Artificial intelligence in cardiology in the current era: a narrative review

Abstract

Background and Objective: Artificial intelligence (AI) has transformed cardiovascular healthcare by influencing multimodal imaging, interventional procedures, and remote patient monitoring systems. The rapid growth of available evidence presents challenges for researchers to evaluate the quality of AI systems and their readiness for clinical application. The review examines current AI uses in cardiology by assessing methodological quality to pinpoint key factors for implementing effective AI models. Methods: We performed extensive search on MEDLINE/PubMed, Web of Science, Scopus, EMBASE for original research, clinical trials, and consensus/guideline statements, published in English from January 2015 through December 2025. Key Content and Findings: AI applied to electrocardiography (AI-ECG) technology shows practical success in clinical trials for detecting left ventricular systolic dysfunction (LVSD) and predicting atrial fibrillation (AF) from normal heart rhythms. AI in echocardiography performs automated view detection, chamber measurement, and assists novice users. Cardiac magnetic resonance (CMR) uses deep learning (DL) for reconstruction and inline perfusion analysis. Coronary computed tomography angiography (CCTA) employs automated coronary artery calcium (CAC)/plaque analysis and machine learning (ML)-based fractional flow reserve computed tomography (FFRct), while nuclear cardiology uses AI to reduce doses and enhance image quality. The cath lab benefits from AI-assisted intravascular imaging, which helps optimize percutaneous coronary intervention (PCI) planning, robotics systems that lower operator radiation exposure, and angiography-derived physiology [e.g., quantitative flow ratio (QFR)], enabling wire-free ischemia assessment with context-dependent value. Challenges to AI adoption include limited external validation data, unclear calibration methods, and gaps in addressing subgroup fairness, decision impact, and post-deployment performance monitoring. Conclusions: AI technology has advanced from experimental development to practical solutions for specific cardiology procedures that improve operational efficiency and standardization. Turning accuracy into patient benefits requires researchers to conduct external validation studies, prospective impact assessments, cost-effectiveness evaluations, equity-focused design, workflow integration, and ongoing governance systems. Implementing AI as an assistive tool will enhance healthcare accessibility, reduce unnecessary treatment disparities, and improve outcomes through clinical support, while automation continues to streamline routine tasks.

Ähnliche Arbeiten