3D spinal and rib cage predictors of brace effectiveness in adolescent idiopathic scoliosis

• Published in: BMC musculoskeletal disorders 2019-08, Vol.20 (1), p.384-384, Article 384
• Main Author: Pasha, Saba
• Format: Article
• Language: English
• Subjects: Adolescent; Adolescent idiopathic scoliosis; Aircraft; Braces; Care and treatment; Child; Child health; Diagnosis; Equipment Design; Female; Health aspects; Humans; Imaging, Three-Dimensional; Kyphosis; Kyphosis - diagnostic imaging; Kyphosis - therapy; Lordosis; Lordosis - diagnostic imaging; Lordosis - therapy; Lumbosacral Region - diagnostic imaging; Male; Models, Biological; Pediatrics; Predictive model; Prognosis; Radiography; Regression analysis; Retrospective Studies; Rib Cage - diagnostic imaging; Sagittal profile; Scoliosis; Scoliosis - diagnostic imaging; Scoliosis - therapy; Thoracic Vertebrae - diagnostic imaging; TLSO brace; Transverse plane; Treatment Outcome; Youth
• ISSN: 1471-2474; 1471-2474
• DOI: 10.1186/s12891-019-2754-2

Scoliotic braces are the standard of curve for management of moderate spinal deformities in pediatric patients. The effectiveness of this treatment method has been shown; however, the spinal and rib cage parameters, in the three anatomical planes, that are associated with bracing outcome in adolescent idiopathic scoliosis (AIS) are not fully identified. A total number of 45 right thoracic AIS patients who had received a thoraco-lumbo-scaral brace for the first time were included retrospectively. For each patient, radiographic images at three visits, pre-brace, in-brace, and at least 1 year after the first brace fit were included. Age, sex, Risser sign, and curve type at pre-brace, and thoracic and lumbar frontal and sagittal Cobb angles, thoracic and lumbar apical rotations, sagittal and frontal balances at pre-brace and in-brace were determined. Two sagittal curve types (hypothoracolumbar and normal/hyperthoracolumbar kyphosis), two rib cage types based on the costovertebral joints (drooping and horizontal), and two axial shapes of the spine (S shaped and V shaped) were used to stratify the patients. Feature selection and linear regression with regularization determined the parameters and the interaction terms that predicted the brace effectiveness significantly. Smaller in-brace thoracic Cobb and larger in-brace lordosis predicted brace effectiveness, p