Laparoscopic pancreaticoduodenectomy : CUSUM analysis in a developing single surgeon

Abstract

Introduction: Postoperative pancreatic fistula (POPF) remains the most serious complication after pancreaticoduodenectomy (PD) and the risk of developing POPF is more or less equal with different techniques of reconstruction. Various risk calculation models have been developed for predicting / calculating risk of pancreatic fistula. Fistula risk score (FRS), described in 2013 is the most commonly used scoring system for this purpose. Recently alternative fistula score (a-FRS), was developed and validated in western population. Both FRS & a-FRS are developed and validated in western population with predominantly pancreatic head tumors; whereas, in India we mostly encounter periampullary tumors. Aim: to compare the a-FRS and FRS models for predicting POPF. Method: We performed a retrospective analysis from 1 st January 2011 to 31 st March 2018. All the patients undergoing PD with pancreatico-enteric reconstruction by PJ method were included and patients undergoing distal /median/ total pancreatectomy were excluded. A-FRS and FRS scores were calculated and compared with the actual POPF rates (defined by 2016 ISGPS definition). Result: study cohort contained 825 patients with male preponderance. Median age of the study cohort was 55 years. 61.8% tumors were located in periampullary region and 77.4% were adenocarcinoma. Overall CR-POPF rate in our study population was 16.8% and was represented with more accuracy by a-FRS model (p value-0.003). The performance of each model was assessed byarea under the receiver operating curve (AUC) and calibration plots. Conclusion: a-FRS model performed better in as compared to FRS model (AUC-0.70 Vs 0.65, p value-0.003) for prediction of POPF but, AUC's of both the models were still not the best for the perfect prediction model.