Objective: To develop a predictive model for the incidence of Endoscopic Retrograde Cholangiopancreatography (ERCP) following emergency laparoscopic cholecystectomy, utilising advanced machine learning techniques. Also, the associated factors between these procedures are to be identified. Background: Laparoscopic cholecystectomy is the preferred treatment for symptomatic cholelithiasis and acute cholecystitis, with increasing applications even in severe cases. The necessity for postoperative ERCP to manage choledocholithiasis or biliary injuries poses significant clinical challenges. This study aims to develop a predictive model for the incidence of ERCP following emergency laparoscopic cholecystectomy using advanced machine learning techniques. Method: We conducted a retrospective cohort study utilising the Tokushukai Medical Database, which includes data from 42 hospitals over a decade in Japan. The study population consisted of adult patients undergoing emergency laparoscopic cholecystectomy. We employed four machine learning models—logistic regression, random forest, gradient-boosting decision trees (GBDT), and multilayer perceptrons - on a dataset divided into training/validation and testing groups. We also calculated Shapley additive explanation values for the GBDT to identify the significant variables. Result: Out of 9,695 patients, 8,854 met the inclusion criteria. The incidence of postoperative ERCP was 5.7% and 6.4% in the training/validation and testing datasets, respectively. The GBDT demonstrated superior performance, with the highest predictive capacity for postoperative ERCP. Significant predictors identified included common bile duct dilatation, serum albumin, and lactate dehydrogenase levels. Conclusion: This study successfully established a robust predictive model for ERCP following emergency laparoscopic cholecystectomy and identified associated factors with the outcome.

Background: Real-time surgical computer vision (CV-AI) is technically feasible, although real-world actionability and failure modes remain poorly characterised. This study quantified live performance and mapped dominant failure mechanisms during intraoperative deployment in laparoscopic cholecystectomy (LC). Methods: A real-time multi-task CV-AI was implemented prospectively (surgeon-blinded) during 100 consecutive LCs. Live video inference evaluated (1) five-phase workflow recognition, (2) inflammatory grade (1-4), (3) Rouviere’s sulcus (RS) identification, and (4) drain detection performance. Outputs were compared with intraoperative observer annotations. Discordant cases underwent structured observer/surgeon feedback integrated with quantitative error signatures to derive a task-by-mechanism failure taxonomy. Results: Phase recognition showed greatest potential actionability. Across 316 eligible transitions, median absolute timing error was 22s (IQR 9-91); only 55% met the predefined ±30 s actionability threshold. Large delays clustered around branched workflows, particularly IOC/CBDE (median error 588s vs 23s). Inflammatory grading showed modest ordinal agreement (κw 0.39), improving when restricted to a post-liver-lift assessment window (κw 0.47). RS outputs were dominated by early false positives (sens 100%, spec 0%). Drain detection performance supported automated documentation (sens 90.9%, spec 83.1%). Mixed-methods synthesis identified unstable/limited visual access, non-linear workflow sequences, instrument-driven proxy triggering, and documentation-reference discordance as predominant failure mechanisms. Conclusion: Real-time CV-AI can generate clinically-relevant intraoperative signals, but utility is constrained by timeliness and predictable context-dependent failure modes rather than headline accuracy alone. Mapping failure mechanisms to quantitative signatures identifies actionable targets for iterative model redesign, including branch- and uncertainty-aware workflow modelling, before progression to higher-stakes decision-support.

Purpose Accurate pre-operative diagnosis and prognostication are critical in suspected pancreatic adenocarcinoma (PDAC) and cholangiocarcinoma (CCA), where surgery carries substantial morbidity and mortality yet may not confer survival benefit for all patients. Artificial intelligence (AI) has shown potential to improve cyto- and histopathology performance in this area. This scoping review aimed to map the current evidence on AI approaches applied to digitised pre-operative PDAC/CCA tissue and to summarise reported diagnostic and prognostic performance. Methodology This study was conducted in accordance with PRISMA extension for scoping reviews. A systematic search was conducted in five databases. Articles published between 2015 and 2025 assessing AI-models on digitised pre-operative tissues of suspected PDAC or CCA were extracted. Results Of the 601 articles screened, 12 met inclusion criteria. Eight studies analyzed fine needle aspirates (FNA), two bile duct brushings (BDB) and one fine needle biopsy (FNB). Nine studies were single centre, and three multicentre. Convolutional neural networks were the predominant model architecture. Common technical strategies included augmentation, segmentation and Z-stacking. Model performance varied across studies and reporting was heterogeneous. Two studies directly compared AI performance with human interpretation. Fang et al., (2025) concluded that AI diagnostic accuracy (90.0%) exceeded intermediate (88.3%) and junior (76.7%) cytopathologists but was lower than senior cytopathologists (95.0%). Marya et al., (2024) evaluated a computer-aided detection approach in which AI highlighted regions of interest for cytopathologists; diagnostic accuracy was similar to the original cytology interpretation, while workflow efficiency was reported to improve. Conclusion Early studies suggest AI has potential to play a valuable role in pre-operative tissue analysis for suspected PDAC and CCA in clinical practice.