Funding outcome

Adolescent Idiopathic Scoliosis (AIS) is a three-dimensional (3D) spine deformity. Approximately 30% of children with AIS will require treatment. Brace treatment is the most effective non-surgical treatment to stop curve progression. There is a relatively short window of opportunity during the rapid growth spurt of adolescence for brace treatment to be effective. If brace treatment fails and the curve progresses significantly, surgery will be required. Scoliosis surgery has a life-altering impact and is a complicated operation undertaken with significant risks, including death, paraplegia, and a 6-9 month recovery period. The effectiveness of brace treatment depends on 3 factors: 1) remaining growth and curve progression risk, 2) the design of the brace, which affects the amount of in-brace correction, and 3) patients' compliance, which is affected by the comfort of the brace and the patient's belief in the likelihood of success or failure. Currently, no accurate prediction model can estimate brace treatment outcomes at the early stage. Most models predict final brace treatment outcomes, and the accuracy is only between 60 to 75%. However, predicting brace treatment outcome at the first follow-up out-of-brace clinic, 6 months after the brace is prescribed, is important because a) early prediction allows clinicians to personalize treatment plans more effectively, and b) early feedback may convince patients to achieve better adherence to the brace wear to improve brace effectiveness. This study aims to apply artificial intelligence technologies to develop a prediction model based on spine curvature, in-brace correction, brace wear compliance, and patients' demographic information at the initial visit and first in-brace scoliosis clinic, usually 2 months after the brace is prescribed. The outcomes of this research can assist the scoliosis team in delivering a more effective treatment plan for children who require brace treatment.