Diagnostic Accuracy of Artificial Intelligence in Detecting Pleural Effusion on Ultrasound Imaging: A Systematic Review and Meta-Analysis.

Artificial intelligence (AI) interpretation of ultrasound (US) images is promising, yet its accuracy in diagnosing pleural effusions remains unclear. We conducted a comprehensive database search which identified 84 studies, of which 6 met eligibility criteria. We included 2951 patients with 58 392 images and 12 069 ultrasound clips. Bivariate analysis revealed a pooled sensitivity of 0.92 (95% CI: 0.85-0.96, p < 0.05) and specificity of 0.96 (95% CI: 0.88-0.99, p < 0.05), with a high AUC of 0.98 under the sROC curve, although included studies had significant heterogeneity, I 2 > 99%. AI demonstrates high diagnostic accuracy in detecting pleural effusion on ultrasound.