It is possible to predicte pancreatic fistula with preoperative imaging or other perioperative parameters after pancretic surgery?

Abstract

Indroduction: Postoperative pancreatic fistula (POPF) poses an unsolved problem in pancreatic surgery. Prediction of the risk of POPF can optimize individual treatment decisions for patients undergoing pancreas surgery. The aim of this study is to investigate perioperative variables to predict risk factors of developing a clinically relevant POPF after pancreatic surgery. Methods: We reviewed 324 consecutive patients undergoing pancreatic surgery between July 2007 and March 2016 in our hospital. We investigated perioperative parameters. Particularly the pancreatic anastomosis region was segmented manually using OsiriX in axial CT slices. Pixel values from the segmentation regions were exported as a single 3D VOI region. Histograms of the 3D VOIs were obtained and histogram statistics calculated using Prism. The histogram skewness values were used for calculations. Results: Postoperative pancreatic fistula occurred in 52/324 patients (31,7%). Parameters like age, gender, body mass index, the American Society of Anesthesiologists (ASA) score, medical comorbidities, preoperative biliary drainage, type of anastomosis, and pancreatic duct size were analyzed.In our study main pancreatic duct (MPD) diameter on computed tomography (CT) scan <1,95 mm, high BMI, duration of the operation, proof of Enterococcus faecium in drain fluid were statistically significant predictors for clinically relevant POPF. Analyzing the skewness values of contrast-enhanced CT scan we found no significance for developing pancreatic fistula. Conclusions: Perioperative risk factors like main pancreatic duct (MPD) diameter, BMI, duration of the operation, proof of Enterococcus faecium in drain fluid predicts POPF after pancreatic surgery.