Related Experiment Video
Updated: Jul 11, 2025
09:15
Creation of Patient-Specific Silicone Cardiac Models with Applications in Pre-surgical Plans and Hands-on Training
Published on: February 10, 2022
3.5K
SDF4CHD: Generative Modeling of Cardiac Anatomies with Congenital Heart Defects.
Fanwei Kong1, Sascha Stocker2,3, Perry S Choi4
1Department of Pediatrics, Cardiovascular Institute, Stanford University, Stanford.
Arxiv
|November 14, 2023
Summary
This study introduces a novel deep learning method to generate diverse virtual congenital heart disease (CHD) anatomies. This approach aids in creating patient-specific cardiac models for improved diagnosis and treatment planning.
Area of Science:
- Medical imaging and computational anatomy
- Artificial intelligence in healthcare
- Cardiovascular research
Background:
- Congenital heart disease (CHD) presents unique anatomical challenges requiring personalized treatments.
- Deep learning (DL) automates cardiac segmentation for normal anatomies but struggles with rare CHD variations.
- Existing generative models for cardiac anatomy are not optimized for the topological diversity of CHDs.
Conclusions:
- This generative approach offers a powerful tool for creating virtual cohorts of CHD anatomies.
- It has the potential to significantly enhance cardiac segmentation and computational modeling for CHD patients.
- The method facilitates improved diagnosis, treatment planning, and simulation for rare cardiovascular structural abnormalities.
