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.

Randomized trial of cholangioscopy-guided laser lithotripsy versus conventional therapy for large bile duct stones (with videos).

BACKGROUND AND AIMS: Bile duct stones >1 cm have a decreased incidence of successful endoscopic extraction and often require lithotripsy. Although previous guidelines suggested mechanical lithotripsy for large common bile duct stones, current guidelines suggest cholangioscopy-guided lithotripsy as an adjunct with or without balloon dilation or mechanical lithotripsy. However, no randomized trials have assessed the usefulness of this practice. METHODS: Patients with bile duct stones >1 cm in diameter were randomized in a 2:1 ratio to cholangioscopy-guided laser lithotripsy versus conventional therapy only. Conventional therapies such as mechanical lithotripsy or balloon dilation were also allowed in the laser lithotripsy group. Randomization was stratified by history of ERCP in the past 3 months. The primary outcome was endoscopic clearance of the bile duct stones. RESULTS: Endoscopic clearance was achieved in 39 (93%) of 42 patients treated with cholangioscopy-guided laser lithotripsy and 12 (67%) of 18 treated with conventional therapy only (P = .009). The 9 patients in whom ERCP was unsuccessful underwent surgical common duct exploration with stone removal. Mean procedure time was 120.7 ± 40.2 minutes for the cholangioscopy-guided laser lithotripsy group compared with 81.2 ± 49.3 minutes for the conventional therapy group (P = .0008). There was no significant difference in fluoroscopy time, number of procedures, or adverse events (cholangitis) (cholangioscopy, 2; conventional, 1) and post-ERCP pancreatitis (cholangioscopy, 2; conventional, 1). CONCLUSION: Cholangioscopy-guided laser lithotripsy increases the incidence of endoscopic clearance of large bile duct stones and decreases the need for surgery compared with conventional therapy alone. However, it is associated with longer procedure times. (Clinical trial registration number: NCT0175997.).

Comparative performance of PET tracers in biochemical recurrence of prostate cancer: a critical analysis of literature.

Positron emission tomography (PET) with a number of tracers targeted to particular biological features of cancer has been explored for the imaging evaluation of patients with biochemical recurrence of prostate cancer after curative primary treatment. However, these reports are often heterogeneous in study design, patient cohorts, standards of reference for the imaging findings, data analysis, and data reporting. The aim of our study was to address these limitations by extracting and re-analyzing the PET detection data only from studies that satisfied pre-defined sets of patient selection criteria and verification standards. Our investigation analyzed the effects of 5 tracers ((18)F-fluorodeoxyglucose (FDG), (11)C-acetate (ACET), (11)C- or (18)F-choline (CHOL), anti-1-amino-3-(18)F-fluorocyclobutane-1-carboxylic acid (FACBC), and radiolabeled ligand targeted to prostate-specific membrane antigen (PSMA)), 2 treatment types (radical prostatectomy and radiation therapy), and whether the detected disease was local or metastatic, including lesion type (bone, lymph node, soft tissue). FDG exhibited the lowest detection rate for any suspected disease. ACET tended to be advantageous over CHOL in detecting local recurrence and lymph node lesions, even though the difference was not statistically significant. FACBC had greater likelihood of detecting local recurrence, when compared to CHOL, though this difference was not statistically significant. PSMA tended to show a higher proportion of patients with suspected disease compared to the other four tracers. Patients treated with radiation therapy had greater odds of displaying local recurrence on PET than those treated with radical prostatectomy. We also provide suggestions for future investigations that facilitate communication and the impact of the findings.