J Ultrasound Med. 2025 May 28. doi: 10.1002/jum.16721. Online ahead of print.
ABSTRACT
OBJECTIVE: Congenital heart disease (CHD) is the most common birth defect and the leading cause of infant death from congenital anomalies. Limitations in standard-of-care fetal echocardiography lack hemodynamic insight. Cardiovascular computational modeling methods have been developed to simulate patient-specific morphology and hemodynamics, but are limited in applications for fetal diagnosis, as existing pipelines depend upon 3D CMR imaging data. There is no existing workflow for converting 2D echocardiograms into models of the fetal aorta. We aim to develop a methodology to create pulsatile 3D-aortic models from standard-of-care 2D echocardiograms to supplement fetal imaging with noninvasive predictions of hemodynamics in CHD diagnosis.
METHODS: Utilizing 2D fetal echocardiograms, edge detection algorithms are applied to delineate vessel boundaries. Cross-sectional diameters along the aortic arch and branch centerlines were segmented, integrated into 3D geometric models, and reconstructed using SimVascular. Patient-specific simulations were developed for three false-positive coarctation of the aorta (CoA) fetuses and 3 true positive CoA fetuses (postnatally confirmed), using echocardiogram and Doppler source data.
RESULTS: We propose a modeling methodology and set of boundary conditions that generate physiologically reasonable and cross-validated quantifications of fetal hemodynamics. Noninvasive predictions of fetal aortic pressures, flow streamlines, and vessel displacement offer insight into real-time hemodynamics and the stress of abnormal morphology on flow directions in the prenatal aorta.
CONCLUSIONS: We present a clinically useful pipeline for generating simulations of flow in the fetal aorta that capture fluid-structure interactions and generate noninvasive predictions of diagnostic hemodynamic indicators that could not previously be captured prenatally. This pipeline integrates into clinical diagnosis and offers insight into patient-specific physiology beyond a visualization of cardiac morphology alone, offering the potential to enhance the diagnostic precision of CHDs.
PMID:40434286 | DOI:10.1002/jum.16721
