Utility of T2 mapping in the staging of thyroid-associated ophthalmopathy: efficiency of region of interest selection methods

Abstract

Background

Discriminating the stage of thyroid-associated ophthalmopathy (TAO) is crucial for the treatment strategy and prognosis prediction. Utility of conventional magnetic resonance imaging in the disease staging is limited.

Purpose

To investigate the performance of T2 mapping based on different region of interest (ROI) selection methods in the staging of TAO.

Material and Methods

Thirty-two patients with TAO were retrospectively enrolled. Two radiologists independently measured the T2 relaxation time (T2RT) of extraocular muscles using two different ROIs (hotspot [ROI_HS]: T2RT-hot; single-slice [ROI_SS]: T2RT-mean, T2RT-max, T2RT-min). Independent-samples t test, Wilcoxon signed rank test, Spearman correlation analysis, receiver operating characteristic (ROC) curves analyses, multiple ROC comparisons, and intra-class correlation coefficient (ICC) were used for statistical analyses.

Results

No significant difference was found in the measuring time between ROI_HS and ROI_SS methods (P = 0.066). T2RT-mean demonstrated the highest ICC for measurement, followed by T2RT-max and T2RT-min, and T2RT-hot showed the poorest reproducibility. Active TAOs showed significantly higher values for all the T2RTs than inactive mimics (all P < 0.001). Significant positive correlations were found between T2RTs and CAS (all P < 0.005). T2RT-hot and T2RT-max showed significantly higher areas under the curve than that of T2RT-mean (P = 0.013 and 0.024, respectively), while the difference between T2RT-hot and T2RT-max was not significant (P = 0.970).

Conclusion

The T2RTs derived from both ROI selection methods could be useful for the staging of TAO. The results of measuring time, reproducibility, and diagnostic performance suggest that T2RT-max would be the optimal indicator for staging.

Keywords

Thyroid-associated ophthalmopathy T2 mapping magnetic resonance imaging region of interest staging

Introduction

Thyroid-associated ophthalmopathy (TAO), also known as Graves’ orbitopathy, is considered to be the most common and important extra-thyroidal manifestation of hyperthyroidism (1). During its acute active phase, pathological inflammatory changes including mononuclear cell infiltration, edema, and fibroblast proliferation in orbital tissues can be present, and generally show good response to anti-inflammatory treatments (2). However, the contrary chronic and inactive phase is usually characterized by interstitial fibrosis with collagen deposition and fat infiltration, and surgical procedure is the only treatment option (3). Adequate and timely treatment in the active phase is critical for limiting both the severity of its chronic fibrotic sequelae and the progression of compressive sight-threatening optic neuropathy (4). Therefore, the accurate and prospective differentiation between the two phases is of highly therapeutic importance (5,6).

With the superiority of high soft-tissue resolution and no ionizing radiation, magnetic resonance imaging (MRI) has been increasingly applied in the evaluation of TAO. Thickness and volume measurement of orbital tissues can help to inspect the degree of involvement and establish the diagnosis of TAO (7,8). Previous studies have indicated that the semi-quantitative signal intensity ratio (SIR) of extraocular muscles (EOMs) and lacrimal glands measured with T2-weighted (T2W) imaging was correlated with the clinical activity and could be useful for the staging of TAO (8 –10). However, the overall assessing accuracy is still limited.

T2 mapping based on multi-echo spin-echo pulse sequences which quantifies native T2 relaxation time (T2RT) is a technique used for enabling visualization as well as quantification of water content and collagen structure, reflecting the histopathological changes of certain diseases (11). It has already been applied in the early assessment of cartilage degeneration and myocardial abnormalities (12 –15). Characterized by acute edema and chronic fibrosis, EOMs of TAOs shared the similar microstructural pathological changes which could also be detected by the T2RT derived from T2 mapping.

In addition, throughout the previous studies of SIR measurements in TAO, mainly two kinds of region of interest (ROI) selection methods were used: (i) ROI covering the “hotspot” within the EOMs; and (ii) ROI covering the single slice at the hotspot in which the EOMs were most inflamed (8,16,17). However, the hotspot ROI may lead to significant inter- and intra-observer variabilities (18). Single-slice ROI can largely overcome this problem, but it may take more time to complete the measurement process (19). To the best of our knowledge, no study has so far addressed the issue of the effect of different ROI selection methods on T2RT measurement and its value in the staging of TAO.

Therefore, the purpose of this study was to assess the performance of T2 mapping image based on the two different ROI selection methods in the staging of TAO.

