Artificial Intelligence-Enabled ECG for Elevated E/e' on Echocardiography: Hemodynamic Relevance and Prognostic...

Area of Science:

• Cardiology
• Artificial Intelligence
• Medical Diagnostics

Background:

• Left ventricular filling pressure is crucial for heart failure management but lacks noninvasive, accessible estimation tools.
• Echocardiography's E/e' ratio is a surrogate for filling pressure, but its serial assessment is challenging.
• An artificial intelligence (AI) model was developed to estimate E/e' using standard 12-lead ECGs.

Purpose of the Study:

• To develop and validate an AI model capable of detecting increased left ventricular filling pressure from ECG data.
• To assess the prognostic significance of the AI-ECG model's output in diverse patient cohorts.
• To evaluate the correlation of AI-ECG-derived filling pressure estimates with invasive measurements.

Main Methods:

• An AI model was trained on over 1 million ECGs and fine-tuned using data from 92,775 patients across two hospitals.
• Model performance was validated in internal (n=9,278) and external (n=17,926) test cohorts.
• Prognostic value was assessed in hospital cohorts and the UK Biobank (n=43,347); invasive validation used cardiac catheterization data (n=60).

Main Results:

• The AI-ECG model achieved an AUC of 0.868 (internal) and 0.850 (external) for detecting increased E/e'.
• AI-ECG output strongly correlated with invasively measured left ventricular end-diastolic pressure (r=0.655).
• The AI-ECG output was significantly associated with incident heart failure and mortality.

Conclusions:

• The AI-ECG model shows promise for identifying patients with elevated left ventricular filling pressure noninvasively.
• This AI tool provides significant prognostic information regarding heart failure and mortality risk.
• Further prospective studies are needed to confirm the clinical utility of the AI-ECG model.