Perioperative medicine & outcomes
Artificial Intelligence–Assisted Point-of-Care Ultrasound for Evaluating Left Ventricular Ejection Fraction: A Systematic Review of Prospective Observational Studies
Authors' conclusion
AI-assisted POCUS can approximate reference LVEF and reasonably classify reduced LVEF at the bedside, supporting its use as a screening and triage tool in settings in which comprehensive echocardiography is not immediately available.
Study context
- Population
- Patients in intensive care unit (ICU), ED, perioperative, ward, home or community settings12 studies
- Comparison
- Real-time artificial intelligence (AI)-assisted point-of-care ultrasound (POCUS) vs reference standard for left ventricular ejection fraction (LVEF)
- Outcome
- Agreement between AI-derived and reference LVEF
Methodological detailsnarrative synthesis, QUADAS-2 risk of bias appraisal, PubMed/MEDLINE, Embase and Cochrane searched from inception to June 11, 2025
- Narrative synthesis
- QUADAS-2 risk of bias appraisal
- PubMed/MEDLINE, Embase and Cochrane searched from inception to June 11, 2025
Pooled result
As reported in the abstract
- Correlation (r), range
- 0.56 to 0.92
- Intraclass correlation coefficients, range
- 0.84 to 0.94
- Bland-Altman limits of agreement, broad, often AI underestimating LVEF
- ≈−31.8% to +20.0%
- Sensitivity at 50% threshold, range
- 70% to 93%
- Specificity at 50% threshold, range
- 89% to 100%
- AUC at 50% threshold, range
- 0.85 to 0.98
- Weighted κ, range
- 0.49 to 0.83
Further pooled results
- No study achieved a low risk of bias across all QUADAS-2 domains
Original publication
Artificial Intelligence–Assisted Point-of-Care Ultrasound for Evaluating Left Ventricular Ejection Fraction: A Systematic Review of Prospective Observational Studies. J Cardiothorac Vasc Anesth. Published 12 February 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →