Prediction of clinically relevant postoperative pancreatic fistula using radiomic features and preoperative data

Abstract

Clinically relevant postoperative pancreatic fistula (CR-POPF) can significantly affect the treatment course and outcome in pancreatic cancer patients. Preoperative prediction of CR-POPF can aid surgical decision-making process and lead to better perioperative management of patients. In this retrospective study, we have developed radiomic risk models that use preoperative computed tomography (CT)-based radiomic features in a cohort of 148 patients having undergone pancreaticoduodenectomy. The developed models were compared with risk models using preoperative clinical features and mesh-based volumes of the annotated structures. The model signatures were analysed in detail by visualising feature expression maps and comparing significant features to the standard CR-POPF risk measures. The best average area under receiver operating characteristic curve (AUC) of 0.81 on validation data was achieved by the radiomic signature model retrained using the most important features. The following features showed significant correlation with outcome (p <0.05)-texture and morphology of healthy pancreatic segment, intensity volume histogram-based feature of pancreatic duct segment and morphology of combined segment; mesh-based volumes of healthy pancreatic segment, segment containing arteries, combined segment and pancreatic duct segment; BMI and sex. The results of the study indicate that preoperative data can be useful for an accurate risk assessment of CR-POPF.

Artificial intelligence in surgery

“…The prediction of complications after surgery has been the goal of many academic studies 139,140 and presents a formidable challenge in a complex postoperative setting, with myriad variables affecting care and outcomes. However, the early detection of complications 138,141 -in particular, devastating outcomes such as anastomotic leaks after rectal cancer surgery and postoperative pancreatic fistulas after pancreatic surgery 142,143 -is likely to have a substantial impact on the ability of healthcare systems to reduce mortality following complications 144,145 . MySurgeryRisk represents one of the few advances in complication prediction, using a machine learning algorithm 50 .…”