T2-relaxation mapping and fat fraction assessment to objectively quantify clinical activity in Thyroid Eye Disease: an initial feasibility study Tilak Das PhD FRCR,1 Jonathan C P Roos PhD FRCOphth,2,3 Andrew J Patterson PhD,4 Martin J Graves PhD4 & Rachna Murthy BSc FRCOphth2,3 1Dept. of Radiology, Cambridge University Hospitals NHS Foundation Trust, Cambridge, UK 2 Dept. of Ophthalmology, Cambridge University Hospitals NHS Foundation Trust, Cambridge, UK  3Department of Clinical Neurosciences, University of Cambridge, Cambridge UK 4Dept. of Radiology, University of Cambridge, Cambridge, UK Corresponding author: Dr Tilak Das BM BChir PhD MRCP FRCR Consultant Neuroradiologist Department of Radiology Addenbrooke’s Hospital, Cambridge, CB2 0QQ, UK Tel: ++441223245151 Email: tilak.das@addenbrookes.nhs.uk Declaration of interest statement: All authors are doctors who assess patients with Thyroid Eye disease either clinically or radiologically. There are no other competing interests and a unified declaration form is available on request. Declaration of Funding: The project was supported by the Addenbrooke’s Charitable Trust and the NIHR comprehensive Biomedical Research Centre award to Cambridge University Hospitals NHS Foundation Trust in partnership with the University of Cambridge. Contributions Statement: The team conceived of the idea for this study and all authors have contributed to the data gathering, analysis and writing. Ethics Statement: This study was approved by the local Research Ethics Committee. License: The corresponding author has the right to grant on behalf of all authors and does grant on behalf of all authors, an exclusive licence on a worldwide basis to permit this article to be published. Guarantor: Dr Das serves as guarantor of this work. It is an honest, accurate, and transparent account of the study being reported; no important aspects of the study have been omitted. Key words: MRI, T2 relaxation time, Orbitopathy, Fat Fraction, TSH Receptor antibody, Clinical Activity Score, Graves’ disease, Thyroid Eye Disease Abstract Background: Imaging in Thyroid Eye Disease (TED) is used to exclude other diagnoses, assess for apical crowding and plan surgery. But to quantify TED activity objectively, subjective clinical scoring assessments remain the norm. Magnetic Resonance Imaging (MRI) T2-relaxation times correlate with extra-ocular muscle (EOM) inflammation but are confounded by signal from fat. Aim: We investigated whether T2-relaxation mapping in combination with fat fraction (FF) measurements could quantify disease activity in EOMs objectively. Methods: Sixty-two TED patients and six controls were enrolled for coronal STIR, T2 multi-echo fast-spin echo and multi-echo fast-gradient echo MRI of the orbits. STIR signal intensity ratios (SIR), T2-relaxation times and percentage FF were derived for inferior, lateral, superior and medial recti bilaterally. Twelve patients were re-scanned following immunosuppressive treatment. Results: There was a positive correlation for all subjects between T2 and SIR (p< 0.001) but only mean T2 differed significantly between patients and controls (p<0.001). We measured FF in EOMs for the first time and found it greater in TED (p<0.001). There was also a significant reduction in mean T2 after treatment, with a corresponding reduction in the clinical activity score (CAS) in almost all patients. Conclusions: T2-relaxation times differentiate between normal and inflamed extra-ocular muscles and are responsive to treatment. Combined, uniquely, with fat fraction measurement in EOMs, an objective, quantitative marker of inflammation in TED-affected muscles could be derived. T2 relaxation times mirrored improvements in CAS after treatment, occasionally preceding them. Rarely, they diverged, suggesting limitations in the CAS as a disease burden marker. Main Body Introduction Assessing TED & the role of imaging A wide array of imaging techniques, summarised in Table 1, have been proposed to aid in the diagnosis and management of TED.[endnoteRef:2] Whilst most TED patients may not require any imaging, it can be necessary for assessment of orbital apex crowding affecting the optic nerve; planning decompression surgery; and to exclude other orbital diseases. This is particularly so with asymmetrical presentations when myositis, lymphoma, carotico-cavernous fistulae, varices, IgG4 disease, meningiomas and other malignancies should be considered.