Pakistan: Researchers have found in a new study that machine learning models show strong promise in predicting postoperative complications after laparoscopic cholecystectomy and in accounting for intraoperative and perioperative factors that affect patient safety and recovery. However, more research with larger, diverse patient groups is required to validate their clinical reliability.
Laparoscopic cholecystectomy is generally a safe procedure, but postoperative infections, bile duct injuries, bleeding, and perioperative issues can still occur. Early identification of patients at higher risk can significantly improve outcomes. With recent advances in artificial intelligence, researchers have increasingly explored the potential of machine learning algorithms to support clinical decision-making. Until now, however, there had been no comprehensive review summarizing how well these models perform in LC.
The authors conducted the systematic review following PRISMA guidelines and screened studies from four major databases—PubMed, Embase, Scopus, and Web of Science—covering the years 2010 to 2024. Only studies that applied ML techniques to predict postoperative or perioperative complications in LC were included. Six studies met the criteria, and their quality was assessed using the Newcastle-Ottawa Scale. Due to significant variations in study design, patient characteristics, and outcome measures, a quantitative meta-analysis was not feasible; instead, the researchers presented a narrative synthesis.
Key Findings:
Multiple machine learning methods were assessed across the included studies, such as decision trees, artificial neural networks (ANN), deep learning models, and Adaboost algorithms.
ANN models showed the highest predictive accuracy for postoperative quality-of-life outcomes, with MAPE values ranging between 4.20% and 8.60%.
Deep learning approaches performed well for intraoperative evaluations, achieving a balanced accuracy of 71.4% in assessing the critical view of safety (CVS).
Adaboost algorithms proved useful in identifying significant risk factors for hepatic fibrosis in post-cholecystectomy patients.
Despite these advancements, the review noted several limitations. In particular, models aimed at predicting surgical adverse events performed less reliably due to the low incidence of such complications, which reduces the amount of data available for algorithm training. Most existing studies also involved small sample sizes, limiting the generalizability of the findings.
The authors concluded that while ML-based models hold significant potential for enhancing perioperative planning, risk stratification, and postoperative care in laparoscopic cholecystectomy, broader validation is still needed. Future research involving larger, diverse populations and standardized outcome reporting will be critical to determining whether these tools can be integrated into real-world surgical practice.
Reference:
Leghari, S., Tausif, M., Rehan, R. et al. Predictions of postoperative and perioperative complications of laparoscopic cholecystectomy using machine learning algorithms: systematic review. BMC Surg (2025). https://doi.org/10.1186/s12893-025-03035-z
Disclaimer: This website is primarily for healthcare professionals. The content here does not replace medical advice and should not be used as medical, diagnostic, endorsement, treatment, or prescription advice. Medical science evolves rapidly, and we strive to keep our information current. If you find any discrepancies, please contact us at corrections@medicaldialogues.in. Read our Correction Policy here. Nothing here should be used as a substitute for medical advice, diagnosis, or treatment. We do not endorse any healthcare advice that contradicts a physician's guidance. Use of this site is subject to our Terms of Use, Privacy Policy, and Advertisement Policy. For more details, read our Full Disclaimer here.
NOTE: Join us in combating medical misinformation. If you encounter a questionable health, medical, or medical education claim, email us at factcheck@medicaldialogues.in for evaluation.