Orbital MRI and clinical activity score for predicting steroid treatment response in moderate-to-severe Graves'...

To evaluate the potential role of pre-treatment clinical activity score (CAS) and extraocular muscle (EOM) signal intensity ratio (SIR) values on orbital MRI in predicting the steroid response in patients with moderate-to-severe active Graves' orbitopathy (GO).

The data of 51 patients with moderate to severe GO (CAS ≥ 3) were retrospectively evaluated. The patients were categorized into two groups: steroid-resistant (n = 25) and steroid-responsive (n = 26). Demographic and clinical characteristics, CAS, VISA (vision, inflammation, strabismus, appearance), laboratory data and orbital MRI measurements were compared. Two standardized MRI sequences were utilized for analysis: T2-weighted short-tau inversion-recovery (T2w-STIR) and contrast-enhanced T1-weighted fat-suppressed (T1w-CE) imaging. Predictors of steroid resistance were identified through multivariate logistic regression analysis, and ROC curve was performed to determine predictive performance and optimal cut-off values.

Pre-treatment CAS (5.05 ± 1.07 vs 4.00 ± 0.93, p = 0.002) and T2w-STIR SIR (4.77 ± 0.93 vs 3.95 ± 1.32, p = 0.031) were both significantly higher in the steroid-resistant group. Higher pretreatment CAS (OR 2.380, p = 0.001) and T2w-STIR SIR (OR 1.862, p = 0.040) predicted steroid resistance. ROC analysis demonstrated good discriminative performance for CAS (AUC = 0.786); with a cut-off value > 4 yielding 71.4% sensitivity and 68.2% specificity. T2w-STIR SIR with a threshold value > 3.6 provided high sensitivity (94.4%) but limited specificity (37.5%), indicating moderate overall accuracy (AUC = 0.673).

Pre-treatment CAS and T2w-STIR SIR values may serve as potential predictors of steroid resistance in moderate-to-severe active GO. Early identification of high-risk patients may facilitate consideration of alternative treatments such as immunosuppressive agents or radiotherapy. Larger scale prospective studies are required to validate optimal T2w-STIR SIR cut-off values and the role of imaging biomarkers in risk stratification.