Publications by authors named "D Dunaway"
Head shape changes following spring-cranioplasty for craniosynostosis (CS) can be difficult to predict. While previous research has indicated a connection between surgical outcomes and calvarial bone microstructure ex-vivo, there exists a demand for identifying imaging biomarkers that can be translated into clinical settings and assist in predicting these outcomes. In this study, ten parietal (8 males, age 157 ± 26 days) and two occipital samples (males, age 1066 and 1162 days) were collected from CS patients who underwent spring cranioplasty procedures.
View Article and Find Full Text PDFArticle Synopsis
- * The researchers utilized a dataset of 3D head shapes, enhanced using a new data augmentation method, to train the SD-VAE model, which allows for detailed analysis of both overall head shapes and specific anatomical regions.
- * The findings enable syndrome classification and help to predict outcomes of craniofacial surgeries, thus improving diagnostic techniques and surgical planning, with the code shared on GitHub for further research.
Article Synopsis
- Advancements in AI, specifically the Swap Disentangled Variational Autoencoder (SD-VAE), allow for objective assessment of changes in head shape and facial morphology following craniofacial surgery.
- The study analyzed data from 56 patients with Apert and Crouzon syndromes who underwent midfacial procedures, comparing their post-surgery shape changes to a healthy population using 3D mesh analysis.
- The findings suggest that AI can improve the evaluation of surgical outcomes by quantifying regional and global shape changes, ultimately enhancing decision-making in surgical practices.
Objective: To evaluate optic nerve head (ONH) morphology in children with craniosynostosis versus healthy controls.
Design: Single-center, prospective cohort study.
Methods: Handheld optical coherence tomography (OCT) was performed in 110 eyes of 58 children (aged 0-13 years) with craniosynostosis.
Article Synopsis
- The study focuses on using AI to assist in diagnosing syndromic craniosynostoses like Apert, Crouzon, Muenke, Pfeiffer, and Saethre Chotzen syndromes from facial photographs.
- Researchers analyzed 2,228 photos from 541 patients over 44 years, aiming to identify features that distinguish these syndromes from non-syndromic cases.
- The AI model successfully diagnosed 70.2% of cases with a significant correlation between certain genotypes and milder disease phenotypes in Crouzon-Pfeiffer syndrome, suggesting new diagnostic avenues.