DefinePK

Abstract

Background Minimally invasive cardiac interventions (MICIs), including transcatheter aortic valve replacement (TAVR), percutaneous coronary intervention (PCI), and catheter ablation, have significantly improved the management of cardiovascular diseases by offering reduced morbidity and faster recovery. However, the complex anatomical landscape and the dynamic nature of the heart create challenges in achieving accurate imaging and navigation. Emerging artificial intelligence (AI) technologies, including machine learning and augmented reality (AR), offer potential solutions, yet a comprehensive synthesis of their impact in MICIs remains limited.
Objective This systematic review aims to evaluate the effectiveness of AI-assisted systems in enhancing preoperative planning, intraoperative navigation, and postoperative monitoring in MICIs, with a focus on procedural accuracy, safety, and efficiency.
Methods A systematic review was conducted following PRISMA guidelines. Databases including PubMed, Scopus, EMBASE, and IEEE Xplore were searched from January 2013 to March 2025. Studies were included if they reported clinical applications of AI in imaging or navigation during MICIs. Data extraction and risk of bias assessment were independently performed by two reviewers using standardized tools (RoB 2.0, ROBINS-I). Studies involving animal models, simulations without clinical validation, and non-English articles were excluded.
Results Out of 276 records, 43 studies met the inclusion criteria. AI technologies, particularly convolutional neural networks and AR-based guidance systems, demonstrated consistent improvements in anatomical visualization, procedural success, and navigation precision (mean accuracy within 2 mm). Compared to conventional approaches, AI-assisted procedures showed reduced fluoroscopy time (28% decrease), radiation dose (40% decrease), and overall procedure duration (18.4% reduction). AI was also associated with fewer complications and improved clinical outcomes in TAVR, PCI, and ablation procedures.
Conclusion AI-assisted imaging and navigation significantly enhance the precision, safety, and operational efficiency of MICIs. Despite promising findings, variability in study designs and the need for high-quality, diverse datasets highlight the necessity for further large-scale, multicenter research. Future work should also focus on regulatory frameworks and seamless clinical integration.