Abstract
Clinical risk stratification for postoperative recurrence in patients with pathological stage II (pStage II) colorectal cancer (CRC) is essential for guiding the use of postoperative adjuvant chemotherapy (ACT). In this study, we identified novel prognostic gene expression biomarkers in patients with pStage II CRC and developed a new risk stratification framework for ACT decision‐making. First, genome‐wide biomarker discovery was conducted to identify prognostic gene expression biomarkers associated with recurrence risk in pStage II CRC. This analysis identified 10 differentially expressed genes as potential biomarkers for recurrence. The efficacy of these biomarkers was then tested using 188 clinical surgical specimens obtained from patients with pStage II CRC. A predictive panel was developed using qRT‐PCR and used to assess 93 clinical specimens with an area under the curve (AUC) of 0.82, and its performance was further validated in an independent cohort (
n
= 95). By incorporating key clinicopathological features, a Gene expression‐based Prediction of Recurrence in pStage II CRC (GPRSC) signature was developed, which robustly predicted postoperative recurrence (AUC: 0.80). Finally, combining the GPRSC signature, microsatellite instability status, and conventional criteria, we developed a novel risk stratification system for postoperative ACT decision‐making in pStage II CRC. Overall, we identified novel gene expression biomarkers and developed a prognostic signature that informs clinical decision‐making regarding postoperative ACT in patients with pStage II CRC.