Related Experiment Video
Updated: Aug 8, 2026
06:09
Optic Nerve Sheath Point of Care Ultrasound: Image Acquisition
Published on: August 18, 2023
Automated Semisupervised Measurement of Optic Nerve Sheath Diameter From CT Following Traumatic Brain Injury
Emily Wittrup1, Alan Kay1, Elizabeth Lombard1
1Gilbert S. Omenn Department of Computational Medicine and Bioinformatics, University of Michigan, Ann Arbor, Michigan, USA, umich.edu.
International Journal of Biomedical Imaging
|August 7, 2026
Summary
This study introduces a novel automated pipeline for measuring optic nerve sheath diameter (ONSD) using computed tomography (CT) scans. The semisupervised deep learning approach enhances accuracy, even with limited training data.
Area of Science:
- Medical Imaging
- Artificial Intelligence in Medicine
- Ophthalmology
Background:
- Automated measurement of optic nerve sheath diameter (ONSD) is crucial in clinical settings.
- Deep learning for ONSD estimation faces challenges due to limited expert-labeled data.
Purpose of the Study:
- To develop a fully automated pipeline for ONSD estimation from CT scans.
- To leverage semisupervised learning to overcome data scarcity in medical image segmentation.
Main Methods:
- A modular pipeline combining deep learning and semisupervised learning was developed.
- The pipeline includes automated slice selection, optic nerve segmentation using labeled and unlabeled data, and ONSD measurement.
- Utilized axial head CT scans for validation.
Main Results:
- The semisupervised approach demonstrated superior generalization compared to supervised methods.
- Achieved an intersection over union (IoU) of 0.561 ± 0.141 and Dice coefficient of 0.707 ± 0.131 on unseen data.
- Mean absolute errors for ONSD measurement were as low as 1.779 mm (right) and 1.899 mm (left).
Conclusions:
- Semisupervised deep learning offers a viable solution for fully automated ONSD measurements.
- The framework is adaptable for medical imaging tasks with limited ground truth data.
- Highlights the potential for AI in improving diagnostic efficiency in ophthalmology.
