Deep Learning-Based Approach to Automatically Assess Coronary Distensibility Following Kawasaki Disease

Area of Science:

• Cardiovascular Research
• Pediatric Vasculitis
• Medical Imaging

Background:

• Kawasaki disease (KD) is a leading cause of acquired heart disease in children, characterized by acute vasculitis affecting coronary arteries (CA).
• Coronary artery aneurysms are a known complication, but the long-term impact on CA distensibility and its correlation with imaging findings remain unclear.
• Optical coherence tomography (OCT) reveals CA wall damage in KD, necessitating investigation into its relationship with vessel stiffness over time.

Purpose of the Study:

• To link OCT findings of CA damage with vessel distensibility changes in KD patients.
• To analyze the temporal evolution of coronary stiffness using a novel deep learning segmentation system.
• To evaluate the correlation between coronary stiffness and the severity of KD-related vascular damage.

Main Methods:

• A deep learning system was developed for automated coronary artery segmentation from angiography and OCT data.
• Coronary artery caliber changes over the cardiac cycle were measured to assess distensibility.
• Longitudinal analysis of distensibility was performed on KD patients with serial CA angiography over an average of 14.5 years.

Main Results:

• Coronary artery distensibility in KD patients negatively correlated with the severity of OCT-identified vascular damage.
• KD patients exhibited significantly increased CA wall stiffness at 1 year post-diagnosis, which subsequently plateaued.
• Persistent CA aneurysms were associated with a statistically significant increase in wall stiffness over time compared to regressed aneurysms.

Conclusions:

• Automated deep learning analysis of coronary artery distensibility correlates with OCT findings in KD patients.
• Coronary artery stiffness in KD patients peaks at 1 year post-diagnosis and shows long-term persistence.
• Increased coronary stiffness is more severe and progressive in KD patients with persistent aneurysms, indicating significant long-term vascular sequelae.