The standard deviation of CT value may become a new way to predict pancreatic fistula: a single-center large sample data study

Abstract

Abstract Background Postoperative pancreatic fistula (POPF) after pancreaticoduodenectomy is one of the most difficult problems for surgeons. The main purpose of this study was to use preoperative imaging and other clinical data to better predict the occurrence of POPF. Methods From January 2017 to January 2022, the baseline data, intraoperative and postoperative outcomes of patients who underwent laparoscopic pancreaticoduodenectomy (LPD) in the Division of Pancreatic Surgery, Department of General Surgery, West China Hospital were retrospectively analyzed. Univariate and multivariate logistic regression analyses were used to analyze the predictive factors of pancreatic fistula after LPD. The receiver operating characteristic curve (ROC) was drawn and the area under the curve (AUC) was calculated to evaluate the predictive value of the predictors. Results A total of 381 patients' data were included in the final study, of which 35 patients (9.2%) had a clinically relevant POPF(CR-POPF). Univariate analysis showed the CR-POPF (+) group had higher BMI (23.69 ± 3.33 vs. 21.96 ± 3.01), larger pancreatic duct diameter (5.62 ± 4.58 mm vs. 4.17 ± 2.39 mm), and larger standard deviation of pancreatic CT value (16.26 ± 34.44 vs. 9.82 ± 2.14), compared with the CR-POPF (-) group ( all P < 0.05). Multivariate analysis showed that only BMI was an independent risk factor for CR-POPF (OR (1.345), P = 0.002)). The AUC for the standard deviation of pancreatic CT value to CR-POPF was 0.706. Conclusions BMI is an independent predictor of CR-POPF after LPD. The standard deviation of pancreatic plain CT may be a new and simple way to predict the risk of CR-POPF after LPD.