A combined score based on anterior chest wall CT attenuation to aid recognition of SAPHO syndrome: a...

To investigate whether quantitative anterior chest wall CT attenuation can aid recognition of SAPHO syndrome, to develop a simple combined score, and to evaluate its discriminatory performance in a validation set.

We retrospectively enrolled 260 patients with SAPHO syndrome and 260 age- and sex-matched healthy controls. CT attenuation values were measured at five sites: the left and right clavicular heads, the left and right upper manubrium, and the sternal angle, and were normalized to the CT attenuation of an internal reference vertebra. Using stratified random sampling (7:3), the total sample was divided into a training set (364 cases: 182 SAPHO and 182 controls) and a validation set (156 cases: 78 SAPHO and 78 controls). In the training set, ROC analysis was used to compare individual ratios and additive combinations, establish the combined score, and determine the optimal cutoff. The fixed cutoff was then applied to the validation set. In addition, a subgroup analysis was performed in SAPHO patients without obvious anterior chest wall CT morphological changes.

In the training set, the combined score achieved an AUC of 0.980 (95% CI: 0.969-0.991). At the optimal cutoff of >5.412, sensitivity and specificity were both 93.41%. In the validation set, using the same cutoff, the AUC was 0.976 (95% CI: 0.958-0.993), with a sensitivity of 92.3% and a specificity of 84.6%. Among 96 SAPHO patients without obvious anterior chest wall CT morphological changes, 81 (84.4%) had combined scores above the fixed cutoff; the AUC for distinguishing this subgroup from healthy controls was 0.956 (95% CI: 0.935-0.974).

The combined score based on multi-site anterior chest wall CT attenuation demonstrated high discriminatory ability between SAPHO patients and healthy controls and showed consistent performance in the internal validation set. In patients without obvious anterior chest wall CT morphological changes, the score provided additional quantitative imaging information and may serve as an objective complement to conventional morphological assessment.