Truace
Subscribe

Diagnostic performance of the ECG-SMART AI score for detecting angiographically significant coronary artery disease in patients undergoing elective coronary angiography

The 12-lead ECG is central to the evaluation of patients with suspected coronary artery disease (CAD) but has limited diagnostic accuracy for identifying significant obstructive disease prior to elective coronary angiography. We previously developed an artificial intelligence-enhanced ECG (AI-ECG) model for detecting acute coronary occlusion. Whether this model generalizes to identifying angiographically significant CAD in symptomatic patients undergoing elective evaluation remains unknown. In this prospective c…

By the Truvace desk1 min read1 sourcePeer-reviewedTRV-2026-1712· cite & verify
Diagnostic performance of the ECG-SMART AI score for detecting angiographically significant coronary artery disease in patients undergoing elective coronary angiography
Hospital Universitari Doctor Peset, València 10 by 19Tarrestnom65. CC BY-SA 4.0 · https://creativecommons.org/licenses/by-sa/4.0

In brief

Researchers tested an AI-enhanced ECG model, originally built to detect acute coronary occlusion, on symptomatic outpatients at a tertiary center who were scheduled for elective coronary angiography. Standard 12-lead ECGs taken beforehand were analyzed offline and patients were sorted into low or intermediate-to-high risk groups.

The result suggests an existing AI tool may carry over to a different cardiac question and help refine pretest probability of obstructive disease. It was a single-center study, so how it performs in other settings or alongside routine care is still unestablished.

Main points

  1. The study enrolled 363 patients with a mean age of 58.9 years, and 55.1% were male.
  2. 36.1% of patients had angiographically significant coronary artery disease.
  3. The AI-ECG labeled 56% of patients low risk and 44% intermediate-to-high risk.
  4. Intermediate-to-high risk had an adjusted odds ratio of 3.12 for significant CAD.
  5. The model reached an AUROC of 0.79 with 60% precision at 80% recall.

The gain

An AI-ECG model first built for acute coronary occlusion identified angiographically significant coronary artery disease in symptomatic outpatients with good discrimination, adding noninvasive information to pretest assessment.

The rundown

The 12-lead ECG is central to evaluating suspected coronary artery disease but has limited accuracy for identifying significant obstructive disease before elective angiography. The team asked whether their earlier acute-occlusion model would generalize to this setting.

The study was a prospective cohort at a tertiary center. Significant CAD was defined as at least 70% stenosis in a major epicardial vessel or at least 50% stenosis of the left main coronary artery.

Authors conclude the findings support generalizability beyond acute coronary occlusion and that AI-ECG may complement existing clinical assessment as a noninvasive input, rather than replace it.

Sources

  1. Peer-reviewedPLOS Digital Health2026-10-08

The debate