Detailed Abstract
[BP Best Oral Presentation - Pancreas Disease/Surgery]
[BP BEST OP 3] Nomogram for predicting postoperative pancreatic fistula (POPF)
In Woong HAN1, Yunghun YOU1,2, Dong Wook CHOI1, Seong Ho CHOI1, Jin Seok HEO1, Sunjong HAN1, Youngju RYU1, Dae Joon PARK1
1Departments of Surgery, Samsung Medical Center Sungkyunkwan University College of Medicine, Korea
2Department of Surgery, Konkuk University Choongju Hospital, Konkuk University School of Medicine, Korea
Introduction : Previous studies analyzed risk factors for postoperativce pancreatic fistula (POPF) and develop risk prediction tool using scoring system. However, there was no study to build nomogram based on individual risk. This study evaluated individual risks of POPF and proposed a nomogram for predicting POPF.
Methods : From 2007 to 2016, the medical records of 1771 patients at Samsung Medical Center undergoing pancreaticoduodenctomy were reviewed retrospectively. Variables with p < 0.05 in the multivariate logistic regression analysis were included in the nomogram. Internal performance validation was executed using a 5-fold cross validation method.
Results : The rate of POPF was 12.5%. On multivariate stepwise logistic regression analysis, body mass index (BMI) (p < 0.001), preoperative level of albumin (p = 0.035), pancreatic duct size (p = 0.002), sex (p = 0.004), ASA (American Society of Anesthesiologists) score (p = 0.039), and location of tumor (p < 0.001) were identified as independent predictors for POPF. A POPF nomogram was based on these 6 variables. The area under the curve (AUC) estimated from the receiver operating characteristic (ROC) graph was 0.7086 in the train set and 0.6523 the test set.
Conclusions : The Samsung Medical Center POPF nomogram was developed to predict the POPF. This nomogram may be useful in selecting patients who need more intensified therapy and establishing effective treatment strategy
Methods : From 2007 to 2016, the medical records of 1771 patients at Samsung Medical Center undergoing pancreaticoduodenctomy were reviewed retrospectively. Variables with p < 0.05 in the multivariate logistic regression analysis were included in the nomogram. Internal performance validation was executed using a 5-fold cross validation method.
Results : The rate of POPF was 12.5%. On multivariate stepwise logistic regression analysis, body mass index (BMI) (p < 0.001), preoperative level of albumin (p = 0.035), pancreatic duct size (p = 0.002), sex (p = 0.004), ASA (American Society of Anesthesiologists) score (p = 0.039), and location of tumor (p < 0.001) were identified as independent predictors for POPF. A POPF nomogram was based on these 6 variables. The area under the curve (AUC) estimated from the receiver operating characteristic (ROC) graph was 0.7086 in the train set and 0.6523 the test set.
Conclusions : The Samsung Medical Center POPF nomogram was developed to predict the POPF. This nomogram may be useful in selecting patients who need more intensified therapy and establishing effective treatment strategy
SESSION
BP Best Oral Presentation
Room B 3/30/2018 1:10 PM - 2:00 PM