Detection of spondylosis deformans in thoracolumbar and lumbar lateral X-ray images of dogs using a deep learning network

Junseol Park , Hyunwoo Cho , Yewon Ji , Kichang Lee , Hakyoung Yoon 

Background: Spondylosis deformans is a degenerative condition of the spine that causes bony spurs and bridges between the vertebrae. It can be difficult to diagnose using radiography, and may be associated with disc disease and neurological signs in dogs.

Study: The authors aimed to develop a deep learning model based on the attention U-Net algorithm to segment the vertebral body and detect spondylosis deformans in thoracolumbar and lumbar lateral X-ray images of dogs.

Methods: A total of 265 X-ray images from 162 dogs were used to train, validate, and test the model. The images were manually labeled by 13 veterinary clinicians, and augmented by various techniques. The model performance was evaluated by the dice similarity coefficient (DSC) and Cohen’s kappa coefficient, and compared with the manual evaluation by the clinicians. The relationship between spondylosis deformans and clinical signs was also investigated.

Results: The model achieved high DSC values for the segmentation of the vertebral body, intervertebral disc space, and foramen, and high kappa values for the detection of spondylosis deformans, showing almost perfect agreement with the clinicians. The model was also fast and accurate in detecting spondylosis deformans in abdominal lateral X-ray images. The most common sites of spondylosis deformans were T12-T13 and L2-L3, and no significant association was found between spondylosis deformans and clinical signs.

Limitations: The study had some limitations, such as the small number of images for some grades of spondylosis deformans, the lack of histopathological confirmation of the lesions, the use of cropped images for the test dataset, and the possible bias of the manual labeling by the clinicians.

Conclusions: The study demonstrated the feasibility and usefulness of a deep learning model for the automatic segmentation of the vertebral body and detection of spondylosis deformans in dogs. The model could help veterinarians to quickly and accurately identify sites of disc instability and diagnose disc diseases. The model could also be applied to other vertebral and disc diseases in the future.

Disclaimer: The summary generated in this email was created by an AI large language model. Therefore errors may occur. Reading the article is the best way to understand the scholarly work. The figure presented here remains the property of the publisher or author and subject to the applicable copyright agreement. It is reproduced here as an educational work. If you have any questions or concerns about the work presented here, reply to this email.