Automated Cobb Angle Measurement in Adolescent Idiopathic Scoliosis: Validation of a Previously-Published Deep Learning Method

The severity of scoliosis and surgical decisions are determined based on accurate measurement of the Cobb angle of the spine. There are several previously published deep learning models for automated measurement of Cobb angle, but none are externally validated in severe scoliosis patients. We evaluated the external performance of a previously published deep learning method for Cobb angle measurement in 2278 full-spine X- rays of 860 severe scoliosis patients. The model performed poorly and missed several vertebrae when labelling landmarks. Findings underscore the importance of external validation studies to assess model performance in patient subgroups with varying levels of scoliosis severity.