Standardized 2D Ultrasound Sequences for Fetal Cardiac Screening: A Platform for AI Integration.

Area of Science:

• Medical Imaging
• Cardiology
• Prenatal Diagnosis

Background:

• Accurate identification of fetal cardiac structures is crucial for diagnosing congenital heart defects (CHDs).
• Interobserver variability in fetal echocardiography interpretation poses a clinical challenge.
• Fetal cardiac MRI is limited by acquisition time and availability, necessitating optimized ultrasound techniques.

Purpose of the Study:

• To investigate the reliability of a layered imaging approach with standardized B-mode and color Doppler ultrasound protocols.
• To enhance interobserver agreement in the interpretation of fetal echocardiograms.
• To improve the diagnostic accuracy and standardization of fetal cardiac assessments.

Main Methods:

• Systematic evaluation of 209 B-mode and 205 color Doppler fetal recordings by five expert supervisors.
• Quantification of interobserver agreement using the prevalence-adjusted bias-adjusted kappa (PABAK) coefficient.
• One-way ANOVA and post hoc analysis to assess consistency across anatomical layers.

Main Results:

• Most fetal cardiac anatomical structures and functional features were consistently identified across observers.
• Certain challenging structures were identified due to smaller size, lower visibility, or presentation variability.
• The study explored the potential for improved diagnostic accuracy through structured imaging and statistical analysis.

Conclusions:

• Integrating structured imaging sequences and advanced statistical methods can enhance diagnostic accuracy in fetal cardiac assessments.
• This study provides insights for developing more reliable fetal echocardiography protocols.
• The findings support the development of standardized approaches to reduce interobserver variability in CHD diagnosis.