TPTh 4.06 Using Machine Learning to Predict the Difficulty Laparoscopic Cholecystectomy
British Journal of Surgery · 05.08.2026 · 113 (Supplement_8)
DOI: 10.1093/bjs/znag087.458 ↗Kurz gefasst
Fachartikel in der Zeitschrift British Journal of Surgery, online erschienen am 05.08.2026. Verfasst von Nadia Matias und 4 weiteren Autor:innen.
Abstract (Originalsprache)
Aim
Acute biliary disease accounts for a third of emergency General Surgery presentations. Early laparoscopic cholecystectomy (LC) is now recommended for symptomatic gallstone disease with the establishment of ‘hot gallbladder’ pathways in acute centres. Predicting the operative difficulty of cholecystectomy is of major benefit for pre-operative planning and the identification of high-risk cases. This study aims to evaluate whether artificial intelligence and machine learning (ML) can accurately predict the level of operative difficulty based on patient demographics, clinical observations, and blood results.
Methods
A retrospective cohort of patients admitted between 2020 and 2022 with acute abdominal pain at a busy university teaching hospital who underwent LC were identified and included in the study. Operation notes were reviewed and scored according to the Surgical Cholecystectomy Score (SCS) and categorised into ‘Easy’ (score<2), and ‘Difficult’ (Grade >=2) cases. ML models were trained using R and a Random Forest Classifier. Inputs included patient demographics, clinical observations and laboratory blood results were extracted, and SCS categories as outputs.
Results
There were 417 patients included in the final cohort (33.6% males), which was split into training (70%) and test (30%) sets. The model achieved an accuracy of 0.67, sensitivity of 0.96, and specificity of 0.11.
Conclusion
ML models show promise for early prediction of difficult LC. However, this study highlights the poor specificity of ML when training only using patient demographics, clinical observations and blood results. Additional data such as imaging data should be used in future work to improve model accuracy.
Zitieren
APA
Matias, N., Martin, G., Brass, A., Ang, Y., & Chan, A. (2026). TPTh 4.06 Using Machine Learning to Predict the Difficulty Laparoscopic Cholecystectomy. British Journal of Surgery, 113(Supplement_8). https://doi.org/10.1093/bjs/znag087.458
Vancouver
Matias N, Martin G, Brass A, Ang Y, Chan A. TPTh 4.06 Using Machine Learning to Predict the Difficulty Laparoscopic Cholecystectomy. British Journal of Surgery. 2026;113(Supplement_8). doi:10.1093/bjs/znag087.458
BibTeX
@article{Matias2026,
author = {Matias, Nadia and Martin, Glen and Brass, Andrew and Ang, Yeng and Chan, Anthony},
title = {TPTh 4.06 Using Machine Learning to Predict the Difficulty Laparoscopic Cholecystectomy},
journal = {British Journal of Surgery},
year = {2026},
volume = {113},
number = {Supplement_8},
doi = {10.1093/bjs/znag087.458}
}
EUVIMED