International Validation of Echocardiographic Artificial Intelligence Amyloid Detection Algorithm
Authors
Discussion
Loading...
Member takes
Implication
Multisite retrospective study evaluates the AI algorithm's performance in distinguishing cardiac amyloidosis, suggesting high specificity and predictive value.
Key Points
- EchoNet-LVH demonstrated strong discrimination ability with an AUC of 0.896 in diagnosing cardiac amyloidosis.
- At the prespecified model threshold, sensitivity was 0.644 and specificity was 0.988, along with a positive predictive value of 0.968.
- Retrospective case-control study across multiple international sites provided robust evidence without performance heterogeneity.
- AI-assisted echocardiographic evaluations can lead to earlier and more accurate diagnoses of cardiac amyloidosis.
Cite This Study
Duffy et al. (2025) studied this question.