The study focused on the development of a new way of predicting the aggressiveness of a pancreatic tumor before surgery using a special dual-energy CT scan and computerized image analysis. The research team created a model that combined information from CT images with patients' clinical data, specifically body mass index and the level of the CA 125 marker. This combined model achieved high accuracy in predicting the degree of tumor aggressiveness, with AUC values of 0.836 in the test group and 0.862 in the externally validated group. The results were verified using decision curve analysis and calibration graph, which confirmed the reliability and consistency of the model. A nomogram created from this model could help doctors better plan the treatment of patients with pancreatic ductal adenocarcinoma before surgery. The study included patients from two centers, which increases the credibility of the results.