Combined radiomics and body analysis enhance predictive accuracy for pancreatic surgery complications

A new model integrating radiomics, pancreatic duct diameter, and body composition analysis effectively predicts clinically relevant postoperative pancreatic fistula in patients undergoing pancreaticoduodenectomy. This combined approach achieved an area under the ROC curve of 0.839, outperforming existing models. Internal validation through 1,000 bootstrap resamples demonstrated the model’s reliability, culminating in the development of a web-based calculator for clinical use. Future work will focus on expanding sample sizes to further validate this promising predictive tool.