Background: Colorectal cancer (CRC) is a leading cause of cancer-related mortality, and the occurrence of lymph node metastasis (LNM) and postoperative gastrointestinal dysfunction (POGD) significantly dictates long-term survival and quality of life. Accurate preoperative risk assessment is crucial for optimizing surgical planning and perioperative management. This study aimed to evaluate the predictive value of preoperative inflammation-nutrition indicators for LNM and POGD in CRC, and to construct nomogram prediction models.

Methods: A retrospective analysis was conducted on the data from 200 CRC patients who underwent radical surgery. Multiple preoperative inflammation-nutrition indicators were collected. Logistic regression was used to screen for independent predictors of LNM and POGD, and nomogram models were constructed accordingly. Multicollinearity was assessed using variance inflation factors (VIFs), and the linearity-in-the-logit assumption for continuous predictors was evaluated using the Box-Tidwell test. Discrimination was evaluated using receiver operating characteristic (ROC) curves and the area under the curve (AUC). Model calibration, internal validation via Bootstrap resampling (1000 iterations), and clinical net benefit were assessed using calibration curves and decision curve analysis (DCA).

Results: The incidence rates of LNM and POGD were 39.0% and 32.0%, respectively. The independent predictors for LNM were neutrophil percentage-to-albumin ratio (NPAR) (per 1-unit increase; OR = 1.24, 95% CI 1.08–1.42), cT3–4 stage (OR = 3.20, 95% CI 1.62–6.33), history of smoking (OR = 3.11, 95% CI 1.50–6.46), and fibrinogen-to-albumin ratio (FAR) (per 1-unit increase; OR = 1.22, 95% CI 1.03–1.45). The independent predictors for POGD were NPAR (per 1-unit increase; OR = 1.21, 95% CI 1.08–1.35), open surgery (OR = 2.60, 95% CI 1.14–5.93), advanced age (per 1-year increase; OR = 1.09, 95% CI 1.04–1.13), and prolonged operative time (per 1-min increase; OR = 1.04, 95% CI 1.02–1.06). The AUC of the LNM predictive model was 0.80 (95% CI 0.74–0.87), and the adjusted AUC after internal validation using Bootstrap was 0.78 (95% CI 0.69–0.87); The AUC of the POGD prediction model was 0.80 (95% CI 0.73–0.87), and the adjusted AUC was 0.77 (95% CI 0.66–0.87). DCA indicated that it has a certain clinical net benefit.

Conclusion: The NPAR-based nomogram model shows promising predictive performance, and preoperative NPAR may serve as a key indicator for predicting LNM and POGD in CRC patients.