen. Methods: Among 1535 patients enrolled in the trial, total RNA was extracted from 798 formalin-fixed, paraffin-embedded specimens. Gene expression levels of 11 enzymes (TS, DPD, TP, OPRT, FPGS, GGH, DHFR, MTHFR, MTHFD, FOLRA, GART) related to 5-FU and folic acid metabolism were analyzed by quantitative real-time reverse transcription - TaqMan PCR. Impact of the mRNA expression level of each gene on RFS was investigated. A multivariate Cox regression with elastic net penalization was employed using treatment-by-gene interaction terms extracted by Tian’s transformation (Tian L, et al.; J Am Stat Assoc. 109, 2014). To evaluate the predictive accuracy of this scoring system method, a double 10-fold cross-validation was performed to stratify all patients based on the cross-validated score. Finally, the prediction algorithm was applied to the entire cohort data to build the scoring system for use in future patients. Results: FPGS and GGH significantly correlated with RFS (HR; 1.289 [95%CI; 1.054-1.577], p= 0.014 and HR; 0.908 [95%CI; 0.832-0.991], p= 0.031) and TP seemed to correlate with RFS (HR; 0.906 [95%CI; 0.816-1.007], p= 0.068). In the S-1 treated group, TP and FPGS significantly impacted on RFS (HR; 0.853 [95%CI; 0.731-0.995], p= 0.043 and HR; 1.591 [95%CI; 1.212-2.089], p< 0.001). In the UFT/LV treated group, there was no significant correlation between the gene expression levels of 5-FU/ folate metabolizing enzymes and RFS. In the cross-validation analysis with the penalized Cox regression scoring method, it was demonstrated that a third part of patients with the lowest cross-validated scores had improved RFS by receiving S-1 treatment. The final scoring system developed using the entire patient cohort was composed of TS, TP, and FPGS. Conclusions: Gene expression level of FPGS, GGH and TP were correlated with the prognosis of stage III colon cancer. The predicting scoring system using the gene expression level of TS, TP and FPGS seemed to be useful for the selection of oral fluoropyrimidine. S-1 might be effective for patients with the low-score colon cancers. Citation Format: Toshiaki Ishikawa, Kenta Murotani, Megumi Ishiguro, Eiji Nakatani, Hiroyuki Uetake, Shigeyuki Matsui, Kenichi Sugihara. Novel multi-gene classifier for prediction of relapse-free survival after FU-based adjuvant chemotherapy for stage III colon cancer: A biomarker study of the ACTS-C trial, a randomized controlled study [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 3131.