Skip to content

Development of a Deep Learning-Based Vision Foundation Model for Cardiac Magnetic Resonance and Efficacy Evaluation in Clinical Auxiliary Diagnosis

Development of a Deep Learning-Based Vision Foundation Model for Cardiac Magnetic Resonance and Efficacy Evaluation in Clinical Auxiliary Diagnosis

Status
Active, not recruiting
Phases
Unknown
Study type
Observational
Source
ChiCTR
Registry ID
ChiCTR2600127917
Enrollment
Unknown
Registered
2026-07-09
Start date
2025-09-01
Completion date
Unknown
Last updated
2026-07-13

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Various cardiac diseases, including myocardial infarction, hypertrophic cardiomyopathy, dilated cardiomyopathy, myocarditis, hypertensive heart disease, and other common conditions.

Interventions

Gold Standard:Professional radiologists manually delineate the endocardium and epicardium based on cardiac magnetic resonance imaging to measure cardiac functional parameters and combine clinical data
Index test:The left ventricular function parameters and prediction results derived from the model are compared and analyzed against manual measurements and diagnostic outcomes. The evaluation metrics

Sponsors

Shanghai Tenth People's Hospital
Lead Sponsor

Eligibility

Sex/Gender
All
Age
18 Years to No maximum

Inclusion criteria

Inclusion criteria: 1.Adult patients aged 18 years or older; 2.Complete standardized cardiac magnetic resonance examination; 3.The examination includes a complete short-axis cine sequence;

Exclusion criteria

Exclusion criteria: 1.Poor image quality or incomplete sequence; 2.Severe Artifact Impact Analysis;

Design outcomes

Primary

MeasureTime frame
The intraclass correlation coefficient (ICC) between the automatically calculated left ventricular ejection fraction (LVEF) by the model and the manually measured results.;

Secondary

MeasureTime frame
The area under the receiver operating characteristic curve (AUC) of a cardiac disease auxiliary diagnostic model.;The average absolute error (MAE) and relative error between the automatically computed cardiac function parameters by the model and the manually measured results.;The Bland–Altman agreement limits between automatically calculated cardiac function parameters by the model and manually measured results.;Model computational efficiency metrics;

Countries

China

Contacts

Public ContactDan Mu

Shanghai Tenth People's Hospital

mudan118@126.com+86 21 6630 1604

Outcome results

None listed

Source: ChiCTR (via WHO ICTRP) · Data processed: Jul 23, 2026