The Foundation And Verification Of Nomogram Of Difficulty Prediction In Laparoscopic Cholecystectomy

Objective: Laparoscopic Cholecystectomy is now a very developed technique.However,while applying to patients with difficult situations during surgeries,surgeons usually turn to conversion to solve the problem.Thus it may prolong the time and add the cost of surgeries.The aim of this study was to collect the clinic data of the patients who received laparoscopic cholecystectomy in order to find a predictive model which can evaluate the difficulty of laparoscopic cholecystectomy before its implementation.Provide guidance for surgeon and improve the efficiency of treatment,and reduce the possibility of the conversion.Method: Retrospectively research 773 patients who underwent laparoscopic or open cholecystectomy during 2016.11-2018.11.The exclusion criterion contained,1.Surgery operated on other organs besides gallbladder.2.Fast frozen pathology was implemented.Patients were divided into two groups by a ratio of 3:1.We evaluated the difficulty of the surgeries subjectively and objectively.Normal surgery was defined as that surgeon encountered no difficulties during the operation.The average time of surgery duration was defined as standard time.Difficult surgery was defined as those time of surgery duration exceeded 1.5 times of standard time or the surgeries which had conversion.580 patients went to training group,193 patients went to validation group.Among training group,128 patients were recorded as difficult surgeries and they were put into difficult surgery group.The other 452 patients were put in normal surgery group.Collect their clinical data such as gender,age,medical history and laboratory test and imaging examination.Use univariate and multivariate logistic regression to analyze and build the prediction model of multivariate logistic regression.Draw the receiver operating characteristic(ROC)curve based on the prediction model among training group and validation group.Calculate the area under the curve(AUC).Plot the nomogram of difficulty prediction in laparoscopic cholecystectomy with the help of prediction model of multivariate logistic regression and R software.Internal and external validation of nomogram were performed in validation group.C-index and calibration curve of both validation were recorded.Result: Elder age,history of abdominal pain within a week,elevation of fibrinogen,increase of length of gallbladder,thickening of gallbladder wall,calculus incarcerated in neck of gallbladder,history of upper abdominal surgery were the independent risk factors contributing to the difficulty of laparoscopic cholecystectomy(P<0.05).According to the prediction model of multivariate logistic regression,the area under the curve(AUC)were 0.808 in training group and 0.809 in validation group.Nomogram of difficulty prediction in laparoscopic cholecystectomy was build,the nomogram contained every independent risk factors and corresponding score,the difficulty of laparoscopic cholecystectomy was evaluated by the total score of the nomogram.The C-index of internal validation was 0.796,and 0.809 in external validation.The calibration curve demonstrated good consistency among internal and external validation.Conclusion: The nomogram of difficulty prediction in laparoscopic cholecystectomy was successfully established and possessed a well ability to predict the difficulty of laparoscopic cholecystectomy.For the patients who have a higher score,a well planned surgery or an experienced surgeon could help them reduce the conversion rate of the operation.