EchoNext ECG flag finds severe aortic stenosis, leads to TAVR

AI system

System: EchoNext deep-learning ECG model integrated into the electronic health record
Model: The second-generation EchoNext model.
Release: The second generation of EchoNext.
Application developer: Columbia University (Poterucha, Elias and colleagues)
Role: EchoNext analyzed an ECG and prompted further assessment for structural heart disease.
Configuration: Not assessed
Inputs and outputs: Not assessed
Autonomy: Not assessed
Workflow: EchoNext scoring of a routine ECG. EHR-integrated advisory to the primary team. Cardiology consultation and echocardiography.
Human oversight: Not assessed
Relation to AI: The EHR advisory from EchoNext prompted the echocardiogram in an asymptomatic patient, according to the developers' own case report. The evidence is developer-authored: the case report and the Nature development paper, with disclosed industry research support. No independent account was found.
AI contribution: The EHR advisory from EchoNext prompted the echocardiogram in an asymptomatic patient, according to the developers' own case report. The evidence is developer-authored: the case report and the Nature development paper, with disclosed industry research support. No independent account was found.

Studies: Not assessed
Study design: Not assessed
Measurement: Not assessed
Setting and sample: Not assessed
Measures: echo showed LVEF 40% and severe aortic stenosis
Comparison: cardiologists reading the same ECGs without AI assistance (1,600 reviews of 150 ECGs)
Results: Reference: the accuracy of the cardiologists was 64.0% (95% CI 61.6–66.4%) with sensitivity 61.1% (95% CI 57.3–64.8%) and specificity 66.1% (95% CI 63.0–69.1%). Result: an accuracy of 77.3%, sensitivity 72.6% and specificity of 80.7%.
Additional value from AI: Not assessed
Limitations: Not assessed
Denominators: Not assessed
Materials for replication: Not assessed

Participants: Developer: Columbia University (Poterucha, Elias and colleagues). Operator: Columbia University Irving Medical Center / NewYork-Presbyterian.
Jurisdictions: Benefit: New York, New York, USA. Developer: United States. Operator: United States.
Ownership and control: Not assessed
Dependencies: Not assessed
Cross-border connections: Not assessed
Geopolitical implications: Not assessed

Conditions for replication: Model deployed with alerts in the health record. Clinicians act on the advisory before surgery.
Replication attempts: 1 follow-up entries. Benefit recurred: 1.
Later findings: 23 June 2026. In June 2026 NewYork-Presbyterian announced that EchoNext, commercialized by the spinout Pathway Labs, had received FDA clearance for AI detection of structural heart disease from ECGs, allowing broader deployment beyond the Columbia pilot.
Availability: Not assessed
Institutional decision: Not assessed
Grounds: Not assessed
Action and result: Not assessed
Success criterion: Not assessed

Assessed benefits: People. Organizations. Society.
Depth: People: 3. Organizations: Not assessed. Society: Not assessed.
Reasons: People: Severe aortic stenosis found before non-cardiac surgery, changing perioperative management and leading to TAVR. Organizations: No organizational outcome reported for this case. Society: Single case.
Observed or intended: Not assessed
Rubric: Not assessed
Assessor and date: Not assessed
Alternative assessments: Not assessed
Fill: The makers' claim

Kind: Not assessed
Related entries: Not assessed
Field: Health; cardiology. Hospital patients with an ECG and no known structural heart disease.
Purpose: Screen ECGs for unrecognised structural heart disease and alert clinicians through the electronic health record
Dates: Date basis unassessed: 2026 (precision unassessed). Updated: 21 October 2026.
Stage: Not assessed
Version: SHA-256: ded723cc3c3b…