Quantitative evaluation of synthetic MRI for assessing disease activity in thyroid eye disease

Abstract Objective Conventional magnetic resonance imaging (MRI) for thyroid eye disease (TED) is limited by long scan time, restricting quantitative assessment. Synthetic MRI techniques, such as magnetic resonance image compilation (MAGiC), allow for the simultaneous acquisition of T1, T2, and proton density (PD) maps in a single, rapid scan. This study evaluated MAGiC MRI as a quantitative imaging approach for TED disease activity assessment. Materials and methods A total of 113 TED patients (49 active, 64 inactive) underwent orbital MRI using a 3-T scanner with MAGiC sequences. Quantitative T1, T2, and PD maps were analyzed for correlations with the clinical activity score (CAS). Image quality of MAGiC_STIR and IDEAL IQ_WP sequences was assessed by two radiologists. Reproducibility was evaluated using intraclass correlation coefficients (ICCs). Logistic regression and receiver operating characteristic (ROC) analyses were performed to identify independent predictors and evaluate diagnostic performance. Results CAS showed significant correlations with multiple MAGiC parameters, especially MR_T1 and MR_PD ( p = 0.018, 0.040, respectively). MAGiC_STIR achieved comparable image quality to IDEAL IQ_WP (64.6% versus 67.3%). Quantitative parameters demonstrated excellent reproducibility (ICC = 0.755–0.985). IR_T1, SR_T1, and IR_PD independently predicted TED activity. The combined T1 + PD model achieved the highest accuracy (area under the ROC curve = 0.898; sensitivity = 0.816; specificity = 0.875), and adding T2 did not result in a significant improvement ( p = 1.000), as assessed by the DeLong test. Conclusion Synthetic MAGiC MRI provides fast and reproducible quantitative assessment of TED activity. The T1 + PD model offers an efficient, objective biomarker for disease staging and treatment planning. Key Points Question Can synthetic MRI provide rapid, reproducible quantitative biomarkers for accurately differentiating active and inactive TED within a single acquisition? Findings Synthetic MRI independently predicted disease activity, and the combined model achieved excellent diagnostic performance. Relevance Statement This study demonstrates that synthetic MRI can effectively assess the disease activity of TED, providing reproducible quantitative biomarkers to enhance disease staging and assist treatment decisions more efficiently.