Skip to main content

SYSTEMATIC REVIEW article

Front. Cardiovasc. Med.
Sec. Pediatric Cardiology
Volume 12 - 2025 | doi: 10.3389/fcvm.2025.1473544
This article is part of the Research Topic Imaging, diagnosis and interventional treatment of congenital heart disease in children View all articles

DIAGNOSTIC ACCURACY OF ARTIFICIAL INTELLIGENCE MODELS IN DETECTING CONGENITAL HEART DISEASE IN THE SECOND-TRIMESTER FETUS THROUGH PRENATAL CARDIAC SCREENING: A SYSTEMATIC REVIEW AND META-ANALYSIS

Provisionally accepted
Dina Lies Dina Lies 1,2*Yosilia Nursakina Yosilia Nursakina 3
  • 1 RSUPN Dr. Cipto Mangunkusumo, Jakarta, Indonesia
  • 2 National Cardiovascular Center Harapan Kita (Indonesia), Jakarta, Jakarta, Indonesia
  • 3 School of Public Health, Faculty of Medicine, Imperial College London, London, England, United Kingdom

The final, formatted version of the article will be published soon.

    Background Congenital heart disease (CHD) is responsible for a significant portion of morbidity, infant mortality, and the highest burden on healthcare costs on a global scale. Early diagnosis and prompt treatment of CHD contribute to enhanced neonatal outcomes and survival rates, but there is a low rate of proficient examiners in remote regions. Therefore, artificial intelligence (AI)-empowered ultrasound provides a potential solution in improving the diagnostic accuracy of fetal CHD screening.Methods Literature searching is done across 7 databases for systematic review. Articles are retrieved based on PRISMA Flow 2020 as well as on inclusion and exclusion criteria. Eligible diagnostic data will be further meta-analyzed and the risk of bias was tested using QUADAS-AI.Findings A total of 374 studies were screened for eligibility, but only 9 studies were included. Most studies utilized deep learning (DL) models using either ultrasound or echocardiograpy images. Overall, the AI models performed exceptionally well in accurately identifying normal and abnormal ultrasound images. Meta-analysis is done for 9 studies about CHD diagnosis, resulting in an pooled sensitivity of 0.89 [0.81 -0.94], specificity of 0.91 [0.87 -0.94], and AUC of 0.952 with random effects model.Although there are several limitations to overcome to be able to implement AI models in clinical practice, overall AI showed promising results in CHD diagnosis. Nevertheless, prospective studies comparing AI algorithms to conventional methods with bigger datasets and more inclusive population are needed.

    Keywords: artificial intelligence, congenital heart disease, Meta-analysis, Prenatal cardiac examination, Ultrasonography

    Received: 09 Aug 2024; Accepted: 27 Jan 2025.

    Copyright: © 2025 Lies and Nursakina. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

    * Correspondence: Dina Lies, RSUPN Dr. Cipto Mangunkusumo, Jakarta, Indonesia

    Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.