The Application Of Radiomics In Predicting Postoperative Pancreatic Fistula In Patients With Pancreaticoduodenectomy

BACKGROUNDPancreaticoduodenectomy(PD)has become the first choice for the treatment of various benign and malignant diseases that are around the ampulla of the pancreas and the head of the pancreas.PD is an example of a major operation,a complex operation for a general surgeon.Although advances in surgical techniques and perioperative management have made PD as a relatively safe surgical approach with mortality rate less than 5%in large-scale centers,the incidence of surgical complications remains high.Postoperative pancreatic fistula(POPF)is one of the main causes of complications,leading delayed gastric emptying(DGE),postoperative bleeding,intraperitoneal infection,prolonged hospitalization(LOH)and other complications.Several risk factors associated with POPF have been identified,including increased body mass index(BMI),pancreatic texture,pancreatic steatosis,decreased pancreatic duct diameter,and lack of surgical experience.Although many risk factors of POPF have been reported,it is still difficult to integrate various risk factors and significantly predict the accuracy of POPF.A scoring system that can well predict the occurrence of POPF will contribute to the perioperative risk management of PD.Recent studies have shown that several scoring systems had achieved good accuracy in predicting postoperative pancreatic fistula,but they all have some disadvantages.Radiomics is more popular in recent years,which has great application value in cancer diagnosis and treatment and prognosis.Its application in various diseases has been reported in relevant literatures.To our knowledge,there have not previous studies using radiomics to integrate the clinical data into the radiological scoring(rad-score)system for the prediction of postoperative pancreatic fistula in patients undergoing pancreaticoduodenectomy.OBJECTIVETo develop and validate a radiomics-based formula for the preoperative prediction of postoperative pancreatic fistula(POPF)in patients undergoing pancreatoduodenectomy(PD).MATERIALS AND METHODSA total of 117 consecutive patients who underwent PD were enrolled in this retrospective study.Radiomics features were extracted from portal venous-phase computed tomography(CT)of above patients.LASSO logistic regression was used to construct a formula of Rad-score calculation.Then the performance of the formula was assessed with standard pancreatic fistula risk score.RESULTSThe Rad-score could predict POPF with an area under the curve(AUC)of 0.8248 in the training cohort and of 0.7609 in the validation cohort.Patients who had experienced POPF generally had a statistically higher Rad-scores than those who had not experienced POPF in both cohorts.The AUC of the Rad-score was statistically higher than the Fistula Risk Score for predicting POPF in both of the training and validation cohort.CONCLUSIONA novel radiomics-based formula was developed and validated for predicting POPF in patients who underwent PD,which provides a new method for identifying POPF risks and may improve decision support in the prevention of POPF at low cost.