From Electrocardiography to the Catheterization Laboratory: A Multimodal Artificial Intelligence Framework for Acute...

Area of Science:

• Cardiology
• Artificial Intelligence
• Diagnostic Medicine

Background:

• Current acute coronary syndrome (ACS) diagnosis relies on sequential, independent interpretation of electrocardiograms (ECGs), troponin levels, and coronary angiograms.
• This approach may miss opportunities for integrated, real-time diagnostic insights.

Purpose of the Study:

• To review the evidence on advanced diagnostic modalities, particularly machine learning applications, in acute coronary syndrome (ACS) care.
• To assess the diagnostic accuracy and prognostic associations of AI-enabled ECG, troponin interpretation, and angiography-derived physiological methods.

Main Methods:

• A narrative review selectively searched PubMed, EMBASE, Cochrane CENTRAL, and Web of Science (January 2015-February 2026).
• Study selection was performed by a single reviewer, without duplicate screening or formal risk-of-bias assessment.

Main Results:

• Machine learning-enabled ECG achieved an AUC of 0.938, correctly diagnosing 42% of missed cases by experts.
• An AI-enabled troponin interpretation method (CoDE-ACS) reported an AUC of 0.953, increasing rule-out rates from 27% to 61%.
• Angiography-derived physiological methods (e.g., QFR) yielded conflicting prognostic results across trials, with one showing reduced major adverse cardiovascular events (MACE) and another indicating potential harm.

Conclusions:

• AI holds significant potential for enhancing the accuracy and efficiency of ACS diagnostics, particularly with ECG and troponin analysis.
• While AI demonstrates strong diagnostic capabilities, robust clinical trials are needed to confirm its impact on reducing hard clinical outcomes like death and reinfarction in ACS care.