1 in 5
Heart attacks are silent — damage occurs but the patient is unaware.
CDC, Heart Disease Facts
Clinical evidence
The science behind earlier detection.
AI-ECG algorithms surfaced through MyoVista Insights are validated through prospective and retrospective studies at academic medical centers.
23+
Peer-reviewed publications
9+
Academic medical centers
10K+
Patients in published studies
15+
AI-ECG algorithms
The diagnostic gap
What conventional ECG misses.
Standard ECG interpretation focuses on rhythm and gross morphology. The signal contains far more — patterns AI can extract reliably.
1 every 36s
One person dies from cardiovascular disease every 36 seconds.
CFAH, 2024
80%
Of heart-disease cases could be prevented with earlier detection.
WHO
Methodology
How studies are designed and reported.
Each algorithm referenced here was evaluated against an imaging-derived reference standard, in published, peer-reviewed work — with patient counts and AUC reported transparently.
- 01
Reference-standard comparisons
AI-ECG outputs are evaluated against echocardiography or cardiac MRI ground truth — not against other ECG-derived models.
- 02
Multi-center validation
Models are trained at one institution and validated at independent centers (Mount Sinai, Mayo, Rutgers, UK Biobank).
- 03
Prospective study designs
Where feasible, retrospective discovery studies are followed by prospective validation (e.g. MSH validation cohort).
- 04
Reporting transparency
AUC, sensitivity, specificity, and patient counts are reported per cohort with confidence intervals.
Selected studies
A reading list.
Six papers we point clinicians to first.
- 2024Read
Quantitative Prediction of Right Ventricular Size and Function From the ECG
Journal of the American Heart Association · UKBB n=42,938 + MSH n=3,019 + MSH validation n=115 — AUC 0.86
- 2024
Energy-Waveform ECG to Detect Subclinical LV Dysfunction in T2DM
Cardiovascular Diabetology · Training n=178 + validation n=97 — AUC 0.81
- 2023Read
A Foundational Vision Transformer Improves Diagnostic Performance for ECG (HeartBEiT)
npj Digital Medicine · Multi-task SOTA across 5 ECG diagnostic benchmarks
- 2023
Novel ECG-Based Deep Learning Algorithm to Predict Cardiomyopathy in Patients With PVCs
Journal of the American Heart Association · Mount Sinai cohort — predictive of cardiomyopathy onset
- 2022
Multi-Center Retrospective Cohort: Deep Learning of ECG to Identify Left-Heart Valvular Dysfunction
JACC · Multi-center retrospective cohort
- 2021
Machine Learning of ECG Waveforms to Improve Selection for Asymptomatic LV Dysfunction Testing (PROMPT)
JACC · PROMPT prospective study
Research partners
Built with leading academic medical centers.
Mount Sinai
Algorithm development
UWE Bristol
Population studies
Westcliffe
Heart screening
Collaborate
Interested in clinical collaboration?
We partner with academic medical centers and health systems to advance AI-ECG research.