Material and Methods

Patients

This study was approved by our institutional review board and the informed consent requirement was waived due to its retrospective nature. From January to July of 2018, a total of 43 consecutive patients who were clinically diagnosed with TAO based on Bartley’s criteria (20) underwent orbital MRI for pretreatment assessment in our hospital. Eleven patients were excluded based on the following criteria: (i) T2 mapping was not scanned (n = 7); (ii) image quality was not adequate for further analysis (n = 2); (iii) patient who had undergone radiotherapy or surgical decompression (n = 1); and (iv) patient who had other orbital pathologies (n = 1). Finally, 32 patients (8 men, 24 women; mean age = 46.1 ± 15.0 years) were enrolled in this study.

The disease activity was assessed for each unit of eye according to the modified 7‑point formulation of Mourits’ clinical activity score (CAS) (5), which includes: (i) spontaneous retrobulbar pain; (ii) pain on attempted up or down gaze; (iii) redness of the eyelids; (iv) redness of the conjunctiva; (v) swelling of the eyelids; (vi) inﬂammation of the caruncle and/or plica; and (vii) conjunctival edema. Each item presented was given 1 point and CAS is the sum of total given points. Eyes with a CAS of ≥3 were enrolled in the active TAOs group. Otherwise, they were enrolled in the inactive TAOs group. Finally, a total of 34 eyes were defined as active with a mean CAS of 4 ± 1 and the other 30 eyes were defined as inactive with a mean CAS of 1 ± 1.

Image acquisition

All patients were examined by using a 3.0-T MRI system (MAGNETOM Skyra; Siemens Healthcare, Erlangen, Germany) with a 20-channel head coil. Conventional imaging protocols included axial T1-weighted imaging (repetition time [TR]/echo time [TE] = 635/6.7 ms), axial T2W imaging (TR/TE = 4000/117 ms) with fat suppression, and coronal T2W imaging (TR/TE = 4000/75 ms) with fat suppression. Coronal T2 mapping was obtained by using a modified multi-echo spin-echo sequence that implemented both model-based accelerated relaxometry by iterative nonlinear inversion (MARTINI) (21) and generalized autocalibrating partial parallel acquisition (GRAPPA) (22). This technique was named GRAPPATINI (23). The parameters were as follows: TR/TE = 2020/12–192 ms; delta TE = 12 ms; echoes = 16; field of view = 20 cm; slice thickness = 3.0 mm; slice gap = 0.6 mm; slices = 20; matrix = 320 × 320; GRAPPA acceleration factor = 2; MARTINI acceleration factor = 5. The scan duration was 4 min 40 s. During MR scanning, participants were instructed to lie still in supine position and look at a fixed site with both eyes closed in order to reduce eye movement.

Image analysis

All imaging data were analyzed with Siemens workstation for measuring T2RT of EOMs (superior, inferior, medial, and lateral recti) in each unit of eye. Before quantitative measurement, conventional T2W images were overviewed initially to evaluate the area with relatively higher signal intensity. For ROI placement, we applied two methods: hotspot ROI (ROI_HS) and single-slice ROI (ROI_SS). ROI_HS was defined as a small circular ROI placed in the area of the most inflamed muscle with the highest degree of T2RT observed by naked eye. ROI_SS was placed in the whole section of the most inflamed muscle at the hotspot, which was generally the second or third section behind the eyeball. The borders and surrounding fat tissues were excluded by taking the corresponding T2W images as reference. A schematic illustration of ROI selection methods was shown in Fig. 1. Once the ROIs were determined, the mean values of T2RT in ROI_HS (T2RT-hot) as well as the mean, maximum, and minimum values of T2RT in ROI_SS (T2RT-mean, T2RT-max, and T2RT-min) were automatically generated. The time required for measurements with the two different ROI methods was also recorded.

Fig. 1.

Methods for measuring T2RT values of EOMs. Coronal fat-suppressed T2-weighted image (a) in a 63-year-old woman with active-phase TAO showed the enlarged EOMs and increased signal intensity, especially in the bilateral inferior rectus. For the quantitative measurement on T2 mapping image (b), taking the left eye as an example, a small circular hotspot ROI (ROI_HS) was placed in the area of inferior rectus featuring the highest degree of T2RT observed by naked eye (red circle). Concurrently, a single-slice ROI (ROI_SS) was placed in the whole section at the hotspot (yellow circle) with exclusion of the borders and surrounding fat tissues. EOM, extraocular muscle; TAO, thyroid-associated ophthalmopathy; T2RT, T2 relaxation time.

