Quantifying Orbital Segmentation in Thyroid Eye Disease- Fat to Muscle Ratio Correlates With Restrictive Strabismus.

Abstract

To evaluate the contribution and application of orbital fat-to-muscle ratio (FMR) in Thyroid Eye Disease related restrictive strabismus. Retrospective cross-sectional study. Adult patients (≥18 years) with a confirmed diagnosis of TED evaluated at a tertiary referral thyroid eye disease clinic between 2017 and 2025 who had high-resolution orbital CT imaging available for quantitative analysis. Orbital CT scans were analyzed using three-dimensional (3D) volumetric segmentation to quantify orbital fat and extraocular muscle (EOM) volumes. The FMR was calculated for each orbit. A simplified single-slice two-dimensional (2D) FMR was also measured for comparison. Clinical data included presence and magnitude of strabismus, restrictive ocular motility, Clinical Activity Score (CAS), thyroid-stimulating immunoglobulin (TSI) levels, and surgical history. Associations were assessed using correlation analyses, multivariable logistic regression, intraclass correlation coefficients (ICC), and receiver operating characteristic (ROC) analysis. Association of FMR with TED factors associated with disease severity, restrictive ocular motility and strabismus presence. Of 579 screened patients, 197 met inclusion criteria (mean age 54.9 ± 16.3 years; 69.7% female). Restrictive ocular motility was present in 39.1% and was associated with significantly lower FMR compared with nonrestrictive disease (0.69 ± 0.36 vs 0.91 ± 0.42; P < .001). Strabismus angle correlated inversely with FMR (r = -0.169; P = .030), as did the number of EOMs operated during strabismus surgery (r = -0.160; P = .027). FMR was negatively correlated with CAS (r = -0.186; P = .010) and TSI (r = -0.188; P = .047). In multivariable analysis, lower FMR remained an independent predictor of restrictive motility. An FMR cutoff of 0.66 best discriminated the presence of strabismus (sensitivity 0.59; specificity 0.72). Agreement between 2D and 3D measurements was good (ICC = 0.77). Quantitative FMR derived from CT segmentation is a clinically relevant imaging biomarker in TED that reflects a continuous spectrum of muscle and fat involvement. Lower FMR identifies patients at increased risk for restrictive motility, strabismus, and inflammatory activity. Three-dimensional FMR provides superior anatomical and clinical relevance compared with 2D measurements and may enhance risk stratification and longitudinal assessment in TED.