[endnoteRef:3] [2: Pitz S, Muller-Forell W. Orbital Imaging. In Wiersinga WM, Kahaly GJ (eds): Graves’ Orbitopathy: A Multidisciplinary Approach – Questions and Answers. Basel, Karger, 2017 pp61-73.] [3: Konuk O, Anagnostis P. Diagnosis and differential of Graves’ Orbitopathy. In Wiersinga WM, Kahaly GJ (eds): Graves’ Orbitopathy: A Multidisciplinary Approach – Questions and Answers. Basel, Karger, 2017 pp74-92.] A further and emerging role for orbital imaging is to help guide and determine response to treatment. Several principally MRI-based methods have been developed in an effort to quantify disease activity objectively (see Table 2). However, they have drawbacks, including confounding by surrounding tissue effects, the use of semi-quantitative relative values with a lack of day-to-day reproducibility and complexity of sequences required to derive absolute values. Given these limitations, clinicians still rely principally on subjective scoring systems to quantify disease severity and activity clinically; a widely used tool is the clinical activity score (CAS),[endnoteRef:4] but it has been criticised as too binary, subjective and lacking in sensitivity.[endnoteRef:5] It is also poorly designed for monitoring of improvement;[endnoteRef:6] even marked amelioration of any component of the CAS will not improve the actual score unless the feature resolves completely.[endnoteRef:7] Not surprisingly, more complex scoring systems have been introduced such as the VISA,[endnoteRef:8] modified NOSPECS[endnoteRef:9] and EUGOGO[endnoteRef:10] scoring tools. But fundamentally they remain subjective and are cumbersome to administer.[endnoteRef:11] [4: Mourits MP, Koornneef L, Wiersinga WM, Prummel MF, Berghout A, van der Gaag R. Clinical criteria for the assessment of disease activity in Graves' ophthalmopathy: a novel approach. Br J Ophthalmol. 1989;73(8):639-44.] [5: Gorman CA. The measuring of changes in Graves’ ophthalmopathy. Thyroid. 1998;8:539-543.] [6: Dickinson AJ, Hintschich C. Clinical Manifestations. In Wiersinga WM, Kahaly GJ (eds): Graves’ Orbitopathy: A Multidisciplinary Approach – Questions and Answers. Basel, Karger, 2017 pp1-25.] [7: Dickinson AJ, Perros P. Controversies in the clinical evaluation of active thyroid-associated orbitopathy: use of a detailed protocol with comparative photographs for objective assessment. Clin Endocrinol (Oxf). 2001;55(3):283-303.] [8: Dolman PJ. Grading Severity and Activity in Thyroid Eye Disease. Ophthalmic Plast Reconstr Surg. 2018;34(4S Suppl 1):S34-S40. ] [9: Werner SC. Modification of the classification of the eye changes of Graves' disease: recommendations of the Ad Hoc Committee of the American Thyroid Association. J Clin Endocrinol Metab. 1977;44(1):203-4.] [10: See www.eugogo.eu] [11: Dolman PJ. Grading Severity and Activity in Thyroid Eye Disease. Ophthalmic Plast Reconstr Surg. 2018;34(4S Suppl 1):S34-S40. ] Other proposed methods for assessing disease progression include serial binocular single vision mapping, measurements of proptosis[endnoteRef:12] or Hess charts.[endnoteRef:13] But these may not reflect activity or severity dynamically as the active inflammatory phase of TED can be followed by a chronic scarring phase with a static effect on ocular motility. Serial TSH receptor antibody levels have been postulated to provide a further objective and quantitative means by which to assess thyroid eye disease activity but have not yet been widely adopted.[endnoteRef:14] [12: Smith TJ, Kahaly GJ, Ezra DG, Fleming JC, Dailey RA, Tang RA, Harris GJ, Antonelli A, Salvi M, Goldberg RA, Gigantelli JW, Couch SM, Shriver EM, Hayek BR, Hink EM, Woodward RM, Gabriel K, Magni G, Douglas RS. Teprotumumab for Thyroid-Associated Ophthalmopathy. N Engl J Med. 2017;376(18):1748-1761. ] [13: Roos JC, Murthy R. Use of Sirolimus for treatment of refractory Thyroid Eye Disease. 2019 (submitted)] [14: Roos JC, Paulpandian V, Murthy R. Serial TSH-Receptor Antibody levels to guide the management of Thyroid Eye Disease: the impact of smoking, immunosuppression, radio-iodine, and thyroidectomy. Eye (Lond). 