Diagnostic value of MRI and diffusion-weighted imaging in dysthyroid optic neuropathy: a cross-sectional study providing insights into predictive parameters and pathogenesis.

Purpose: To compare magnetic resonance imaging (MRI) and clinical characteristics between patients with moderate-to-severe Graves orbitopathy (GO) and dysthyroid optic neuropathy (DON), and to assess the diagnostic value of these factors for DON.

Methods: Monocentric prospective cross-sectional study. Thirty-four orbits with DON and 62 orbits with moderate-to-severe GO were included. Patients were older than 18 years, had never undergone surgical decompression, and did not receive glucocorticoid or radiotherapy within 6 months prior to participation. A standardized orbital MRI protocol was performed followed by ophthalmic and orthoptic examination within 4 weeks. Main outcome measures were: maximal radial diameter of the extraocular muscles (EOMs), signal intensity ratio of the EOMs (relative to the temporalis muscle), relative apparent diffusion coefficient (rADC) of the EOMs (relative to the temporalis muscle), apical crowding index, optic nerve stretching, clinical activity score (CAS), and duction scores. Data were analyzed using a multivariable generalized linear mixed model to identify measures associated with DON.

Results: The combination of medial rectus diameter and superior rectus rADC resulted in an area under the curve of 0.937 with a sensitivity of 85.2%, specificity of 93.2%, positive predictive value of 85.2%, and negative predictive value of 93.2% at a cutoff of 0.461. Apical crowding and optic nerve stretching were not independently associated with DON.

Conclusions: MRI with diffusion-weighted imaging sequences can be a useful adjunct in the diagnosis of DON. Medial rectus diameter and superior rectus rADC are independently associated with DON. Apical crowding and optic nerve stretching may be less important pathogenically.

Key messages: 1. What is already known? The diagnosis of dysthyroid optic neuropathy (DON) can be challenging due to the lack of validated diagnostic imaging features. 2.

What this study adds: a. Medial rectus diameter and superior rectus relative apparent diffusion coefficient measured by magnetic resonance imaging are independent predictors of DON. b. Apical crowding and optic nerve stretching are not independently associated with DON. 3. How this study might affect research: MRI with diffusion-weighted imaging sequences can be a useful adjunct in the diagnosis of DON.