MedPath

AI-ECG Screening for Left Ventricular Systolic Dysfunction

Not yet recruiting
Conditions
Left Ventricular Systolic Dysfunction
Registration Number
NCT06231797
Lead Sponsor
Seoul National University Hospital
Brief Summary

The purpose of the current study is to verify the effectiveness of the artificial intelligence algorithm applied to the electrocardiogram as a potential screening tool for left ventricular systolic dysfunction.

Detailed Description

The current investigators have developed an artificial intelligence (AI) algorithm based on 12-lead electrocardiogram (ECG) detecting left ventricular systolic dysfunction, through 364,845 ECGs from 148,547 patients. Then, when the model was tested retrospectively on 59,805 ECGs of 24,376 patients, the model performance expressed as an area under the receiver operating characteristic curve was 0.889 (95% CI 0.887-0.891).

The investigators are planning to prospectively validate the model's effectiveness as a potential screening tool for left ventricular systolic dysfunction.

Recruitment & Eligibility

Status
NOT_YET_RECRUITING
Sex
All
Target Recruitment
1530
Inclusion Criteria
  • Individuals or those whose legal representative agree to participate in the study, and sign the consent form
  • Can complete both 12-lead electrocardiogram and transthoracic echocardiography
Exclusion Criteria
  • Individuals whose age is less than 18 year-old.
  • Individuals who do not agree to participate in the study
  • Patients who are unable to participate in clinical trials at the discretion of the investigator

Study & Design

Study Type
OBSERVATIONAL
Study Design
Not specified
Primary Outcome Measures
NameTimeMethod
Area under the receiver operating characteristic curve (AUROC)Through study completion, an average of 1 year

AI model performance detecting LVSD, expressed as an AUROC. As a diagnostic assistance for LVSD, an ROC curve expressed as sensitivity to (1-specificity) will be presented, and the accuracy of prediction will be confirmed by calculating the AUROC, which is the area below.

Secondary Outcome Measures
NameTimeMethod
SensitivityThrough study completion, an average of 1 year

AI model sensitivity detecting LVSD

Positive predictive valueThrough study completion, an average of 1 year

Positive predictive value in the recruited patient population

Negative predictive valueThrough study completion, an average of 1 year

Negative predictive value in the recruited patient population

SpecificityThrough study completion, an average of 1 year

AI model sensitivity detecting patients with normal left ventricular systolic function

© Copyright 2025. All Rights Reserved by MedPath