Two radiologists (observer 1: with eight years of experience in head and neck radiology; observer 2: with five years of experience in head and neck radiology) independently drew the ROIs manually. They were blinded to the clinical information, pathological results and study design. The measurement results of these two observers were used to assess the inter-observer agreement, and the measurement was repeated by observer 1 with a washout period of at least one month, in order to evaluate the intra-observer reproducibility.

Statistical analysis

All numeric data were reported as mean ± SD, and the Kolmogorov–Smirnov test was used for normality distribution analysis. The time required for outlining the two different ROIs was compared by using Wilcoxon signed rank test, and T2RT values (T2RT-hot, T2RT-mean, T2RT-max, and T2RT-min) between the two groups were compared by using independent sample’s t test. Spearman correlation analysis was performed to evaluate the correlation between T2RTs and CAS. Receiver operating characteristic (ROC) curves were used to determine the diagnostic performance of each significant T2RT value in differentiating active from inactive TAOs. The sensitivity and specificity were calculated at a cut-off point that maximized the value of Youden index (Youden index = sensitivity + specificity – 1). Multiple ROC curves comparisons according to DeLong et al. (24) were used to compare the diagnostic performance of T2RT values. Inter-observer and intra-observer agreements of T2RT measurements based on the two different ROI selection methods were assessed by using the two-way intraclass correlation coefficients (ICC). The ICC values were in the range of 0–1.00, and values closer to 1.00 represent better reproducibility. They were defined as follows: (<0.40 = poor, 0.40–0.75 = moderate, 0.76–0.90 = good, ≥0.91 = excellent). All statistical analyses were carried out by using two statistical packages (SPSS Version 22.0; SPSS Inc., Chicago, IL, USA and MedCalc version 18.2.1; MedCalc, Ostend, Belgium). A two-sided P value <0.05 was considered statistically significant.

Results

In active TAOs, the most commonly chosen EOM as the most inflamed muscle was inferior rectus (18/34), followed by medial rectus (12/34), and superior rectus (4/34). The average time required for ROI_HS and ROI_SS measurements were 10.5 ± 2.1 s and 11.1 ± 1.9 s, respectively. No significant difference was found in the time required for these two different ROI methods (P = 0.066). Good to excellent reproducibility was obtained during the measurements of all the T2RTs. Both inter-observer and intra-observer ICCs were highest for T2RT-mean, followed by T2RT-max and T2RT-min, and lowest for T2RT-hot (Table 1).

Table 1.

Inter-observer and intra-observer ICCs for measurements of T2RT values with the two different ROI selection methods.

ICC	Inter-observer	Intra-observer
T2RT-hot	0.785 (0.646–0.869)	0.878 (0.799–0.926)
T2RT-mean	0.847 (0.749–0.907)	0.942 (0.904–0.965)
T2RT-max	0.842 (0.740–0.904)	0.941 (0.903–0.964)
T2RT-min	0.836 (0.730–0.900)	0.937 (0.897–0.962)

Values in parentheses are 95% confidence intervals.

ICC, intraclass correlation coefficient; ROI, region of interest; T2RT, T2 relaxation time.

Table 2 summarized the comparisons of T2RT values in active and inactive TAO groups. Active TAOs demonstrated significantly higher values for all the T2RTs than inactive TAOs (T2RT-hot = 133.0 ± 33.0 vs. 86.8 ± 21.7; T2RT-mean = 104.0 ± 26.1 vs. 77.9 ± 18.6; T2RT-max = 142.0 ± 33.1 vs. 96.4 ± 20.4; T2RT-min = 72.1 ± 21.6 vs. 53.9 ± 12.7; all P < 0.001) (Fig. 2). Spearman correlation results demonstrated that there were significant positive correlations between all the T2RTs and CAS (T2RT-hot, r = +0.555; T2RT-mean, r = +0.444; T2RT-max, r = +0.560; T2RT-min, r = +0.352; all P < 0.005).

Table 2.

Comparisons of T2RT values between active and inactive TAO groups

T2RT (ms)	Active TAOs (n = 34)	Inactive TAOs (n = 30)	P value
T2RT-hot	133.0 ± 33.0	86.8 ± 21.7	<0.001
T2RT-mean	104.0 ± 26.1	77.9 ± 18.6	<0.001
T2RT-max	142.0 ± 33.1	96.4 ± 20.4	<0.001
T2RT-min	72.1 ± 21.6	53.9 ± 12.7	<0.001

Values are given as mean ± SD.

