The Efficacy of Artificial Intelligence in Predicting Postoperative Mortality Rate in Patients with Congenital Heart Disease

A Systematic Review and Meta-analysis

Abstract

Background: Congenital heart disease (CHD) is a leading cause of morbidity and mortality in children requiring surgery. Accurate mortality prediction post-surgery remains challenging due to the limitations of traditional risk stratification systems. Artificial intelligence (AI) has emerged as a promising tool for enhancing predictive accuracy in this field.

Objective: This systematic review and meta-analysis aimed to evaluate the efficacy of AI in predicting postoperative mortality in patients with CHD.

Methods: Following PRISMA guidelines and PROSPERO registration (CRD42024557722), five databases were systematically searched up to July 16, 2024. Eligible studies included retrospective, prospective, or cross-sectional designs evaluating AI-based models for predicting mortality after CHD surgery. Data were extracted, and study quality was assessed using the PROBAST tool. Pooled sensitivity, specificity, and area under the curve (AUC) were calculated.

Results: Six studies involving 42,536 patients and 11 AI models were included. The pooled AUC was 0.90 (95% CI: 0.88–0.93), with a pooled sensitivity of 0.43 (95% CI: 0.23–0.65) and specificity of 0.96 (95% CI: 0.92–0.98). Subgroup analysis revealed that XGBoost (AUC: 0.93) and Gradient Boosting Machine (AUC: 0.91) had the highest predictive performance. All studies demonstrated a low risk of bias.

Conclusion: XGBoost and GBM, shows high specificity and promising accuracy in predicting postoperative mortality in CHD patients, outperforming traditional scoring systems. Further multicenter, prospective studies are needed to enhance generalizability and clinical implementation.