Proposal for an Early Diagnosis System for Diffuse Idiopathic Skeletal Hyperostosis Using CT Images of Emergency Room Visitors
BioHealth Informatics · 2026 · 1 (1) · n/a
DOI: 10.67261/bhi.oa.2026-0010 ↗Kurz gefasst
Fachartikel in der Zeitschrift BioHealth Informatics, 2026. Verfasst von Manaho Kato und 4 weiteren Autor:innen. Die Arbeit nennt 10 Literaturangaben.
Abstract (Originalsprache)
Objective
Diffuse idiopathic skeletal hyperostosis (DISH) causes ossification of the anterior longitudinal ligament of the spine and is usually asymptomatic, so patients rarely seek care and early detection is difficult. We propose a system that automatically detects DISH from spinal computed tomography (CT) images of emergency room patients to support early diagnosis.
Methods
Whole-spine, thoracic, and lumbar CT images from 122 emergency room patients (61 with DISH and 61 non-DISH controls) at Wakayama Medical University Hospital were used. An uploaded CT folder is automatically processed by convolutional neural networks that first select spinal CT images and then classify them by imaging range. For each range, a U-Net semantic segmentation model trained on manually labeled ground-truth images generates DISH prediction images. A patient is diagnosed as positive if at least 1 image is detected as positive.
Results
The spinal CT selection and imaging-range classification models achieved accuracies of 91.0% and 82.3%, respectively. Segmentation recall and mean Intersection over Union (IoU) were 96.9% and 80.3% for whole-spine CT, 87.2% and 88.5% for thoracic CT, and 79.3% and 78.6% for lumbar CT. Per-patient diagnostic recall reached 100% for thoracic and lumbar CT and 75% for whole-spine CT.
Conclusion
The proposed U-Net-based system detects DISH with high recall, particularly for thoracic and lumbar CT, and may enable incidental early detection of DISH in emergency settings.
Zitieren
APA
Kato, M., Morikawa, Y., Teraguchi, M., Nakata, T., & Yoshino, T. (2026). Proposal for an Early Diagnosis System for Diffuse Idiopathic Skeletal Hyperostosis Using CT Images of Emergency Room Visitors. BioHealth Informatics, 1(1), n/a. https://doi.org/10.67261/bhi.oa.2026-0010
Vancouver
Kato M, Morikawa Y, Teraguchi M, Nakata T, Yoshino T. Proposal for an Early Diagnosis System for Diffuse Idiopathic Skeletal Hyperostosis Using CT Images of Emergency Room Visitors. BioHealth Informatics. 2026;1(1):n/a. doi:10.67261/bhi.oa.2026-0010
BibTeX
@article{Kato2026,
author = {Kato, Manaho and Morikawa, Yamato and Teraguchi, Masatoshi and Nakata, Tomonori and Yoshino, Takashi},
title = {Proposal for an Early Diagnosis System for Diffuse Idiopathic Skeletal Hyperostosis Using CT Images of Emergency Room Visitors},
journal = {BioHealth Informatics},
year = {2026},
volume = {1},
number = {1},
pages = {n/a},
doi = {10.67261/bhi.oa.2026-0010}
}
EUVIMED