Loss of anterior concavity of the first sacrum can predict spinal involvement in ankylosing spondylitis

In this study, we evaluated the frequency of squaring of the first sacrum (S1), defined as the loss of anterior concavity, in patients with ankylosing spondylitis (AS). We also determined the interobserver reliability in the assessment of S1 squaring and the relationships of S1 squaring with MRI findings and the modified Stoke Ankylosing Spondylitis Spinal Score (mSASSS). To this end, we performed a retrospective study of 100 patients with AS (mean age 33.2 years; range 19–57 years) and 100 control patients (mean age 35.6 years; range 19–50 years). Four experienced radiologists independently assessed the presence of S1 squaring in the AS and control groups. The frequencies of S1 squaring as scored by the four observers were 47, 48, 46, and 42 in the AS group and 3, 6, 4, and 6 in the control group. The interobserver agreement among the four observers with respect to S1 squaring was excellent (κ value 0.80) in the AS group and fair to good (κ value 0.61) in the control group. In patients with AS, the presence of S1 squaring showed fair to good agreement with the MRI changes (κ value 0.74). Moreover, the mSASSSs of patients with versus without S1 squaring were significantly different (mean 23.9 vs 7.0, p < 0.001). In conclusion, S1 squaring is relatively common in patients with AS. Moreover, S1 squaring is closely correlated with MRI changes and significantly associated with the mSASSS. Assessment of S1 squaring could be a simple method that is potentially useful for predicting early spinal structural involvement in patients with AS.