TAO, thyroid-associated ophthalmopathy; T2RT, T2 relaxation time.

Fig. 2.

Bar chart for T2RT values of active (black) and inactive TAOs (gray) based on the two ROI selection methods. The error bars refer to SD. * indicates the difference reached significant level. ROI, region of interest; TAO, thyroid-associated ophthalmopathy; T2RT, T2 relaxation time.

ROC curves analyses results indicated that T2RT-hot demonstrated the highest area under curve (AUC) of 0.869 (cut-off = 115.4 ms; sensitivity = 76.5%; specificity = 93.3%), followed by T2RT-max (AUC = 0.868; cut-off = 121.5 ms; sensitivity = 79.4%; specificity = 93.3%), T2RT-mean (AUC = 0.793; cut-off = 82.0 ms; sensitivity = 82.4%; specificity = 73.3%), and T2RT-min (AUC = 0.731; cut-off = 70.5 ms; sensitivity = 50.0%; specificity = 93.3%). T2RT-hot and T2RT-max showed significantly higher AUCs than that of T2RT-mean (P = 0.013 and 0.024, respectively), while no significant differences were found between T2RT-hot and T2RT-max (P = 0.970). The results of ROC analyses using the above T2RTs to discriminate active from inactive TAOs were shown in Fig. 3.

Fig. 3.

ROC analyses using T2RT-hot (a), T2RT-mean (b), T2RT-max (c) and T2RT-min (d) to differentiate active from inactive TAOs. ROC, receiver operating characteristic; TAO, thyroid-associated ophthalmopathy; T2RT, T2 relaxation time.

Discussion

Discriminating the stage of TAO is of great significance for its treatment strategy and prognosis. Currently CAS is the most widely used metric for staging TAO, but it relies on patients’ subjective symptoms and signs of the anterior visible orbit, which cannot reflect the involvement conditions of retro-ocular tissue (6). In this study, we sought to differentiate active from inactive TAO based on T2 mapping, which is a technique used to construct parametric T2RT image generated by measuring the MRI signal intensity at various echo times (15,25). T2RT is a tissue-specific time constant describing the decay of transverse magnetization of tissues (25); therefore, it is independent of machine and scanning parameters and does not need to be standardized. This feature could improve the accuracy and repeatability of the measurements. Additionally, the uniqueness of this study was the utility of different ROI selection methods in T2RT measurement, taking into consideration of measuring time, reproducibility, and diagnostic performance.

In this study, we obtained T2RT values of EOMs from active and inactive TAOs based on ROI_HS and ROI_SS. The average measuring time required for T2RT-mean, T2RT-max, and T2RT-min with ROI_SS did not differ significantly with that for T2RT-hot with ROI_HS, despite a little longer tendency towards ROI_SS. All derived T2RT values in active TAOs were significantly higher than those in inactive TAOs. T2RT-hot and T2RT-max showed significantly higher AUCs than T2RT-mean, while no significant difference was found between T2RT-hot and T2RT-max. In addition, T2RT-mean demonstrated the highest ICC, followed by T2RT-max and T2RT-min; T2RT-hot showed the poorest reproducibility. Combining the results of reproducibility, measuring time, and diagnostic performance, we suggested that T2RT-max would be the optimal indicator for differentiating the clinical stage of TAO.

The present study showed that all the T2RT values were higher in active TAOs as compared to those in inactive TAOs. Active phase is characterized histopathologically by mononuclear cell infiltration, fibroblast proliferation, and edema in EOMs. In contrast, inactive phase is characterized by interstitial fibrosis and collagen deposition (2,3). Given that T2RT value could increase in edematous muscle and decrease in fibrotic tissue, it would not be surprising that active TAOs had prolonged T2RT of EOMs than that of inactive ones (26). In addition, T2RT-max and T2RT-hot demonstrated better diagnostic performance than other T2RT-derived indicators, while the variation of AUC between T2RT-max and T2RT-hot was of no significance. In our opinion, T2RT-max and T2RT-hot reflect the most inflamed portion of EOMs and could better represent the disease process of TAO.

