Development and validation of machine learning predictive models for gastric volume based on ultrasonography: A multicentre study
Authors' conclusion
We propose a novel machine learning-based predictive model that outperforms Perlas model by incorporating the key features of age, RLD-CSA, and Perlas grade, enabling accurate prediction of gastric volume.
Study context
- Population
- Adults undergoing gastroscopy under intravenous anaesthesia at two hospitalsCohorts A 415, B 179, C 199
- Participants
- 793 adults
- Exposure
- New models vs Perlas model
- Outcome
- Gastric volume measured by endoscopic aspiration immediately after ultrasonographic examination
Methodological detailsmulticentre, prediction model development and validation, internal and external validation
- Multicentre
- Prediction model development and validation
- Internal and external validation
Main finding
As reported in the abstract
Agreement with measured gastric volume (Bland-Altman mean bias), internal validation
New models
−0.1 to 2.0
mean bias · mL
Perlas model
23.5
mean bias · mL
Observational data: associations, not effects.
Further findings
Prediction of medium-high gastric volume (internal validation)
- New models significantly improved prediction vs Perlas model: AUC 0.74-0.77 vs. 0.63
Prediction of high gastric volume (internal validation)
- New models significantly improved prediction vs Perlas model: AUC 0.85-0.94 vs. 0.74
Prediction of medium-high gastric volume (external validation)
- Adaptive boosting model AUC
- 0.81 (95% CI 0.74-0.89)
- Linear regression model AUC
- 0.80 (95% CI 0.72-0.89)
Prediction of high gastric volume (external validation)
- Adaptive boosting model AUC
- 0.96 (95% CI 0.91-1)
- Linear regression model AUC
- 0.96 (95% CI 0.89-1)
Original publication
Development and validation of machine learning predictive models for gastric volume based on ultrasonography: A multicentre study. J Clin Anesth. Published 18 September 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →