EP01.47: Accurate AI measurements of ventricular area in second trimester fetal ultrasound video clips

Screening for congenital heart disease (CHD) and assessing cardiac function are amongst the most difficult tasks in prenatal diagnosis. Ventricular size disproportion is often a sign of underlying CHD or dysfunction. However, such measurements are time consuming and challenging to implement, and, therefore, are rarely performed during fetal heart screening. We aimed at evaluating an AI system for automated measurement of right and left ventricle area in ultrasound video clips. We included 30 video clips (1.5-5.1s duration, 18-24 weeks GA, 10 CHD cases) recorded with a GE Expert 22 or Voluson E8, featuring at least one frame of a 4-chamber (4C) view. The AI system was developed independently from these cases. The AI system uses an entire video clip to measure right and left ventricular chamber end-diastolic area on 4C views, and outputs a delimitation of the ventricles on a representative frame. Three experts evaluated each case. First, they identified a representative image of the 4C view at end-diastole and manually segmented RV and LV endocardium, independently from the algorithm. In addition, experts performed “AI-aided” ventricular measurements of the same cases by reviewing the delimitations generated by the AI and manually editing them if needed, without changing frames. Table 1 shows analysis for measurement comparisons, using the mean of independent experts as reference. AI-aided measurements were similar to AI alone, possibly because experts performed only small adjustments. Similar results held on the subgroup of CHD cases. AI measurement of the ventricles is a promising tool for accurate and reproducible cardiac assessment in fetal ultrasounds. ICC: intraclass correlation coefficient BA: Bland Altman bias and limits of agreement (mm2) Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Authors

Institutions

Publication Details

Journal
Ultrasound in Obstetrics and Gynecology
Published
2024-09-01
DOI
https://doi.org/10.1002/uog.28061
Primary Topic
Fetal and Pediatric Neurological Disorders
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

EP01.47: Accurate AI measurements of ventricular area in second trimester fetal ultrasound video clips

C. Gardella, Jennifer Lam‐Rachlin, Bertrand Stos, Greggory R. DeVore et al.
Ultrasound in Obstetrics and Gynecology
Fetal and Pediatric Neurological Disorders
article

EP01.47: Accurate AI measurements of ventricular area in second trimester fetal ultrasound video clips

C. Gardella, Jennifer Lam‐Rachlin, Bertrand Stos, Greggory R. DeVore, Marilyne Lévy, Nathan S. Fox, Valentin Thorey, Nadav Schwartz, Andrei Rebarber, E. Askinazi, M. De Boisredon
article en

Abstract

Screening for congenital heart disease (CHD) and assessing cardiac function are amongst the most difficult tasks in prenatal diagnosis. Ventricular size disproportion is often a sign of underlying CHD or dysfunction. However, such measurements are time consuming and challenging to implement, and, therefore, are rarely performed during fetal heart screening. We aimed at evaluating an AI system for automated measurement of right and left ventricle area in ultrasound video clips. We included 30 video clips (1.5-5.1s duration, 18-24 weeks GA, 10 CHD cases) recorded with a GE Expert 22 or Voluson E8, featuring at least one frame of a 4-chamber (4C) view. The AI system was developed independently from these cases. The AI system uses an entire video clip to measure right and left ventricular chamber end-diastolic area on 4C views, and outputs a delimitation of the ventricles on a representative frame. Three experts evaluated each case. First, they identified a representative image of the 4C view at end-diastole and manually segmented RV and LV endocardium, independently from the algorithm. In addition, experts performed “AI-aided” ventricular measurements of the same cases by reviewing the delimitations generated by the AI and manually editing them if needed, without changing frames. Table 1 shows analysis for measurement comparisons, using the mean of independent experts as reference. AI-aided measurements were similar to AI alone, possibly because experts performed only small adjustments. Similar results held on the subgroup of CHD cases. AI measurement of the ventricles is a promising tool for accurate and reproducible cardiac assessment in fetal ultrasounds. ICC: intraclass correlation coefficient BA: Bland Altman bias and limits of agreement (mm2) Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Ultrasound in Obstetrics and GynecologyVol. 64(S1)
Institut Necker Enfants Malades (FR), University of Pennsylvania (US), Icahn School of Medicine at Mount Sinai (US)
Fundació Institut de Recerca Hospital Universitari Vall d’Hebron
Zero hunger
Openalex Percentile: Top 76%
Fetal and Pediatric Neurological Disorders
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.