Artificial intelligence, extended reality and computational modelling in cross-sectional cardiovascular imaging in congenital heart disease: a narrative review

Background and Objective: Artificial intelligence (AI), extended reality (XR), and computational modelling are increasingly integrated with cross-sectional cardiovascular imaging, particularly computed tomography (CT) and cardiovascular magnetic resonance (CMR), to address the diagnostic and therapeutic complexity of congenital heart disease (CHD). Given the lifelong and heterogeneous nature of CHD, these technologies have the potential to enhance anatomical assessment, haemodynamic understanding, procedural planning, and personalised care. This narrative review aims to provide a structured overview of current developments, clinical applications, and translational challenges related to AI, XR, and computational modelling in CT- and CMR-based CHD imaging. Methods: A targeted literature search was performed in PubMed to identify representative and clinically relevant publications from January 2014 to October 2025, including early online articles. Search terms included combinations of "artificial intelligence", "deep learning", "extended reality", "virtual reality", "computational modelling", "computational fluid dynamics", "3D printing", "congenital heart disease", "cardiovascular magnetic resonance", and "computed tomography". English-language original articles, reviews, and consensus statements were considered. Given the narrative design, studies were selected based on relevance and methodological contribution rather than formal systematic screening. Key Content and Findings: AI applications now span the full imaging-to-decision continuum, including image reconstruction, automated segmentation, quantitative assessment, phenotype recognition, and decision support. XR technologies enhance spatial understanding and pre-procedural planning, while computational modelling and three-dimensional (3D) printing enable patient-specific haemodynamic simulation and procedural rehearsal. Public datasets and integrated AI-XR-modelling pipelines are emerging. However, most evidence derives from feasibility and single-centre studies, and lesion-specific external validation remains limited. Conclusions: AI, XR, and computational modelling hold significant promise for advancing cross-sectional imaging in CHD. Future progress will depend on multicentre collaboration, lesion-stratified validation, workflow integration, and governance frameworks tailored to lifelong congenital care. Demonstration of clinical outcome benefit will be essential for widespread adoption.