Integrating prognostic nutritional index with CA19-9 improves risk stratification in pancreatic cancer

• The Preoperative Prognostic Nutritional Index is a robust and independent predictor of postoperative complications and overall survival in patients with PDAC. • The PNI consistently outperformed systemic inflammatory indices (NLR, PLR, CRP) and CEA in terms of prognostic accuracy. • The integration of PNI with CA19-9 significantly improved model discrimination, calibration, and clinical applicability. • The combined model provides a comprehensive framework that reflects both nutritional-immune reserves and tumor burden. • These findings clarify the inconsistent evidence in the literature and highlight the need for prospective multicenter validation before clinical implementation. Pancreatic ductal adenocarcinoma (PDAC) is one of the most aggressive malignancies, with a 5-year survival rate rarely exceeding 10% despite surgery. Reliable preoperative biomarkers are required to refine risk stratification. The prognostic nutritional index (PNI), derived from albumin and lymphocyte counts, reflects both nutritional and immune status. Although nutritional and inflammatory indices have been studied, their prognostic performance in PDAC remains unclear. We retrospectively analyzed 112 patients who underwent curative-intent pancreaticoduodenectomy for PDAC (2010 and 2020). Preoperative PNI, neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), C-reactive protein (CRP), and tumor markers (CA19-9, CEA) were evaluated. The outcomes were postoperative complications and overall survival (OS). Prognostic accuracy was assessed using ROC curves, spline modeling, Cox regression, concordance index (C-index), calibration, reclassification (net reclassification improvement, NRI; integrated discrimination improvement, IDI), and decision curve analysis (DCA). Low PNI was associated with major complications (p < 0.001) and shorter OS (median 29.8 vs. 52.9 months, p = 0.012) than high PNI. ROC analysis identified an optimal cutoff of 40.2 (AUC 0.65), whereas spline modeling confirmed a near-linear association between declining PNI and mortality. In the multivariate analysis, low PNI (HR 1.70), pancreaticobiliary localization (HR 3.06), male sex (HR 2.24), and lymph node metastasis (HR 1.20 per node) independently predicted OS. The PNI outperformed the NLR, PLR, CRP, and CEA. The integrated PNI+CA19-9 model achieved superior prognostic performance (C-index 0.73 vs. 0.68 for PNI and 0.63 for CA19-9), with improved calibration, greater net clinical benefit in DCA, and supportive reclassification metrics (NRI 0.21, IDI 0.07). Preoperative PNI is a robust independent predictor of morbidity and survival in patients with PDAC. Integration with CA19-9 further enhances prognostic accuracy, supporting individualized risk stratification for patients with PDAC. Prospective multicenter studies are needed to validate the clinical implementation of this model.