Each ROI selection method has its own merits and limitations. Smaller ROI such as ROI_HS was considered to be a convenient and time-saving option, but the variability during measurement was the main drawback (18). Larger single-slice ROI could break the above limitation; however, it might take more time (19). Nonetheless, in the present study, the average measuring time required for ROI_SS did not differ significantly with that for ROI_HS, although there was a little longer tendency towards ROI_SS. According to our experience during the measurement, the placement of ROI_HS involving the highest degree of T2RT within the most inflamed EOMs by naked eye would not always be decisive, especially when there were a few scattered similar intensity portions. This might be responsible for the non-significant measuring time between ROI_HS and ROI_SS. Thus, taking all the above factors into consideration, the T2RT-max derived from ROI_SS with higher reproducibility would be the most appropriate parameter for differentiating the clinical activity of TAO.

SIR of EOMs or lacrimal glands derived from T2W imaging was thought to be a relatively objective and reproducible indicator and was helpful for the staging of TAO (8 –10). However, as Das et al. (26) pointed out, it suffered from the need to choose an appropriate but arbitrary denominator, such as signal intensity from temporalis muscle or normal appearing white matter. Moreover, it was a semi-quantitative measurement with the potential risk of reduced sensitivity and increased variability. Recently, some studies utilized diffusion-weighted imaging (DWI) for the evaluation of TAO (4,6,27 –29). They found that the apparent diffusion coefficient of EOMs or lacrimal glands could also be useful to evaluate disease activity (4,27,28). However, the relatively lower resolution and air-bone susceptibility artifacts in orbital DWI remained to be troublesome issues during image processing, especially in the evaluation of inferior rectus. In addition, the overall available staging efficacy of previous studies is poorer than ours (AUC = 0.711–0.800 vs. 0.868) (10,27). Thus, our current usage of the absolutely quantitative nature of T2RT with high imaging quality and convenient image processing could be adopted as a highly promising method in improving staging accuracy and reliability of TAO, and furthermore in guiding treatment strategy and assessing response.

The present study had several limitations. First, the number of participants was relatively small. A further larger sample study would strengthen the reliability of our findings and verify our results. Second, the oblique angle of lateral rectus to the coronal plane might lead to certain artifact, although no lateral rectus was included in this study due to its lowest involvement rate. Lastly, further work to integrate other MRI modalities (30) into a single multi-parametric model might further improve the staging accuracy.

In conclusion, the present study demonstrated that the T2RT values derived from both ROI selection methods of ROI_HS and ROI_SS were useful for the staging of TAO. Taking reproducibility, measuring time, and diagnostic performance together, T2RT-max would be the optimal indicator for staging of TAO.

Footnotes

Acknowledgements

The authors thank the members of the Siemens group, Tobias Kober, Yi-Cheng Hsu, and Yi Sun, for their advice and expertise in T2 mapping technology.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received the following financial support for the research, authorship, and/or publication of this article: This work was supported by National Natural Science Foundation of China (NSFC) (81801659 to Hao Hu).

ORCID iDs

Tao Yang

Fei-Yun Wu

References

Şahlı

Gündüz

Thyroid-associated Ophthalmopathy. Turk J Ophthalmol 2017; 47:94–105.

Hiromatsu

Eguchi

Tani

, et al. Graves’ ophthalmopathy: epidemiology and natural history. Intern Med 2014; 53:353–360.

Gould

Roth

Soparkar

CN.

The diagnosis and treatment of thyroid-associated ophthalmopathy. Aesthetic Plast Surg 2012; 36:638–648.

Lingam

Mundada

Lee

Novel use of non-echo-planar diffusion weighted MRI in monitoring disease activity and treatment response in active Grave’s orbitopathy: An initial observational cohort study. Orbit 2018; 37:325–330.

Bartalena

Baldeschi

Boboridis

, et al. The 2016 European Thyroid Association/European Group on Graves’ Orbitopathy Guidelines for the Management of Graves’ Orbitopathy. Eur Thyroid J 2016; 5:9–26.

Politi

Godi

Cammarata

, et al. Magnetic resonance imaging with diffusion-weighted imaging in the evaluation of thyroid-associated orbitopathy: getting below the tip of the iceberg. Eur Radiol 2014; 24:1118–1126.

Aydin

Güven

Sencer

, et al. A new MRI method for the quantitative evaluation of extraocular muscle size in thyroid ophthalmopathy. Neuroradiology 2003; 45:184–187.

Mayer

Fox

Herdman

, et al. Signal intensity, clinical activity and cross-sectional areas on MRI scans in thyroid eye disease. Eur J Radiol 2005; 56:20–24.

Higashiyama

Nishida

Morino

, et al. Use of MRI signal intensity of extraocular muscles to evaluate methylprednisolone pulse therapy in thyroid-associated ophthalmopathy. Jpn J Ophthalmol 2015; 59:124–130.

