T₂‐Weighted MR Imaging‐Derived Radiomics for Pretreatment Determination of Therapeutic Response to Glucocorticoid in Patients With Thyroid‐Associated Ophthalmopathy: Comparison With Semiquantitative Evaluation

Abstract

BackgroundMR imaging has been applied to determine therapeutic response to glucocorticoid (GC) before treatment in thyroid‐associated ophthalmopathy (TAO), while the performance was still poor.PurposeTo investigate the value of T2‐weighted imaging (T2WI)‐derived radiomics for pretreatment determination of therapeutic response to GC in TAO patients, and compare its diagnostic performance with that of semiquantitative parameters.Study TypeRetrospective.PopulationA total of 110 patients (49 ± 12 years; male/female, n = 48/62; responsive/unresponsive, n = 62/48), divided into training (n = 78) and validation (n = 32) cohorts.Field Strength/Sequence3.0 T, T2‐weighted fast spin echo.AssessmentW.C. and H.H. (6 and 10 years of experience, respectively) performed the measurements. Maximum, mean, and minimum signal intensity ratios (SIRs) of extraocular muscle (EOM) bellies were collected to construct a semiquantitative imaging model. Radiomics features from volumes of interest covering EOM bellies were extracted and three machine learning‐based (logistic regression [LR]; decision tree [DT]; support vector machine [SVM]) models were built.Statistical TestsThe diagnostic performances of models were evaluated using receiver operating characteristic curve analyses, and compared using DeLong test. Two‐sided P < 0.05 was considered statistically significant.ResultsThe responsive group showed higher minimum signal intensity ratio (SIRmin) of EOMs than the unresponsive group (training: 1.46 ± 0.34 vs. 1.18 ± 0.39; validation: 1.44 ± 0.33 vs. 1.19 ± 0.20). In both cohorts, LR‐based radiomics model demonstrated good diagnostic performance (area under the curve [AUC] = 0.968, 0.916), followed by DT‐based (AUC = 0.933, 0.857) and SVM‐based models (AUC = 0.919, 0.855). All three radiomics models outperformed semiquantitative imaging model (SIRmin: AUC = 0.805) in training cohort. In validation cohort, only LR‐based radiomics model outperformed that of SIRmin (AUC = 0.745). The nomogram integrating LR‐based radiomics signature and disease duration further elevated the diagnostic performance in validation cohort (AUC: 0.952 vs. 0.916, P = 0.063).Data ConclusionT2WI‐derived radiomics of EOMs, together with disease duration, provides a promising noninvasive approach for determining therapeutic response before GC administration in TAO patients.Level of Evidence3Technical EfficacyStage 4

Magnetic Resonance Imaging in the Management of Thyroid Eye Disease: A Systematic Review

“…Four major MRI sequences were utilized across all included studies: T2-weighted imaging in 21 studies (87.5%), 8,14,23,28–31,37,38,40,43–53 T1-weighted imaging was used in 13 of the included studies (54.2%), 23,24,29,38,40,44,46–50,52,53 diffusion-weighted/diffusion-tensor imaging (DWI/DTI) in 5 studies (20.8%), 38,41,46,47,49 and STIR in 4 studies (16.7%). 12,14,28,48 The most common parameters quantified were signal intensity ratio (SIR) in 10 studies (41.7%), 8,12,14,23,24,28,29,37,40,53 T2RT in 8 studies (33.3%), 8,28,30,31,43,45,50,51 and apparent diffusion coefficient (ADC) in 5 studies (20.8%). 38,40,46,47,49 The following imaging findings were reported to correlate with TED activity across multiple studies: longer T2RT and higher SIR.…”

“…Twenty-four articles 8,12,14,23,24,[28][29][30][31]37,38,40,41,[43][44][45][46][47][48][49][50][51][52][53] were identified as meeting the inclusion criteria, and their full text was subsequently reviewed. The extracted information and outcomes are displayed in Table 1.…”

“…Four major MRI sequences were utilized across all included studies: T2-weighted imaging in 21 studies (87.5%), 8,14,23,[28][29][30][31]37,38,40,[43][44][45][46][47][48][49][50][51][52][53] T1-weighted imaging was used in 13 of the included studies (54.2%), 23,24,29,38,40,44,[46][47][48][49][50]52,53 diffusion-weighted/diffusion-tensor imaging (DWI/DTI) in 5 studies (20.8%), 38,41,46,47,49 and STIR in 4 studies (16.7%). 12,14,28,48 The most common parameters quantified were signal intensity ratio (SIR) in 10 studies (41.7%), 8,12,14,23,24,28,…”

“…Four included studies proposed that changes in SI on MRI can be used to monitor response to local or systemic administration of glucocorticoids. 29,30,37,40 The authors reported that MRI may also predict response to treatment with glucocorticoids and that CAS alone is not an exact surrogate for acute TED-associated myositis. However, glucocorticoids were not the only treatment modality assessed in the included studies.…”

Research progress and application of artificial intelligence in thyroid associated ophthalmopathy

“…Coronal T 2 -weighted MRI images with fat suppression can clearly show the cross-sectional morphology and radiomics features of EOMs. Hu et al (2022) developed three ML-based models to analyze the radiomics data of patients with TAO. In this retrospective study, 110 samples were selected, and GC-responsive ( n = 62) and unresponsive ( n = 48) cases were equally split into training and validation sets.…”