An integrated clinicopathological and genomic model for personalized prognosis in stage II-III colorectal cancer: a real-world study

Background: The prognosis for patients with stage II-III colorectal cancer (CRC) is heterogeneous, with only a subset benefiting from adjuvant chemotherapy. Currently, prognostic models that effectively integrate clinicopathological and genetic factors remain limited. This study aimed to develop a predictive model that accurately forecasts survival and guides personalized treatment decisions. Methods: Data from 322 CRC patients were analyzed. Significant prognostic factors were selected using univariate Cox regression analysis. Subsequently, the least absolute shrinkage and selection operator (LASSO) regression algorithm, coupled with the Cox proportional hazards model, was applied to identify the most parsimonious set of predictors. A nomogram was constructed based on a multivariable Cox regression model to estimate 3- and 5-year overall survival (OS). Predictive performance was assessed using the concordance index (C-index), receiver operating characteristic (ROC) curve analysis, and calibration plots. Decision curve analysis (DCA) was performed to assess the clinical utility of the nomogram. Results: Multivariate analysis identified tumor stage, tumor differentiation grade, lymphovascular invasion, BRAF mutation, and adjuvant chemotherapy as independent predictors of OS. The developed nomogram demonstrated good discrimination, with a C-index of 0.712 and 0.726 for the training and testing cohorts, respectively. Calibration plots showed excellent agreement between predicted and observed 3- and 5-year OS. DCA confirmed that the nomogram provided clinical net benefits. Conclusion: The nomogram, integrating clinicopathological and genetic factors, provides a robust tool for predicting outcomes in patients with stage II-III CRC. It can aid in personalized treatment planning and improve patient management.