The Role of Artificial Intelligence in the Detection of Cardiac Amyloidosis: A Systematic Review.

Abstract

Cardiac amyloidosis (CA) is an underdiagnosed condition that occurs when misfolded amyloid proteins deposit in cardiac tissue causing progressive myocardial dysfunction. Artificial intelligence (AI) technology demonstrates strong potential to enhance the identification of various cardiovascular diseases, including CA. This systematic review examines how AI systems identify CA. This study explores AI's application as a diagnostic tool for CA. We searched five databases (PubMed/MEDLINE, Scopus, Web of Science, Embase, and IEEE Xplore) for relevant studies using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Ten of the 629 studies we discovered were determined to be suitable for the current systematic review. We used the PICOS (P: population, I: intervention or exposure, C: comparison, O: outcome, S: study type) framework to collect relevant data from the included studies. The screening procedure was blinded and used predefined inclusion and exclusion criteria. It was undertaken in two phases. Whenever disagreements emerged, the reviewers deliberated and concluded. After the screening procedure, 10 studies in all were judged suitable for this review. Through the analysis of standard laboratory data, transthoracic echocardiography, electrocardiogram, medical records, cardiac magnetic resonance, and whole-body scintigraphy, these studies assessed the potential usefulness of AI models in diagnosing CA. AI models have been useful as a CA diagnostic tool, outperforming professional cardiologists in one instance or being on par with them in others.