ABSTRACT: Changes in the spine of dogs are usually detected in clinical and in surgical practice. Few studies exist on musculoskeletal ultrasound anatomy of the thoracolumbar and lumbar segments of the normal spine of dogs. This study aimed to compare the normal musculoskeletal ultrasound anatomy of the T10-S1 vertebral segments with images obtained with magnetic resonance imaging (MRI), computed tomography (CT), and anatomical structures, and to establish the ability to identify structures using these modalities. Ultrasound scans allowed visualization of the muscles of the region, articular processes, spinous process, interspinous ligament, and yellow ligament in the lumbosacral window. Computed tomography images provided better bone details, compared to ultrasound images. Low-field MRI allowed the identification of the same structures identified with ultrasound imaging, and allowed the identification of cerebrospinal fluid, transverse processes, and provided improved detail of the intervertebral discs and spinal cord. Knowledge of ultrasound anatomy of the region may allow the the identification of muscle and ligament injuries. Thus, in cities where CT and MRI are inaccessible, ultrasonography of the region could be a good alternative to identify possible changes not observable with radiographic examination or to complement radiographic examination.
Using Convolutional Neural Network for edge detection in musculoskeletal ultrasound images