2019. (in press).] Proposed Quantification of TED by modified MRI protocol There has been great interest in using MRI to quantify disease activity, severity and response to treatment in TED. Some of these are summarized in Table 2. However, there is inherent variability in MRI measurement, influenced by technical factors such as variability in scanners; sequences used; and physiological factors such as changes over time (from diurnal to age-related) or change in disease activity. Signal intensities seen on images are not usually directly proportional to specific tissue properties such as density or molecular composition. This variability in measured MR signal values makes them unsuitable for quantifiable comparison between scans – even within an individual – unless sequences are specifically designed for quantification. To get around this, groups have used signal intensity ratios from standard clinical images such as those generated from STIR (Short Tau Inversion Recovery) sequences. These are T2-weighted fat-suppressed images in which higher intensities reflect higher water content. EOM intensity is compared with other unaffected muscles or even non-muscular tissues,[endnoteRef:15],[endnoteRef:16],[endnoteRef:17] but such ratios are less sensitive. Mathematically, small variabilities can be amplified to significantly alter these ratios. [15: Tortora F, Prudente M, Cirillo M, et al. Diagnostic accuracy of short-time inversion recovery sequence in Graves’ ophthalmopathy before and after prednisone treatment. Neuroradiology 2014;56:353–61.] [16: Pajak M, Loba P, Wieczorek-Pastusiak J, et al. Signal intensity and T2 time of extraocular muscles in assessment of their physiological status in MR imaging in healthy subjects. Polish Journal of Radiology 2012;77:7.] [17: Mayer E, Herdman G, Burnett C, et al. Serial STIR magnetic resonance imaging correlates with clinical score of activity in thyroid eye disease. Eye 2001;15:313–8.] Though the MR signal can be variable, the decay rate (or relaxation time) of the signal is not, being a physical property of the tissue that can be expressed numerically. T2 (or transverse) relaxation times depend strongly on water content of tissue and longer T2 relaxation times correlate with greater extra-ocular muscle inflammation.[endnoteRef:18],[endnoteRef:19] However, signal from fat can confound these quantitative measurements as these also represent a variable combination of water and fat. [18: Mancini L, Rajendram R, Uddin J, Lee RW, Rose GE , Yousry T, Miszkiel K,Thornton JS. Extra-orbital muscle T2 relaxation time and clinical activity in thyroid eye disease. Paper presented at International Society for Magnetic Resonance in Medicine. https://cds.ismrm.org/protected/10MProceedings/files/876_3843.pdf (accessed 28/10/18)] [19: Ohnishi T, Noguchi S, Murakami N, et al. Extraocular muscles in Graves ophthalmopathy: usefulness of T2 relaxation time measurements. Radiology 1994;190:857–62.] Fat fraction mapping is a method that takes advantage of the slightly different resonant frequencies of water and fat to generate images that separate the contribution of water and fat to the signal, allowing quantitative measurements of the fat fraction within a tissue.[endnoteRef:20],[endnoteRef:21] [20: Reeder SB, Sirlin CB. Quantification of liver fat with magnetic resonance imaging. Magn Reson Imaging Clin N Am 2010;18:337–57–ix.] [21: Hussain HK, Chenevert TL, Londy FJ, et al. Hepatic Fat Fraction: MR Imaging for Quantitative Measurement and Display—Early Experience 1. Radiology 2005;237:1048–55.] In this study, we sought to compare T2 relaxation mapping to STIR Signal Intensity Ratio (SIR) and, for the first time, to calculate fat fractions for extra-ocular muscles. The combination of fat fraction data with T2 relaxation times may refine signal quantification and could then represent a robust quantitative measure of disease activity. Methods Between 2014 and 2016, sixty-two patients with TED and six healthy controls were recruited for this study which had received ethical approval from our hospital internal ethics review board. All patients received a full ophthalmological work-up, including determination of CAS and blood tests. MRI of the orbits was performed on a 1.5T MRI scanner (450, GE Healthcare, Waukesha, WI, USA) including coronal STIR, T2-weighted