Establishment of risk prediction model of postoperative pancreatic fistula after pancreatoduodenectomy: 2016 edition of definition and grading system of pancreatic fistula: a single center experience with 223 cases.

OBJECTIVE: To establish a risk prediction model for pancreatic fistula according to the pancreatic fistula standards of the 2016 edition. METHODS: Clinical data from 223 patients with PD admitted to Tianjin Third Central Hospital from January 2016 to December 2020 were retrospectively analyzed. Patients were divided into modeling (January 2016 to December 2018) and validation (January 2019 to December 2020) sets according to the time of admission. The risk factors for postoperative pancreatic fistula (POPF) were screened by univariate and multivariate logistic regression analyses, and a risk prediction model for POPF was established in the modeling set. This score was tested in the validation set. RESULTS: Logistic regression analysis showed that the main pancreatic duct index and CT value were independent risk factors according to the 2016 pancreatic fistula grading standard, based on which a risk prediction model for POPF was established. Receiver operating characteristic curve analysis showed that the area under the curve was 0.775 in the modeling set and 0.848 in the validation set. CONCLUSION: The main pancreatic duct index and CT value of the pancreas are closely related to the occurrence of pancreatic fistula after PD, and the established risk prediction model for pancreatic fistula has good prediction accuracy.