Development and validation of predictive nomograms for survival in early-onset colon cancer patients with II-III stage across various tumor sites.

BACKGROUND

The incidence of colorectal cancer (CRC) has been escalating, with a concurrent rise in early-onset colon cancer (EOCC). Despite this alarming trend, the prognosis of EOCC has been understudied. Our study aims to identify risk factors associated with EOCC and develop nomograms for predicting overall survival (OS) and cancer-specific survival (CSS), with the goal of choosing suitable therapy for various patient subgroups.

METHODS

Utilizing data from the Surveillance, Epidemiology, and End Results (SEER) database, we conducted a comprehensive analysis to elucidate risk factors in EOCC patients. We developed and validated nomograms to predict OS and CSS, stratifying patients into left-sided and right-sided groups and further categorizing them into distinct risk categories. After propensity score matching, we assessed therapeutic benefits of various interventions across subgroups.

RESULTS

We identified T stage, tumor histology, grade, size, N stage, carcinoembryonic antigen (CEA) levels, perineural invasion, tumor deposits, and race as independent risk factors for the left-sided group through univariate and multivariate Cox regression analyses. Those factors were integrated into the survival nomograms for this group. For the right-sided group, tumor histology, grade, N stage, CEA levels, perineural invasion, tumor deposits, radiation, and chemotherapy were identified as independent prognostic factors and were similarly incorporated into the survival nomograms. The concordance index (C-index) for our nomograms was significantly higher than that of the American Joint Committee on Cancer (AJCC) 7th edition staging system across all cohorts. Receiver operating characteristic (ROC) curve analysis demonstrated area under the curve (AUC) values of 0.72, 0.71, and 0.71 for 1-, 3-, and 5-year OS in the development cohort of the left-sided group, with comparable results in the validation cohort. The right-sided groups exhibited similarly favorable AUC outcomes. Calibration plots indicated a strong correlation between predicted and actual outcomes. Decision curve analysis (DCA) revealed the clinical utility of our nomograms to be superior to the AJCC 7th edition staging system. Analyses for CSS yielded analogous results. Kaplan-Meier curves highlighted significant differences in OS and CSS between low and high-risk groups. Notably, the right-sided groups derived greater benefits from adjuvant chemotherapy compared to the left-sided groups, whereas radiation therapy provided no discernible benefits across all subgroups.

CONCLUSIONS

Our study provides a comprehensive prognostic evaluation of EOCC patients and uses nomograms for predicting OS and CSS in left-sided and right-sided groups. Subgroup analyses underscore the potential advantages of adjuvant chemotherapy in high-risk groups of both cohorts and the low-risk group of the right-sided cohort. These findings may inform the optimization of therapeutic strategies for EOCC patients.