multi-echo fast spin echo and multi-echo fast gradient echo sequences[endnoteRef:22] (Fig. 1). The sequences for T2 relaxation and FF mapping took an additional ten minutes in the scanner to acquire. [22: Hussain HK, Chenevert TL, Londy FJ, et al. Hepatic Fat Fraction: MR Imaging for Quantitative Measurement and Display—Early Experience 1. Radiology 2005;237:1048–55.] Twelve patients who required active management with immunosuppression with cyclosporine and prednisolone were re-scanned following their treatment. STIR signal intensity ratios (SIR), T2 relaxation times and percentage fat fraction (FF) were measured by a neuroradiologist with more than seven years’ experience, blinded to the clinical data, from regions of interest in the muscle belly of inferior, lateral, superior and medial rectus muscles bilaterally in all individuals. Thereafter Pearson’s correlation between T2 and SIR was calculated for each individual muscle. Unpaired T-tests were performed to investigate if T2, SIR and FF were different between the thyroid eye disease patients and normal. Paired T-tests were performed to investigate if T2 and SIR were different in patients receiving immunosuppressive treatment. For FF comparisons, to meet the normality assumptions for standard parametric tests, the distributions were log transformed. Statistical analysis was performed in R (The R foundation for statistical computing, Vienna, Austria). Results We initially investigated if the T2 relaxation data correlated with SIR. As shown in figure 2a, for both TED and normal groups, there was a positive correlation between T2 and SIR (r = 0.43, p < 0.001). Mean T2 for each rectus muscle category differed significantly between the TED and normal groups (combined mean T2 TED: 93.92+/-12.80ms [95%CI 92.74-95.11]; normal: 75.30+/-7.96ms [95%CI 73.20-78.25], p <0.001; Fig. 2b). However, there was no significant difference in mean SIR (combined mean SIR TED: 1.71+/-0.32 [95%CI 1.68-1.74]; normal: 1.74+/- 0.25 [95%CI 1.66-1.82], p=0.52; Fig 2c). Overall mean fat fraction was significantly greater in TED patients than in healthy volunteers (mean FF TED: 13.98+/-7.0% [95%CI 13.33-14.64], normal: 9.27+/-3.14% [95%CI 8.38-10.15], p<0.001). There was substantial variability in the mean fat fraction for each muscle in the patient group (Fig. 2d). As shown in Figure 3, there was a significant difference in mean T2 as well as SIR between pre- and post-treatment examinations in twelve TED patients receiving treatment with immunosuppression (mean difference T2: 12.05+/-17.79 ms, p<0.001, mean difference SIR: 0.26+/-0.56, p=0.001). On an individual basis (Fig. 3c), a reduction in muscle T2 was related to a reduction in CAS in patients. Discussion In this feasibility study we sought to evaluate if existing MRI protocols can measure muscle inflammation and disease activity in TED patients. We show that measurement of T2 relaxation time is a more robust method for assessing disease activity than STIR SIR. As fat is implicated in the disease process in TED and is a potential confounder to the measurement of T2 relaxation times, we also describe the novel measurement of fat fractions in extra-ocular muscles. T2 relaxation times could clearly differentiate abnormal from normal populations of extra-ocular muscles, as well as pre-treatment from post-treatment status, in turn correlating with CAS. The correlation between T2 relaxation time and CAS as well as post-treatment status has been investigated previously.[endnoteRef:23],[endnoteRef:24] However, the significant difference demonstrated in this study between TED patients and controls allows normative values to be considered. T2 relaxation time appears to be a reliable measure of disease activity, allowing the assessment of individual muscles. [23: Mancini L, Rajendram R, Uddin J, Lee RW, Rose GE , Yousry T, Miszkiel K,Thornton JS. Extra-orbital muscle T2 relaxation time and clinical activity in thyroid eye disease. Paper presented at International Society for Magnetic Resonance in Medicine. https://cds.ismrm.org/protected/10MProceedings/files/876_3843.pdf (accessed 28/10/18)] [24: Ohnishi T, Noguchi S, Murakami N, et al. Extraocular muscles in Graves ophthalmopathy: usefulness of T2 relaxation time measurements. Radiology 