Changes in multimodal MRI parameters following orbital decompression surgery for thyroid-associated ophthalmopathy and their associations with clinical outcomes.

OBJECTIVE: To evaluate postoperative changes in multimodal quantitative magnetic resonance imaging (MRI) parameters after orbital decompression in thyroid-associated ophthalmopathy and explore their associations with clinical outcomes. METHODS: This retrospective study included 47 patients (61 eyes) assessed before and 6 months after surgery. Eye-level longitudinal outcomes were analyzed using linear mixed-effects models with random intercepts for patients and eyes nested within patients. Graves' Ophthalmopathy Quality of Life questionnaire (GO-QOL) outcomes were analyzed using patient-level mixed-effects models. MRI-clinical associations were assessed using generalized estimating equations or multivariable linear regression. Benjamini-Hochberg false discovery rate correction was applied within prespecified hypothesis families. RESULTS: After correction, proptosis and best-corrected visual acuity improved significantly, as did visual-function, appearance, and total GO-QOL scores. Multiple structural and diffusion-related MRI parameters also changed significantly after surgery. Greater reductions in corneal protrusion distance were associated with greater reductions in proptosis (q = 0.008). At the patient level, greater reductions in the vertical orbital apex area index and fat space apparent diffusion coefficient were associated with greater improvements in visual-function and total GO-QOL scores. No MRI parameter remained associated with appearance improvement after correction. A significant time-by-disease-subtype interaction was observed only for best-corrected visual acuity. CONCLUSION: Multimodal quantitative MRI detected postoperative structural and diffusion-related changes after orbital decompression. Reductions in the vertical orbital apex area index and fat space apparent diffusion coefficient showed exploratory associations with patient-reported visual-function improvement, although robustness varied across sensitivity analyses. These measures may complement, but not replace, clinical assessment.