Craniorate: An Image-based, Deep-phenotyping Analysis Toolset, Repository, And Online Clinician Interface For Craniosynostosis
Justin Beiriger, BSE1, Madeleine Bruce, BA1, Wenzheng Tao, MS2, Ross Whitaker, PhD2, Jesse Goldstein, MD1.
1UPMC, Pittsburgh, PA, USA, 2University of Utah, Salt Lake City, UT, USA.
PURPOSE: The diagnosis and management of craniosynostosis involves subjective decision-making by craniofacial and neurosurgeons at the point of care. The purpose of this work is to provide a quantitative severity metric and point-of-care user interface to aid clinicians in the management of metopic craniosynostosis and to provide a platform for future research through deep phenotyping.METHODS: Ten craniofacial surgeons and 8 pediatric neurosurgeons rated the severity of over 80 CT scans of patients with metopic craniosynostosis and normal controls to establish a gold standard. A machine-learning algorithm was developed that quantifies the severity of metopic craniosynostosis, generating a Metopic Severity Score (MSS). MSS was compared to other craniometric measurements. CT imaging from multiple institutions were compiled to establish the spectrum of severity.
RESULTS: To date, almost 400 CT scans of patients with metopic synostosis have been contributed to our database. MSS was found to be more predictive of severity in metopic craniosynostosis than intra-frontal angle. Additionally, we have built a point-of-care user interface (craniorate.org) to accept contributions from collaborating institutions.
CONCLUSION: The resulting quantification of severity using MSS (Figure) has shown an improved capacity, relative to conventional measures, to automatically classify normal controls versus MS patients. We have mathematically described, in an objective and quantifiable manner, the distribution of phenotypes in metopic craniosynostosis with the potential to extend our work to other craniofacial abnormalities. 
Back to 2022 Posters