1994;190:857–62.] Although T2 relaxation time does correlate with SIR, which has also previously been shown to correlate with CAS and post-treatment status,[endnoteRef:25],[endnoteRef:26],[endnoteRef:27] the lack of a statistically significant difference in SIR between TED patients and controls at the level of individual muscles is striking and suggests that SIR is less discriminatory and unsuitable for quantitative assessment of disease activity. SIR measurements suffer from the need to choose an appropriate but arbitrary denominator, such as signal intensity from temporalis muscle or normal appearing white matter on the same image as the measured extra-ocular muscle. T2 relaxation time, as an estimate of an absolute value, is independent of machine and sequence-specific parameters that may affect SIR measurements and a more robust measure for application in longitudinal and cross-sectional comparisons. [25: Politi LS, Godi C, Cammarata G, 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;24:1118–26.] [26: Tortora F, Prudente M, Cirillo M, et al. Diagnostic accuracy of short-time inversion recovery sequence in Graves’ ophthalmopathy before and after prednisone treatment. Neuroradiology 2014;56:353–61.] [27: Mayer EJ, Fox DL, Herdman G, et al. Signal intensity, clinical activity and cross-sectional areas on MRI scans in thyroid eye disease. Eur J Radiol 2005;56:20–4.] In the temporal correlation between maximal T2 relaxation times and CAS (Fig. 3c), there were at least three cases with contradictory assessments (highlighted as darker shade). In each of these, there is a decrease in CAS despite an apparent increase in T2 relaxation time that suggested increased disease activity. Interestingly, clinical follow-up supported the radiological assessments, suggesting that T2 relaxation is a more sensitive marker of inflammation and more dynamic assessment of change. Specifically, in case 5, MRI picked up disease activity before it was reflected in the CAS. In case 7, the MRI results are seen to contradict the high initial CAS – however this was a falsely positive high CAS due to anterior segment inflammation, not due to TED. Finally, in case 9, the MRI detected myositis before it became apparent on the CAS. In addition, case 8 in the series demonstrates an overall correlation between T2 and CAS. However, the decrease in T2 is measurable prior to the decrease in CAS (when this was still 7) and only later reduced to 2. As discussed earlier, this is because of the lack of sensitivity of the CAS to change- there is only a reduction in the score when signs or symptoms are entirely resolved, not when there is a substantial improvement with residual symptoms. This study demonstrates the feasibility of measuring fat fraction in extra-ocular muscles, thereby quantifying the relative contribution of fat and water to the T2 signal. The increased FF within EOMs in TED patients is consistent with the suggestion that there is fatty infiltration and/or adipogenesis within the muscles in the course of the disease.[endnoteRef:28],[endnoteRef:29] Interestingly, the data also demonstrate that EOMs in healthy volunteers contain a measurable proportion of fat. Ultimately, higher fat fractions may reflect chronicity of the disease. Thus, the ability to differentiate between the contribution played by water and hyaluronic acid (presumed to equate to inflammation) and fat (presumed to equate to chronicity), may help guide decisions regarding treatment. [28: Li H, Fitchett C, Kozdon K, Jayaram H, Rose GE, Bailly M, Ezra DG. Independent adipogenic and contractile properties of fibroblasts in Graves' orbitopathy: an in vitro model for the evaluation of treatments. PLoS One. 2014;9(4):e95586.] [29: Lee BW, Kumar VB, Biswas P, Ko AC, Alameddine RM, Granet DB, Ayyagari R, Kikkawa DO, Korn BS. Transcriptome Analysis of Orbital Adipose Tissue in Active Thyroid Eye Disease Using Next Generation RNA Sequencing Technology. Open Ophthalmol J. 2018;12:41-52] The limitations of this approach include the manual placement of regions of interest within a single slice through each EOM. Superior and lateral rectus muscles have smaller cross-sectional areas, the superior rectus muscle is not easily separated from levator palpabrae superioris and lateral rectus lies at an oblique