Related Experiment Video
Updated: Dec 8, 2025
Author Spotlight: Three-Dimensional Cephalometric Landmark Annotation Demonstration on Human Cone Beam Computed Tomography Scans
Published on: September 8, 2023
Combining deep learning with 3D stereophotogrammetry for craniosynostosis diagnosis
Guido de Jong1, Elmar Bijlsma2, Jene Meulstee3,4
1Department of Neurosurgery, Radboudumc, Nijmegen, The Netherlands. Guido.deJong@radboudumc.nl.
Insights
A deep learning algorithm accurately classified infant head shapes from 3D images, distinguishing craniosynostosis subtypes from healthy controls. This AI tool aids in the swift diagnosis of premature cranial suture fusion in infants.
Area of Science:
- Medical imaging
- Artificial intelligence
- Pediatric surgery
Background:
- Craniosynostosis involves premature fusion of infant cranial sutures, impeding brain and skull growth.
- Early diagnosis of craniosynostosis is crucial for managing cosmetic and functional issues.
- Accurate classification of craniosynostosis subtypes is essential for appropriate treatment.
Purpose of the Study:
- To evaluate a deep learning algorithm's ability to classify infant head shapes.
- To differentiate between healthy controls and three craniosynostosis subtypes: scaphocephaly, trigonocephaly, and anterior plagiocephaly.
- To assess the accuracy of AI in diagnosing craniosynostosis subtypes using 3D stereophotographs.
Main Methods:
- Collected 3D stereophotographs of infants with scaphocephaly (n=76), trigonocephaly (n=40), anterior plagiocephaly (n=27), and healthy controls (n=53).
- Employed a deep learning network to classify cranial shape data.
- Utilized stratified tenfold cross-validation for training and testing the algorithm.
Main Results:
- The deep learning algorithm achieved 99.5% accuracy in classifying cranial shapes.
- 195 out of 196 3D stereophotographs were correctly classified during testing.
- The AI demonstrated high accuracy in discriminating between craniosynostosis subtypes and healthy controls.
Conclusions:
- Deep learning algorithms trained on 3D stereophotographs can accurately classify craniosynostosis subtypes.
- AI-based analysis of cranial shape data offers a promising tool for rapid diagnosis.
- This technology can significantly aid in identifying infants with craniosynostosis, enabling timely intervention.
Abstract:
Craniosynostosis is a condition in which cranial sutures fuse prematurely, causing problems in normal brain and skull growth in infants. To limit the extent of cosmetic and functional problems, swift diagnosis is needed. The goal of this study is to investigate if a deep learning algorithm is capable of correctly classifying the head shape of infants as either healthy controls, or as one of the following three craniosynostosis subtypes; scaphocephaly, trigonocephaly or anterior plagiocephaly. In order to acquire cranial shape data, 3D stereophotographs were made during routine pre-operative appointments of scaphocephaly (n = 76), trigonocephaly (n = 40) and anterior plagiocephaly (n = 27) patients. 3D Stereophotographs of healthy infants (n = 53) were made between the age of 3-6 months. The cranial shape data was sampled and a deep learning network was used to classify the cranial shape data as either: healthy control, scaphocephaly patient, trigonocephaly patient or anterior plagiocephaly patient. For the training and testing of the deep learning network, a stratified tenfold cross validation was used. During testing 195 out of 196 3D stereophotographs (99.5%) were correctly classified. This study shows that trained deep learning algorithms, based on 3D stereophotographs, can discriminate between craniosynostosis subtypes and healthy controls with high accuracy.
