A prognostic nomogram for predicting recurrence-free survival of stage I–III colon cancer based on immune-infiltrating Treg-related genes

Abstract

Abstract Purpose A high postoperative recurrence rate seriously impedes colon cancer (CC) patients from achieving long-term survival. Here, we aimed to develop a Treg-related classifier that can help predict recurrence-free survival (RFS) and therapy benefits of stage I–III colon cancer. Methods A Treg-related prognostic classifier was built through a variety of bioinformatic methods, whose performance was assessed by KM survival curves, time-dependent receiver operating characteristic (tROC), and Harrell’s concordance index (C-index). A prognostic nomogram was generated using this classifier and other traditional clinical parameters. Moreover, the predictive values of this classifier for immunotherapy and chemotherapy therapeutic efficacy were tested using multiple immunotherapy sets and R package “pRRophetic". Results A nine Treg-related classifier categorized CC patients into high- and low-risk groups with distinct RFS in the multiple datasets (all p < 0.05). The AUC values of 5-year RFS were 0.712, 0.588, 0.669, and 0.662 in the training, 1st, 2nd, and entire validation sets, respectively. Furthermore, this classifier was identified as an independent predictor of RFS. Finally, a nomogram combining this classifier and three clinical variables was generated, the analysis of tROC, C-index, calibration curves, and the comparative analysis with other signatures confirmed its predictive performance. Moreover, KM analysis exhibited an obvious discrepancy in the subgroups, especially in different TNM stages and with adjuvant chemotherapy. We detected the difference between the two risk subsets of immune cell sub-population and the response to immunotherapy and chemotherapy. Conclusions We built a robust Treg-related classifier and generated a prognostic nomogram that predicts recurrence-free survival in stage I–III colon cancer that can identify high-risk patients for more personalized and effective therapy.