angle to the coronal plane in which images are acquired. The regions of interest could also include voxels at the edges of muscle, where there may be artefact related to the interface between muscle and orbital fat. These factors may account for some of the variability in measurements of fat fraction in both patients and controls. However, the variability is also consistent with the asymmetric and heterogeneous nature of the disease. There may also be an unknown effect of age as the control population are not age-matched. Furthermore, the patient population was unselected in terms of clinical severity and chronicity. In conclusion, quantitative T2 relaxation time measurements of EOMs in TED correlate with CAS and MRI SIR measurements. However, T2 relaxation measurements better discriminate abnormal from normal populations. Although this is an attractive objective measure of disease activity, it may be confounded by the presence of fat within muscles. We address this by concurrent measurement of fat fractions such that this combined MR approach may allow for improved quantification of disease activity and for monitoring within individual patients for treatment effect. Our physiological assessments of inflammation occasionally contradicted CAS scores, raising questions about the validity of clinical scoring of disease activity. Figures & Tables Table 1. Imaging Modalities for Thyroid Orbitopathy & their characteristics Imaging Modality Type of Signal Dangers Utility in TED Advantages Disadvantages Provides a Quantifiable Marker of disease activity? CT X-Ray Radiation & iodine based contrast Better than MRI at identifying enlarged muscles.[endnoteRef:30] Density of muscles can correlate with disease[endnoteRef:31] [30: Polito E, Leccisotti A . MRI in Graves orbitopathy: recognition of enlarged muscles and prediction of steroid response. Ophthalmologica. 1995;209(4):182-6.] [31: Ozgen A, Alp MN, Ariyürek M, Tütüncü NB, Can I, Günalp I. Quantitative CT of the orbit in Graves' disease. Br J Radiol. 1999;72(860):757-62.] Fast. Better bone resolution than MRI. Good for assessment of apical crowding[endnoteRef:32] [32: Pitz S, Muller-Forell W. Orbital Imaging. In Wiersinga WM, Kahaly GJ (eds): Graves’ Orbitopathy: A Multidisciplinary Approach – Questions and Answers. Basel, Karger, 2017 pp61-73.] Radiation. Volumetry of muscles may not correlate with disease activity[endnoteRef:33] [33: Bijlsma WR, Mourits MP. Radiologic measurement of extraocular muscle volumes in patients with Graves' orbitopathy: a review and guideline. Orbit. 2006;25(2):83-91] Mainly proptosis. MRI Nuclear magnetic resonance Patients with certain active or passive implants may be contraindicated. Better than CT at identifying areas of inflammation in muscle Better contrast between soft tissues than CT. Different sequences provide specific anatomical or physiological information Slow. Acoustic noise. Costly. Movement artefacts Potentially, but most clinical protocols provide qualitative or semi-quantitative data Ultrasound Sound echo None Limited Rapid. Available. Can perhaps exclude scleritis and intraocular pathology User-dependent. Limited depth. No visualisation of apex. Limited reproducibility No Doppler Sound echo None Limited clinical utility but blood flow in superior ophthalmic vein is reduced in TED.[endnoteRef:34] [34: Alp MN, Ozgen A, Can I, Cakar P, Gunalp I. Colour Doppler imaging of the orbital vasculature in Graves' disease with computed tomographic correlation. Br J Ophthalmol. 2000;84(9):1027-30.] Rapid. Available Difficult to perform. Inter-observer variability. Not specific for TED. No Octreoscan with 111In Gamma-ray Scintigraphy High Radiation Limited Orbital uptake of this labelled somatostatin analogue is greater in TED Non-specific. High cost. Poor availability. Requires careful standardisation. No Octreoscan with 99Tm Gamma- ray Scintigraphy Lower Radiation Limited. But significant correlation was found between CAS and the orbital uptake. Identifies active disease Lower cost. Greater availability. Higher energy so improved resolution. Shorter acquisition time Radiation. Invasive No Gallium-67 Scintigraphy Gamma- ray Scintigraphy Radiation Limited. Able to detect response to treatment Equivalent positive predictive value to octreotide and T2 relaxation time on MRI Invasive. Requires careful standardisation No, too difficult FDG-PET/CT Positron scintigraphy Radiation Limited but able to identify activity even when MRI normal[endnoteRef:35] [35: Kuo PH, Monchamp T, Deol P. Imaging of inflammation in Graves' ophthalmopathy by positron emission tomography/computed tomography. Thyroid. 