Computed tomography-based radiomics and body composition analysis for predicting clinically relevant postoperative pancreatic fistula after pancreaticoduodenectomy.

Background: Preoperative risk assessment of clinically relevant postoperative pancreatic fistula (CR-POPF) is still lacking. This study aimed to develop and validate a combined model based on radiomics, pancreatic duct diameter, and body composition analysis for the prediction of CR-POPF in patients undergoing pancreaticoduodenectomy (PD).

Methods: Multivariable logistic regression was used to construct a combined model in conjunction with radiomics score (Rad-score), pancreatic duct diameter, and visceral fat area/total abdominal muscle area index (VFA/TAMAI). The models were internally validated using 1,000 bootstrap resamples. The predictive performance of these models was assessed using receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA).

Results: The preoperative combined model was validated by 1,000 bootstrap resampling with the area under the ROC curve (AUC) of 0.839 (95% confidence interval: 0.757-0.907). The calibration curves and DCA showed that the combined model outperformed the clinical model and radiomics model. The combined model was presented as a web-based calculator (https://whyyjyljz.shinyapps.io/DynNomapp/).

Conclusions: We explored a method of combining radiomics features, pancreatic duct diameter, and body composition analysis predictors in preoperative assessment for risk of CR-POPF and developed a combined model that showed relatively good performance, but future studies with a larger sample size are needed to verify the stability and generalizability of this model.