10.

Hu H, Xu XQ, Wu FY, et al. Diagnosis and stage of Graves’ ophthalmopathy: Efficacy of quantitative measurements of the lacrimal gland based on 3-T magnetic resonance imaging. Exp Ther Med 2016;12:725–729.

11.

Lee

Song

, et al. Quantitative T2 mapping of knee cartilage: comparison between the synthetic MR imaging and the CPMG sequence. Magn Reson Med Sci 2018; 17:344–349.

12.

Kester

Carpenter

, et al. T1ρ/T2 mapping and histopathology of degenerative cartilage in advanced knee osteoarthritis. World J Orthop 2017; 8:350–356.

13.

Peng

Sperling

Biochemical magnetic resonance imaging of knee articular cartilage: T1rho and T2 mapping as cartilage degeneration biomarkers. Ann N Y Acad Sci 2016; 1383:34–42.

14.

Kim

Hong

, et al. Myocardial T1 and T2 mapping: techniques and clinical applications. Korean J Radiol 2017; 18:113–131.

15.

Lota

Gatehouse

Mohiaddin

RH.

T2 mapping and T2* imaging in heart failure. Heart Fail Rev 2017; 22:431–440.

16.

Rodríguez-González

Pérez-Rico

López-Para Giménez

, et al. Short-tau inversion-recovery (STIR) sequence magnetic resonance imaging evaluation of orbital structures in Graves’ orbitopathy. Arch Soc Esp Oftalmol 2011; 86:351–357.

17.

Tortora

Cirillo

Ferrara

, et al. Disease activity in Graves’ ophthalmopathy: diagnosis with orbital MR imaging and correlation with clinical score. Neuroradiol J 2013; 26:555–564.

18.

Qian W, Xu XQ, Hu H, et al. Dynamic contrast-enhanced MRI in orbital lymphoproliferative disorders: Effects of region of interest selection methods on time efficiency, measurement reproducibility, and diagnostic ability. J Magn Reson Imaging 2018;47:1298–1305.

19.

Xu XQ, Hu H, Su GY, et al. Diffusion weighted imaging for differentiating benign from malignant orbital tumors: diagnostic performance of the apparent diffusion coefficient based on region of interest selection method. Korean J Radiol 2016;17:650–656.

20.

Bartley

Gorman

CA.

Diagnostic criteria for Graves’ ophthalmopathy. Am J Ophthalmol 1995; 119:792–795.

21.

Sumpf

Uecker

Boretius

, et al. Model-based nonlinear inverse reconstruction for T2 mapping using highly undersampled spin-echo MRI. J Magn Reson Imaging 2011; 34:420–428.

22.

Griswold

Jakob

Heidemann

, et al. Generalized autocalibrating partially parallel acquisitions (GRAPPA). Magn Reson Med 2002; 47:1202–1210.

23.

Hilbert

Sumpf

Weiland

, et al. Accelerated T2 mapping combining parallel MRI and model-based reconstruction: GRAPPATINI. J Magn Reson Imaging 2018; 48:359–368.

24.

DeLong

Clarke-Pearson

DL.

Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988; 44:837–845.

25.

Glaser

New techniques for cartilage imaging: T2 relaxation time and diffusion-weighted MR imaging. Radiol Clin North Am 2005; 43:641–653.

26.

Das

Roos

JCP

Patterson

AJ.

T2-relaxation mapping and fat fraction assessment to objectively quantify clinical activity in thyroid eye disease: an initial feasibility study. Eye (Lond) 2019; 33:235–243.

27.

Razek

El-Hadidy

Moawad

, et al. Assessment of lacrimal glands in thyroid eye disease with diffusion-weighted magnetic resonance imaging. Pol J Radiol 2019; 84:e142–e146.

28.

Kilicarslan

Alkan

Ilhan

, et al. Graves’ ophthalmopathy: the role of diffusion-weighted imaging in detecting involvement of extraocular muscles in early period of disease. Br J Radiol 2015; 88:20140677.

29.

Abdel Razek

El-Hadidy

Moawad

, et al. Performance of apparent diffusion coefficient of medial and lateral rectus muscles in Graves’ orbitopathy. Neuroradiol J 2017; 30:230–234.

30.

Han

Seo

Lee

, et al. Fractional anisotropy and diffusivity changes in thyroid-associated orbitopathy. Neuroradiology 2016; 58:1189–1196.