Artificial Intelligence-Based Modeling to Aid in the Identification of Genetic Conditions in Infants with Congenital Heart Disease
- 1Baylor College of Medicine, Texas Children's Hospital
Abstract
BackgroundGenetic conditions are common in congenital heart disease (CHD) and affect care, but are often missed or diagnosed late. We asked whether routine clinical data can shortlist the most likely genetic conditions.
MethodsRetrospective single-center cohort of children with CHD and an echocardiogram who had either a confirmed genetic diagnosis or negative cytogenetic testing and CHD gene sequencing. Echocardiogram images and serial medical record data were added as model inputs in the order they become clinically available. Class-balanced logistic regression (LR) and a multimodal deep learning fusion network were fitted to the same inputs at each step. The primary metric was balanced top-3 accuracy (correct diagnosis among the top three, averaged across conditions), with bootstrap intervals.
ResultsThe cohort included 1,720 children (988 with common CHD-associated genetic conditions, 231 rare diagnoses, 501 negative testing). Using only inputs generated at the echocardiogram visit, LR reached a balanced top-3 accuracy 0.758 (chance 0.429). Serial records, radiology notes and ICD-10 codes, raised this to 0.809. Multimodal fusion performed similarly at each stage. With a genetic condition withheld from training, the fusion embedding placed affected children nearest to others sharing a gene group in 0.37 of cases versus 0.10 without the embedding (+0.268 [95% CI +0.242, +0.295]).
ConclusionsRoutine clinical data available at first echocardiogram can help identify likely genetic diagnoses, and later coded inputs ceiling. Multimodal fusion added no accuracy over LR, but places children sharing a genetic cause near one another, making nearest-neighbor retrieval informative for undiagnosed children.
Related
- Artificial Intelligence‐Optimized Image Biomarker Identification for Sinonasal Inverted Papilloma Diagnosis Using Image‐Enhanced Versus White‐Light Endoscopy
- Identification of adventitious lung sounds in children with respiratory illnesses in Bangladesh using artificial intelligence and digital auscultation
- Artificial intelligence-driven identification and mechanistic exploration of synergistic anti-aging compounds from Dengzhan Shengmai formulation
- Identification and control of potential occupational safety and health hazards of artificial intelligence: a scoping review
- Pipeline to enable guideline question identification from GoogleTrends with artificial intelligence
BibTeX
@misc{caiola2026artificial,
title = {{Artificial Intelligence-Based Modeling to Aid in the Identification of Genetic Conditions in Infants with Congenital Heart Disease}},
author = {M. Caiola and T. Thompson and S. B. Stephens and B. Salamat and K. Harris and C. Lara and Z. Coban-Akdemir and L. Zhang and M. Roth and A. Milosavljevic and A. C. Weitz and S. A. Morris},
howpublished = {medRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.12.26365328},
url = {https://www.medrxiv.org/content/10.64898/2026.10.12.26365328v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Artificial Intelligence-Based Modeling to Aid in the Identification of Genetic Conditions in Infants with Congenital Heart Disease
AU - M. Caiola
AU - T. Thompson
AU - S. B. Stephens
AU - B. Salamat
AU - K. Harris
AU - C. Lara
AU - Z. Coban-Akdemir
AU - L. Zhang
AU - M. Roth
AU - A. Milosavljevic
AU - A. C. Weitz
AU - S. A. Morris
PY - 2026
DA - 2026/10/12
PB - medRxiv
DO - 10.64898/2026.10.12.26365328
UR - https://www.medrxiv.org/content/10.64898/2026.10.12.26365328v1
AB - BackgroundGenetic conditions are common in congenital heart disease (CHD) and affect care, but are often missed or diagnosed late. We asked whether routine clinical data can shortlist the most likely genetic conditions. MethodsRetrospective single-center cohort of children with CHD and an echocardiogram who had either a confirmed genetic diagnosis or negative cytogenetic testing and CHD gene sequencing. Echocardiogram images and serial medical record data were added as model inputs in the order they become clinically available. Class-balanced logistic regression (LR) and a multimodal deep learning fusion network were fitted to the same inputs at each step. The primary metric was balanced top-3 accuracy (correct diagnosis among the top three, averaged across conditions), with bootstrap intervals. ResultsThe cohort included 1,720 children (988 with common CHD-associated genetic conditions, 231 rare diagnoses, 501 negative testing). Using only inputs generated at the echocardiogram visit, LR reached a balanced top-3 accuracy 0.758 (chance 0.429). Serial records, radiology notes and ICD-10 codes, raised this to 0.809. Multimodal fusion performed similarly at each stage. With a genetic condition withheld from training, the fusion embedding placed affected children nearest to others sharing a gene group in 0.37 of cases versus 0.10 without the embedding (+0.268 [95% CI +0.242, +0.295]). ConclusionsRoutine clinical data available at first echocardiogram can help identify likely genetic diagnoses, and later coded inputs ceiling. Multimodal fusion added no accuracy over LR, but places children sharing a genetic cause near one another, making nearest-neighbor retrieval informative for undiagnosed children.
ER -