A machine learning-based choledocholithiasis prediction tool to improve ERCP decision making: a proof-of-concept study.

BACKGROUND: Previous studies demonstrated limited accuracy of existing guidelines for predicting choledocholithiasis, leading to overutilization of endoscopic retrograde cholangiopancreatography (ERCP). More accurate stratification may improve patient selection for ERCP and allow use of lower-risk modalities. METHODS: A machine learning model was developed using patient information from two published cohort studies that evaluated performance of guidelines in predicting choledocholithiasis. Prediction models were developed using the gradient boosting model (GBM) machine learning method. GBM performance was evaluated using 10-fold cross-validation and area under the receiver operating characteristic curve (AUC). Important predictors of choledocholithiasis were identified based on relative importance in the GBM. RESULTS: 1378 patients (mean age 43.3 years; 61.2% female) were included in the GBM and 59.4% had choledocholithiasis. Eight variables were identified as predictors of choledocholithiasis. The GBM had accuracy of 71.5% (SD 2.5%) (AUC 0.79 [SD 0.06]) and performed better than the 2019 American Society for Gastrointestinal Endoscopy (ASGE) guidelines (accuracy 62.4% [SD 2.6%]; AUC 0.63 [SD 0.03]) and European Society of Gastrointestinal Endoscopy (ESGE) guidelines (accuracy 62.8% [SD 2.6%]; AUC 0.67 [SD 0.02]). The GBM correctly categorized 22% of patients directed to unnecessary ERCP by ASGE guidelines, and appropriately recommended as the next management step 48% of ERCPs incorrectly rejected by ESGE guidelines. CONCLUSIONS: A machine learning-based tool was created, providing real-time, personalized, objective probability of choledocholithiasis and ERCP recommendations. This more accurately directed ERCP use than existing ASGE and ESGE guidelines, and has the potential to reduce morbidity associated with ERCP or missed choledocholithiasis.

Transpapillary Gallbladder Stents Can Stabilize or Improve Decompensated Cirrhosis in Patients Awaiting Liver Transplantation.

GOALS: To describe the short-term and long-term outcomes in 34 consecutive decompensated cirrhotic patients with symptomatic gallbladder disease undergoing transpapillary gallbladder stent (TGS) placement. BACKGROUND: Endoscopic TGS placement is a minimally invasive means of treating symptomatic gallbladder disease in poor surgical candidates. STUDY: Between June 2005 and June 2011, 34 patients with cirrhosis and symptomatic gallbladder disease underwent attempted TGS placement. RESULTS: Median patient age was 52 years, 56% were hospitalized, and 48% were listed for liver transplantation. The median model for end-stage liver disease (MELD) score was 15 (range, 6 to 40) and 88% were Child-Turcotte-Pugh class B/C. A double pigtailed stent was successfully placed in 94% of the patients. At 1-month follow-up, clinical improvement was noted in 82% of the treated subjects and the MELD scores in 14 of 22 (64%) evaluable subjects improved or stabilized. Actuarial transplant-free survival was 53% in the liver transplant candidates with a mean follow-up of 352 days, whereas survival was 44% in the 18 nontransplant candidates with a mean follow-up of 1.5 years. Periprocedural complications included pancreatitis in 5 patients, cholangitis in 3, and 1 patient with cystic duct perforation. In addition, 2 subjects had symptomatic bleeding from traumatic duodenal ulcerations 2 years after TGS placement that necessitated stent removal. CONCLUSIONS: Endoscopic TGS placement was technically feasible in 94% of decompensated cirrhotics and was associated with a relatively low rate of periprocedural (26%) and long-term complications (6%). Stabilization or improvement in clinical status and MELD scores was seen in the majority of treated patients.

Do larger periprocedural fluid volumes reduce the severity of post-endoscopic retrograde cholangiopancreatography pancreatitis?

OBJECTIVE: Fluid therapy is a cornerstone of the early treatment of acute pancreatitis (AP), but data are conflicting on whether it affects disease severity. Administering greater fluid volumes (FVs) during induction of experimental AP preserves pancreatic perfusion and reduces severity but does not prevent onset of AP. We hypothesized that administering larger FV during endoscopic retrograde cholangiopancreatography (ERCP) associates with less severe post-ERCP pancreatitis (PEP). METHODS: In a retrospective cohort study, we identified 6505 patients who underwent 8264 ERCPs between January 1997 and March 2009; 211 of these patients developed PEP (48 mild, 141 moderate, and 22 severe). Data for FVs were available for 173 patients with PEP. RESULTS: In univariable analysis, only 1 of 16 variables was significantly associated with moderate to severe PEP--larger periprocedural FV was protective (0.94 T 0.3 L vs 0.81 T 0.4 L; P = 0.0129). Similarly,multivariable analysis of moderate to severe PEP identified 1 independent predictor-- larger periprocedural FV was protective (odds ratio, 0.20; 95% confidence interval, 0.05-0.83). Conversely, moderate to severe disease correlated with larger FV administered after PEP diagnosis(reflecting treatment decisions). CONCLUSIONS: This hypothesis-generating study suggests that administering larger periprocedural FVs is protective against moderate to severe PEP. Prospective studies on this topic are warranted.