A novel model for predicting prognosis in pancreatic cancer patients: a retrospective study.

[BACKGROUND] Pancreatic cancer is notoriously associated with a poor prognosis and limited survival. We aim to develop a simple and accessible model that can accurately predict the prognosis of pancreatic cancer patients.

[METHODS] This study retrospectively analyzed the blood indicators and overall survival of 500 pancreatic cancer patients. The median value was used as the cutoff for univariate and multivariate analyses. To address the limitations of the median value, receiver operating characteristic analysis was performed, and the optimal cutoff value (the highest Youden index) was determined, followed by univariate and multivariate analyses. Prognostic LASSO coefficient screening was performed to establish a pancreatic cancer prognostic prediction model. Risk factor diagram, Kaplan-Meier curve and prognostic calibration curve were plotted to validate the efficacy of the model.

[RESULTS] Multivariate regression analysis showed that neutrophils (hazard ratio (HR) = 1.416, 95% confidence interval (CI) = 1.037-1.932, = 0.028), lymphocytes (HR = 0.625, 95% CI = 0.462-0.846, = 0.002), Carcinoembryonic Antigen (CEA) (HR = 1.820, 95% CI = 1.315-2.518, < 0.001), CA125 (HR = 1.392, 95% CI = 1.001-1.936, = 0.049), TNM stage (I vs. III: HR = 3.052, 95% CI = 1.900-4.905, < 0.001; I vs. IV: HR = 4.815, 95% CI = 2.504-9.258, < 0.001) and Neutrophil-to-Lymphocyte Ratio (NLR) (HR = 1.748, 95% CI = 1.210-2.525, = 0.003), Lymphocyte-to-Monocyte Ratio (LMR) (HR = 0.597, 95% CI = 0.430-0.829, = 0.002), Neutrophil-to-Macrophage Ratio (NMR) (HR = 2.065, 95% CI = 1.331-3.206, = 0.001), and Systemic Immune-Inflammation Index (SII) (HR = 1.751, 95% CI = 1.244-2.466, = 0.001) were independent risk factors for OS. We have developed a new model incorporating gender, age, treatment, TNM stage, pathological grade, CEA, CA125, and NLR. The model demonstrates good predictive performance, with a C-index of 0.73.