Orbital CT deep learning models in thyroid eye disease rival medical specialists’ performance in optic neuropathy prediction in a quaternary referral center and revealed impact of the bony walls
Orbit · 2025
DOI: 10.1080/01676830.2025.2521868 ↗Kurz gefasst
Fachartikel in der Zeitschrift Orbit, 2025. Verfasst von Si Wei Kheok und 10 weiteren Autor:innen. Beteiligte Einrichtung: Department of Neuroradiology, Singapore General Hospital, Singapore, Singapore. Hauptthemen laut MEDLINE: Graves Ophthalmopathy, Deep Learning, Tomography, X-Ray Computed, Optic Nerve Diseases. Die Arbeit nennt 36 Literaturangaben.
Abstract (Originalsprache)
PURPOSE: To develop and evaluate orbital CT deep learning (DL) models in optic neuropathy (ON) prediction in patients diagnosed with thyroid eye disease (TED), using partial versus entire 2D versus 3D images for input. METHODS: Patients with TED ±ON diagnosed at a quaternary-level practice and who underwent orbital CT between 2002 and 2017 were included. DL models were developed using annotated CT data. The DL models were used to evaluate the hold-out test set. ON classification performances were compared between models and medical specialists, and saliency maps applied to randomized cases. RESULTS: 36/252 orbits in 126 TED patients (mean age, 51 years; 81 women) had clinically confirmed ON. With 2D image input for ON prediction, our models achieved (a) sensitivity 89%, AUC 0.86 on entire coronal orbital apex including bony walls, and (b) specificity 92%, AUC 0.79 on partial axial lateral orbital wall only annotations. ON classification performance was similar (p = 0.58) between DL model and medical specialists. CONCLUSIONS: DL models trained on 2D CT annotations rival medical specialists in ON classification, with potential to objectively enhance clinical triage for sight-saving intervention and incorporate model variants in the workflow to harness differential performance metrics.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Neuroradiology, Singapore General Hospital, Singapore, Singapore.
- SingHealth Duke-NUS Radiological Sciences Academic Clinical Programme, Singapore, Singapore.
- NUS Graduate School, National University of Singapore, Singapore, Singapore.
- Save Sight Institute, University of Sydney, Camperdown, Australia.
- Department of Oculoplastic Surgery, Sydney Eye Hospital, Sydney, Australia.
- Department of Oculoplastic Surgery, Singapore National Eye Centre, Singapore, Singapore.
ORCID-Profile: Si Wei Kheok, Guoyu Hu, Ming-Han Hugo Lee, Chen Pong Wong, Kaiping Zheng, Hla M Htoon, Zhu Lei, Alexander Sheng Ming Tan, Ling Ling Chan, Beng Chin Ooi, Lay Leng Seah
Zitieren
APA
Kheok, S. W., Hu, G., Lee, M. H., Wong, C. P., Zheng, K., Htoon, H. M., Lei, Z., Tan, A. S. M., Chan, L. L., Ooi, B. C., & Seah, L. L. (2025). Orbital CT deep learning models in thyroid eye disease rival medical specialists’ performance in optic neuropathy prediction in a quaternary referral center and revealed impact of the bony walls. Orbit. https://doi.org/10.1080/01676830.2025.2521868
Vancouver
Kheok SW, Hu G, Lee MH, Wong CP, Zheng K, Htoon HM, et al. Orbital CT deep learning models in thyroid eye disease rival medical specialists’ performance in optic neuropathy prediction in a quaternary referral center and revealed impact of the bony walls. Orbit. 2025. doi:10.1080/01676830.2025.2521868
BibTeX
@article{Kheok2025,
author = {Kheok, Si Wei and Hu, Guoyu and Lee, Ming-Han Hugo and Wong, Chen Pong and Zheng, Kaiping and Htoon, Hla M and Lei, Zhu and Tan, Alexander Sheng Ming and Chan, Ling Ling and Ooi, Beng Chin and Seah, Lay Leng},
title = {Orbital CT deep learning models in thyroid eye disease rival medical specialists’ performance in optic neuropathy prediction in a quaternary referral center and revealed impact of the bony walls},
journal = {Orbit},
year = {2025},
doi = {10.1080/01676830.2025.2521868}
}
EUVIMED