Funding outcome

Adolescent Idiopathic Scoliosis is a three-dimensional abnormal spinal curvature. It affects 3% of adolescents. Females have a higher risk of having larger curvature. Early diagnosis and treatment are the key to controlling or reducing the curve progression. The traditional method to estimate the risk of progression is based on a) the severity of the curvature in all 3 planes (coronal, transverse, and sagittal), b) bone age, c) curve type, and d) gender. The first 3 parameters are measured from radiographs requiring regular X-rays taken at every clinic visit. Exposing children to ionizing radiation is undesirable. My supervisor's team has developed innovative imaging methods including ultrasound and surface topography to capture curvature information. In addition, the bone quality of the spine can be extracted from ultrasound images. However, the current ultrasound system with built-in position and orientation sensor is bulky, and the laser scanner which is used to capture surface topography is not portable. The objectives of my project are 1) to develop a low-cost portable and user-friendly imaging tool, which is to integrate a wireless ultrasound scanner with a small stereo camera, to capture internal spine structure and back surface images, 2) to apply AI to reconstruct 3D ultrasound spine and surface images automatically in real-time, 3) to extract curvature parameters and bone quality automatically, and 4) to utilize AI to develop and validate a prediction model to estimate the risk of progression of scoliosis. In total, 150 patient datasets will be collected and used for both model development and validation.