Machine learning algorithms as early diagnostic tools for prolonged operative time in patients with fluorescent laparoscopic cholecystectomy: a retrospective cohort study
BackgroundThe purpose of this study was to explore the risk factors for prolonging the operative time of fluorescence laparoscopic cholecystectomy (LC). In addition, we aimed to construct predictive models to identify patients with potentially prolonged operative times (OT) using machine learning (M...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2025-06-01
|
| Series: | Frontiers in Surgery |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fsurg.2025.1582425/full |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | BackgroundThe purpose of this study was to explore the risk factors for prolonging the operative time of fluorescence laparoscopic cholecystectomy (LC). In addition, we aimed to construct predictive models to identify patients with potentially prolonged operative times (OT) using machine learning (Ml) methods.MethodsClinical data of patients who underwent fluorescent LC for gallbladder stones in the Department of Hepatobiliary Surgery at our hospital from April 2023 to July 2024 were retrospectively analyzed, with the 75th percentile of operative time as the cut-off point. Parameters screened by univariate and multifactor analysis and LASSO regression were incorporated into the model, and the optimal model was analyzed and determined by integrating 11 Ml classification models.ResultsThe 85 min or more was defined as prolonged OT, and 29% (223/726) of patients had prolonged OT. The variables screened by univariate, multivariate analysis and lasso regression included type of cholecystitis, number of puncture ports, gallbladder adhesion, conservative antibiotic treatment before surgery, gallbladder thickness (mm). The above five parameters were incorporated into the Ml model. Comprehensive analysis revealed that the Light Gradient Boosting Machine (LightGBM) classification model was the optimal model, with the area under the curve (AUC) of the validation cohort was 0.876, the 95% confidence interval was 0.8139–0.938, the accuracy was 0.843, the sensitivity was 0.805, and the specificity was 0.857, with AUC of validation cohort was 0.876. The calibration curves showed good agreement between the actual and predicted probabilities of the LightGBM classification model; The decision curve analysis showed that the model had good net clinical benefit in most of the threshold probability range.ConclusionsWe created a nomogram for assessing the risk of prolonged fluorescent LC time using the LightGBM classification model, which may help surgeon identify patients whose OT may be prolonged. |
|---|---|
| ISSN: | 2296-875X |