Data for AI in Congenital Heart Defects: Systematic Review

Area of Science:

• Medical Imaging
• Artificial Intelligence in Healthcare
• Prenatal Diagnostics

Background:

• Congenital heart disease (CHD) detection rates prenatally are low, posing a challenge for obstetric care.
• Artificial Intelligence (AI) presents a potential solution for enhancing the accuracy of prenatal CHD predictions.
• Existing research on AI for prenatal CHD detection requires systematic evaluation.

Purpose of the Study:

• To systematically review and analyze the application of AI in prenatal detection of congenital heart disease (CHD).
• To identify the predominant AI methodologies and data modalities used in this field.
• To provide insights for advancing AI-driven CHD detection in clinical practice.

Main Methods:

• A systematic literature review was conducted following PRISMA guidelines.
• Searches were performed across PubMed, Embase, and Web of Science databases.
• 621 articles were screened, with 28 studies included in the final analysis.

Main Results:

• Deep Learning (DL) is the most frequently employed AI technique for prenatal CHD detection.
• Ultrasound and Magnetic Resonance Imaging (MRI) sequences are the primary data types utilized.
• The review identified a growing body of research in AI for prenatal CHD screening.

Conclusions:

• AI, particularly DL, holds significant potential for improving prenatal CHD detection rates.
• Standardization of data types and AI approaches is needed for clinical translation.
• Further research is warranted to optimize AI algorithms for real-world obstetric applications.