Key result
Machine learning models detect carotid plaque on ultrasound with high accuracy and a 0.98 AUROC.
Why the study?
Early ultrasound detection of carotid plaques is limited by manual analysis, and the comparative diagnostic performance of automated machine learning models remains underexplored.
Do machine learning models provide accurate automated detection of carotid artery plaques on ultrasound compared to reference standards?
Meta-Analysis
Do machine learning models provide accurate automated detection of carotid artery plaques on ultrasound compared to reference standards?
Odds Ratio: 302 (95% CI 54–1684)
Machine learning models demonstrate high diagnostic accuracy for detecting carotid artery plaques on ultrasound, though significant heterogeneity highlights the need for standardized validation before clinical integration.
ML models show high diagnostic accuracy for ultrasound carotid plaque detection; leaves open need for prospective external validation before practice change.
BACKGROUND: Carotid artery plaques, a hallmark of atherosclerosis, are key risk indicators for ischemic stroke, a major global health burden with 101 million cases and 6.65 million deaths in 2019. Early ultrasound detection is vital but hindered by manual analysis limitations. Machine learning (ML) offers a promising solution for automated plaque detection, yet its comparative performance is underexplored. This systematic review and meta-analysis evaluates ML models for carotid plaque detection using ultrasound. METHODS: We searched PubMed, Scopus, Embase, Web of Science, and ProQuest for studies on ML-based carotid plaque detection with ultrasound, following PRISMA guidelines. Eligible studies reported diagnostic metrics and used a reference standard. Data on study characteristics, ML models, and performance were extracted, with risk of bias assessed via PROBAST+AI. Pooled sensitivity, specificity, AUROC were calculated using STATA 18 with MIDAS and METADTA modules. RESULTS: Of ten studies, eight were meta-analyzed (200-19,751 patients) Best models showed a pooled sensitivity 0.94 (95% CI: 0.88-0.97), specificity 0.95 (95% CI: 0.86-0.98), AUROC 0.98 (95% CI: 0.97-0.99), and DOR 302 (95% CI: 54-1684), with high heterogeneity (I² = 90%) and no publication bias. CONCLUSION: ML models show promise in carotid plaque detection, supporting potential clinical integration for stroke prevention, though high heterogeneity and potential bias highlight the need for standardized validation.
No takes yet. Share an insight, caveat, or question.
Eini et al. (2025) conducted a meta-analysis in Carotid artery plaques. Machine learning models vs. Reference standard was evaluated on Diagnostic performance for carotid plaque detection (DOR 302, 95% CI 54-1684). Machine learning models for ultrasound-based carotid plaque detection demonstrated high diagnostic accuracy, with a pooled AUROC of 0.98 and a diagnostic odds ratio of 302 (95% CI 54-1684).
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: