A Radiation-Free Classification Pipeline for Craniosynostosis Using Statistical Shape Modeling.

Area of Science:

• Medical imaging
• Computational anatomy
• Pediatric surgery

Background:

• Craniosynostosis, premature fusion of skull sutures, causes abnormal head shape.
• Computed tomography (CT) is standard but involves radiation.
• 3D photogrammetry offers a radiation-free alternative for head shape analysis.

Purpose of the Study:

• To develop and present a novel, automated classification pipeline for diagnosing craniosynostosis.
• To establish the first publicly available statistical shape model for infant head morphology (under 1.5 years).
• To conduct the largest classification study for craniosynostosis to date.

Main Methods:

• A classification pipeline utilizing a statistical shape model built from 3D photogrammetric surface scans.
• Development of pathology-specific submodels.
• Public release of the statistical shape model and submodels.

Main Results:

• Achieved 97.8% accuracy in classifying craniosynostosis, comparable to CT-based methods.
• Demonstrated the statistical shape model's performance is similar to existing head models.
• Established a new benchmark for classification studies in craniosynostosis.

Conclusions:

• The developed shape-model-based approach provides state-of-the-art, radiation-free diagnosis of craniosynostosis.
• The publicly available model facilitates craniosynostosis assessment on both real and synthetic data.
• This work advances non-invasive diagnostic tools for pediatric head shape abnormalities.