A Novel Method for the Prediction of Pancreatic Fistula Following Pancreaticoduodenectomy by the Assessment of Fatty Infiltration

Pancreaticoduodenectomy is the gold standard treatment of neoplastic lesions of periampullary region. It is a well-known fact that postoperative pancreatic fistula is considered as the most common cause of morbidity. We aimed to present a novel technique using preoperative computed tomography (CT) evaluating the ratio of fatty volume to total pancreatic volume. A total of 40 patients who had been operated for periampullary neoplasm in a single tertiary center were included in the study. Study group was composed of 20 patients who had developed clinically relevant pancreatic fistula. Preoperative CT scans were evaluated retrospectively. Analysis of the image data was based on axial images and two-dimensional (2D) and three-dimensional (3D) post-processing images. Lipid component was estimated as the ratio of lipid volume to the total volume of the pancreas. Adenocarcinoma was the most frequent histopathological diagnosis in both groups (n = 33, 82.5%). Due to preoperative CT scan measurements, both pancreatic fat volume and the ratio of fatty volume to the whole pancreatic volume were found to be statistically higher in the study group (p < 0.01). Hospital stay was longer, and mortality rates were higher in study group. A ratio of 15% fat may be counted as a risk factor for PF development with a 95% sensitivity and 70% specificity, according to the ROC curve analysis. Preoperatively obtained triphasic CT can help in the prediction of postoperative pancreatic fistula. The assessment of the ratio of pancreatic lipid component may be a predictive factor (15%) for anastomotic failure, with the help of an experienced radiologist.