Simple radiological parameters predict postoperative pancreatic fistula in pancreatoduodenectomy

Background:

Postoperative pancreatic fistula (POPF) remains the achilles heel of pancreatic surgery. Several risk stratification strategies have been described, but rarely these are preoperatively available. We attempted to predict the development of POPF based on preoperative radiological parameters from cross-sectional imaging.

Methods:

A cohort of n = 254 patients with pancreatoduodenectomy from 2013 to 2018 and available preoperative cross-sectional imaging were included. The following parameters were analyzed: thickness of pancreatic parenchyma at the pancreatic neck, calcifications, diameter of the pancreatic duct, lymphadenopathy, cholestasis, irregularities of the pancreatic duct. A multivariable model was constructed to predict the occurrence of clinically relevant POPF according to ISGPS.

Results:

A multivariable model containing four of the radiological parameters was able to stratify patients into three risk categories, low/medium/high, with 5%, 10% and 40% of clinically relevant POPF. Area under the curve (AUC) of this model was 0.79.

Conclusion:

Preoperative risk stratification based on cross-sectional radiological parameters is possible in pancreatoduodenectomy.