Analysis of pediatric airway morphology using statistical shape modeling

Area of Science:

• Medical Imaging
• Computational Anatomy
• Pediatric Pulmonology

Background:

• Traditional airway morphology studies use limited measurements.
• Understanding pediatric airway tree structure is crucial for disease assessment.

Purpose of the Study:

• To apply statistical shape modeling (SSM) to pediatric airway trees.
• To detect overall airway shape differences between control and cystic fibrosis (CF) populations.

Main Methods:

• Segmented airway centerlines from CT scans of control and CF subjects.
• Utilized SSM to create a skeleton model of the large bronchi.
• Employed multivariate regression and cross-validation for shape analysis.

Main Results:

• SSM successfully modeled pediatric airway trees.
• Simulations confirmed the ability to detect shape differences.
• CF status was discriminated with 0.7 precision and recall; ROC AUC was 0.85.
• Significant SSM parameter differences were found between control and CF groups.

Conclusions:

• SSM is effective for identifying population-level airway shape variations.
• Subtle skeletal shape differences exist in the airways of children with CF compared to controls.