Development and validation of predictive models for disease-free survival and overall survival in non-metastatic right-sided colon adenocarcinoma patients based on inflammatory and nutritional indices.

BACKGROUND

This study developed and validated prognostic nomograms incorporating inflammatory and nutritional biomarkers to predict disease-free survival (DFS) and overall survival (OS) in patients with non-metastatic right-sided colon adenocarcinoma (NRCA).

METHODS

We retrospectively analyzed NRCA patients who underwent curative resection and allocated 70% ( = 406) to training cohort and 30% ( = 172) to internal validation cohort. An external cohort from a secondary institution ( = 103) provided independent validation. Optimal thresholds for neutrophil-to-lymphocyte ratio (NLR) and prognostic nutritional index (PNI) were established via X-tile software. Prognostic factors were identified through univariate and multivariate Cox regression analyses, followed by nomogram construction. Model performance was evaluated via Harrell’s concordance index (C-index), time-dependent receiver operating characteristic (ROC) curves, calibration plots, and decision curve analysis (DCA). Survival differences across risk-stratified subgroups were assessed via Kaplan‒Meier analysis. A comparative performance evaluation against tumor-node-metastasis (TNM) staging was conducted.

RESULTS

The NLR and PNI optimal cutoffs were determined to be 4.6 and 48.8, respectively. Multivariate analysis revealed that age, CEA level, tumor size, perineural invasion, T and N stage, lymph node dissection (≥ 12), chemotherapy, NLR, and PNI were independent predictors of DFS. For OS, significant predictors included age; CEA level; T and N stage; lymph node dissection (≥ 12); NLR; and PNI. The nomograms demonstrated robust discrimination, with C indices for DFS of 0.805 (training), 0.759 (internal validation), and 0.722 (external validation) and for OS of 0.784, 0.729, and 0.719, respectively. Compared with TNM staging, time-dependent ROC analysis revealed superior prognostic accuracy for both endpoints ( < 0.001). The calibration plots revealed excellent agreement between the predicted and observed outcomes, whereas the DCA confirmed the clinical utility of the model. Risk stratification effectively differentiated survival outcomes between subgroups (log-rank  < 0.001). The finalized nomograms are available as web-based tools: DFS (https://right-sided.shinyapps.io/NRCA/) and OS (https://right-sided.shinyapps.io/NRCAOS/).

CONCLUSION

These validated nomograms incorporating NLR and PNI provide enhanced prognostic precision for NRCA patients, offering clinically actionable tools for personalized risk assessment and treatment optimization.