Apparent diffusion coefficient: a predictor of the efficacy of glucocorticoid therapy in active moderate-severe thyroid eye disease: a retrospective cohort study

To assess the association between the apparent diffusion coefficient (ADC) of extraocular muscles (EOMs) and the efficacy of glucocorticoid (GC) therapy in patients with active and moderate-severe thyroid eye disease (TED).

Eighteen active and moderate-to-severe bilateral TED patients (responsive group, n = 8 patients (16 eyes); unresponsive group, n = 10 patients (20 eyes)) who underwent orbital long variable echo‑trains (Resolve)-diffusion tensor imaging (DTI) before GC therapy were retrospectively enrolled. The ADC values of each EOM were collected. The correlation between the ADC of EOMs and the efficacy of GC therapy was analyzed.

Age was significantly younger in the responsive group compared with the unresponsive group (52.13 ± 5.18 years vs. 58.80 ± 9.33 years, P value = 0.02).There were no statistical differences in sex, smoking, TED duration, visual acuity, intraocular pressure, exophthalmos, clinical activity score, the levels of free triiodothyronine, free tetraiodothyronine thyroid-stimulating hormone, and thyroid-stimulating hormone receptor antibody between the responsive group and the unresponsive group. Apparent diffusion coefficient in the responsive group were significantly lower than that in the unresponsive group in superior rectus muscle (1351.17 ± 232.12 vs. 1624.36 ± 193.70, P value = 0.001) and medial rectus muscle (1420.69 ± 192.05 vs. 1697.00 ± 199.89, P value = 0.001). After age, sex and CAS adjustment, ADC of the superior rectus muscle and medial rectus muscle were independent predictors of the response to GC treatment by binary logistic regression analysis. The cut-off value of ADC was 1582.00 × 10^− 3 mm²/s in superior rectus muscle with a sensitivity of 88% and a specificity of 65% respectively [ area under the curve (AUC) = 0.81]. The cut-off value of ADC was 1541.39 × 10^− 3 mm²/s in medial rectus muscle with a sensitivity of 75% and a specificity of 80% (AUC = 0.84). There were no significant difference in the predictive efficacy between the two ADC cut-off values.

Apparent diffusion coefficient values of EOMs from resolve DTI can be used as an indicator to predict the efficacy of GC therapy. Moderate-to-severe active TED patients with low ADC values at baseline in superior rectus muscle or medial rectus muscle may respond well to GC therapy.