Computer aided evaluation of ankylosing spondylitis

Ankylosing spondylitis is a disease of the vertebra where abnormal bone structures (syndesmophytes) grow at intervertebral disk spaces. We provide a quantitative measure of the syndesmophytes using high resolution CT images. The first part of our algorithm segments the whole vertebra using a cascade of successive level sets that address the conflicting requirements of robustness to boundary discontinuities and flexibility needed to capture the fine structures caused by the pathology. The second part of the algorithm segments the syndesmophytes from the vertebral body and quantitates them. We use curvature information to locate the rim and end plates of the vertebra from where syndesmophytes develop. The results of our algorithm were validated by comparison with a semi-quantitative evaluation by a medical expert made by visually inspecting 3-D CT scans. Pearson correlation between the two evaluations is 0.898 (p<0.0001), and Spearman correlation is 0.906 (p<0.0001)