deep-learning-based-segmentation-of-gravity-loaded-human-spine

Deep Learning-Based Segmentation of Gravity-Loaded Human Spine

Accurate analysis of spinal alignment under gravity-loaded conditions is essential for diagnosing and managing musculoskeletal disorders such as scoliosis ...

R&D Center of Shenzhen Angell Technology Co.

The study aimed to introduce a novel imaging method that generates large-coverage, weight-bearing, and 3D images of the whole spine. The proposed system comprises an X-ray tube, a flat panel detector, and a standing platform. The standing platform rotates the imaged subject, allowing for the acquisition of serial fluoroscopic images from different angles which can be used to create 3D images. To increase the longitudinal coverage, we apply a segmental scanning pattern in which the imaged region is scanned in segments and stitched. To address the issue of data inaccuracy between the segments, redundant areas are set at margins of the segmental images, and registration and stitching algorithms are applied. We conducted validation experiments to evaluate radiation dose and image quality. The dose was evaluated using the volume CT dose index (CTDI). For image quality evaluation, we measured the low-contrast and spatial resolution. Additionally, we conducted a clinical study consisting of 30 volunteers with adolescent idiopathic scoliosis who were imaged by our method, and the images were subjectively assessed based on image noise, artifacts, anatomical coverage, diagnostic confidence, and overall quality. The CTDI was 1.23 mGy, and the low-contrast resolution was 0.6% at 4 mm and the spatial resolution was 8 lp/cm. The clinical images were generally of good quality, with high scores for all factors evaluated. Our method successfully generates large-coverage, weight-bearing, and 3D images of the whole spine with high image quality and low radiation dose. It shows potential for wider clinical applications for various musculoskeletal conditions.

Weight-bearing cone-beam CT with extensive coverage for volumetric imaging in adolescent idiopathic scoliosis: system implementation and initial validation.

Zengtong Chen

Biography

JoVE Journal Publications

Other Publications