Derivation and validation of a preoperative prognostic model for resectable pancreatic ductal adenocarcinoma.

Background: The prognosis of patients with pancreatic ductal adenocarcinoma (PDAC) remains poor even after radical pancreaticoduodenectomy (PD). The study aimed to develop and validate a novel preoperative prognostic model to accurately predict the long-term survival of patients with PDAC.

Methods: Patients with PDAC of pancreatic head from Chinese PLA General Hospital were included. The preoperative PDAC model with contour plots was developed using a non-linear model in the training cohort and then tested in the validation cohort.

Results: Of 421 patients who met the inclusion criteria, 280 were in the training cohort and 141 in the validation cohort. Contour plots for preoperative PDAC model were established to visually predict the survival probabilities of these patients, based on preoperative carbohydrate antigen 19-9, preoperative fibrinogen to albumin ratio and pain symptoms. This model stratified patients into low- and high-risk groups with distinctly different long-term survival in the training cohort [median overall survival (OS) 32.1 vs. 17.5 months; median recurrence-free survival (RFS) 19.3 vs. 10.0 months, both P < 0.001] and the validation cohort (median OS 28.3 vs. 19.0 months; median RFS 17.5 vs. 11.2 months, both P < 0.001). Time-dependent receiver operating characteristic and decision curve analyses revealed that the model provided higher diagnostic accuracy and superior net benefit compared to other staging systems.

Conclusions: This study constructed and validated a novel preoperative prognostic model that can accurately and conveniently predict the long-term survival of patients with resectable PDAC of pancreatic head. Besides, the model can screen high-risk patients with poor prognosis, which may provide references for personal treatment strategies in the future.