Validation of a Risk Scoring System for Postoperative Pancreatic Fistula

Abstract

AIM: The aim of this study is to analyze the effects of pancreatic texture and pancreatic duct diameter on postoperative pancreatic fistula (POPF) formation and to define a risk scoring system for identifying high-risk patients for pancreatic surgery. METHODS: The data of 100 consecutively operated patients who underwent pancreatic surgery between May 2017 and June 2018 from seven different centers were collected through a web-based data collection module that is accessible at the website of The Turkish Hepato Pancreatico Biliary Surgery Association and analyzed retrospectively. The patients' data were evaluated according to the International Study Group for Pancreatic Fistula (ISGPF) criteria. The risk scoring system is defined according to the texture of the pancreatic tissue and the diameter of the pancreatic duct. Risk coefficients were distributed as 3, 2, and 1 for soft, intermediate and firm pancreatic tissue, respectively. The risk coefficients for the pancreatic duct diameter were distributed as 3, 2, and 1 for the pancreatic duct diameter as smaller or equal to 3 mm, wider than 3 mm and smaller or equal to 5 mm, and wider than 5 mm, respectively. The total risk score was calculated by multiplying the distributed risk coefficients. RESULTS: Thirteen out of 100 patients were excluded from the study due to missing or incomplete data. 17 of 87 (19.5%) patients had POPF. Six of 17 patients (35%) were in concordance with ISGPF-Grade A. Nine of 17 (53%) patients were in concordance with ISGPF-Grade B, and 2 of 17 (12%) patients were in concordance with ISGPF-Grade C. 22 of 87 (25.4%) patients were in the low-risk group, 18 (20.6%) patients were in the intermediate-risk group and 47 (54%) patients were in the high-risk group. There were no clinically relevant POPF in the low-risk group, while one patient developed an ISGPF-Grade A fistula. Two of 18 patients in the intermediate-risk group had POPF and both were clinically relevant. Fourteen of 47 patients in the high-risk group had POPF and 5 of 14 patients were clinically non-relevant, while 9 of 14 patients had clinically relevant POPF. The sensitivity and the specificity of the scoring system were 82.35% and 52.86% (p = 0.012), respectively. The area under the curve was 0.666 (95% CI: 0.54-0.77). CONCLUSIONS: This study proposes and validates a simple intraoperative risk scoring system, based on pancreatic tissue and duct diameter, for predicting POPF. The scoring system demonstrated a high negative predictive value, allowing clinicians to identify patients less likely to develop POPF. This practical tool may assist in surgical decision-making and in tailoring postoperative management in pancreatic surgery.