Project 441778

Scoliosis is an orthopaedic condition resulting in curvature and rotation of the spine. It typically first appears during adolescence, and affects females more than males. Deformity of the torso is one of the symptoms. Cosmetic improvement of the torso resulting in better symmetry is appreciated by patients, especially in adolescent females. Scoliosis is diagnosed and monitored using x-rays. Unfortunately, x-rays expose young patients to the effects of radiation including a documented increase in cancer risk. Surface topography (ST) is a non-invasive three dimensional (3D) assessment of the torso shape. Using a laser scanner, 3D images of the torso are acquired and the deformity of the torso is measured. The severity of scoliosis is then quantified using indices reflecting the symmetry of the torso. Previous studies using surface topography with 2D measurements instead of the available 3D data were not able to accurately predict the severity of the spinal deformity. In our recent work, we introduced a novel 3D asymmetry technique that does not rely on markers placed on the torso or on simple 2D measurements. Our St measures were able to quantify the severity and progression of scoliosis. In the current proposal, we will develop artificial intelligence techniques to better use the surface topography parameters to estimate the actual shape of the underlying spinal curvature. The developed methods will be designed to ensure that no moderate/severe curves are missed and that all progressing curves are detected to make sure that patients are not missing important treatment opportunities, while dramatically reducing the x-ray radiation exposure to patients.