Artificial Intelligence in Congenital Heart Surgery: A Scoping Review and Primer for Surgeons

Background:
Congenital heart surgery encompasses a wide spectrum of complex cardiac defects, many of which demand specialised perioperative management and tailored surgical planning. Artificial intelligence (AI), including machine learning, is gaining prominence as a tool to optimise clinical decision-making and achieve better outcomes. This scoping review aims to map and summarise the existing applications of AI modalities in congenital heart surgery.

Methods:
A comprehensive search of MEDLINE, Embase, and Web of Science was performed, combining terms for AI with terms for congenital heart disease and surgery.

Results:
2871 articles were retrieved from the search, of which 93 studies were included. The majority of studies focused on outcome prediction and imaging-based applications. Smaller proportions addressed decision making and data augmentation, omics integration, benchmarking and quality improvement. The majority of studies examined heterogeneous CHD populations, with tetralogy of Fallot and single ventricle physiology most frequently represented.

Conclusions:
AI applications in congenital heart surgery are rapidly expanding across diverse domains, with early studies showing encouraging potential to support diagnostics, guide surgical decision-making, and improve perioperative outcomes. However, most models remain in the preliminary stage with limited external validation. Emerging advances in AI may further accelerate progress, but careful evaluation and integration are essential to translate this promise into tangible clinical benefits.