2006;16(4):419-20.] May detect early/subclinical disease Radiation. Not widely used. No Thermal Heat None Limited. But higher temperatures recorded in TED. Non-invasive. Changes in response to iv steroids[endnoteRef:36] [36: Di Maria C, Allen J, Dickinson J, Neoh C, Perros P. Novel thermal imaging analysis technique for detecting inflammation in thyroid eye disease. J Clin Endocrinol Metab. 2014;99(12):4600-6.] Low specificity. Requires complex equipment to measure accurately No, too much variability Table 2. MRI-based techniques used for TED assessment MRI Sequence What it is Utility in TED Advantages or Disadvantages Provides a Quantifiable Marker for TED? Reference T1 A predominantly anatomical scan Muscle size measurement. Detection of fat (high signal) within muscle (lower signal) No measure of disease activity. Non-quantitative other than anatomical measurements Muscle dimensions (usually 2-dimensional). T1 signal interpretation is qualitative T1 + Gadolinium (Gd) A scan sensitive to abnormal contrast enhancement with Gadolinium Inflamed muscles demonstrate greater enhancement Requires IV contrast medium. Similar information can be obtained from non-contrast sequences such as STIR. No (qualitative due to variability) T1 + Gd SIR A ratio calculated by comparing signal in a muscle after contrast with a non-enhancing structure (eg thalamus) [Ref 23] Differentiates active disease from inactive & healthy. Allows tracking change over time Signal correlates with CAS[endnoteRef:37]. Requires IV contrast medium. Does not differentiate inactive disease from healthy controls [37: Tortora F, Cirillo M, Ferrara M, Belfiore MP, Carella C, Caranci F, Cirillo S. Disease activity in Graves' ophthalmopathy: diagnosis with orbital MR imaging and correlation with clinical score. Neuroradiol J. 2013;26(5):555-64.] Semi-quantitative, using a ratio with an arbitrary denominator Politi et al., 2014[endnoteRef:38] [38: Politi LS, Godi C, Cammarata G, Ambrosi A, Iadanza A, Lanzi R, Falini A, Bianchi Marzoli S. 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(5):1118-26.] Tortora et al., 2013 T2 A scan to mainly identify water content (generally higher intensity a.k.a. longer T2 relaxation time suggests higher water or fat content) Can help detect activity due to higher signal with inflammation Standard protocol for orbital imaging but not for TED. Normal muscle is dark on a bright background of orbital fat. Inflamed muscle may be bright and poorly differentiated from surrounding fat No (qualitative due to variability) T2 mapping Scan to derive quantitative measurements of T2 relaxation time. Highlights oedema and inflammation. Can detect activity and differentiate healthy from active disease. Correlates with response to treatment T2 relaxation time correlates with CAS. Allows tracking change over time. Should be independent of equipment. Sequence may not be readily available. Yes, but possible confounding by signal from fat Ohnishi et al.,[endnoteRef:39] [39: Ohnishi T, Noguchi S, Murakami N, et al. Extraocular muscles in Graves ophthalmopathy: usefulness of T2 relaxation time measurements. Radiology 1994;190:857–62.] Mancini et al.,[endnoteRef:40] [40: Mancini L, Rajendram R, Uddin J, Lee RW, Rose GE , Yousry T, Miszkiel K,Thornton JS. Extra-orbital muscle T2 relaxation time and clinical activity in thyroid eye disease. Paper presented at International Society for Magnetic Resonance in Medicine. https://cds.ismrm.org/protected/10MProceedings/files/876_3843.pdf (accessed 28/10/18)] STIR (Short tau inversion recovery) sequence A T2-weighted sequence with suppression of fat signal Highlights inflammation in muscle and helps to delineate active phase Standard readily-available sequence. Signal correlates with CAS[endnoteRef:41],[endnoteRef:42]. Susceptible to artefact (poor fat suppression, motion, paranasal sinus air, braces etc.) [41: Tortora F, Prudente M, Cirillo M, Elefante A, Belfiore MP, Romano F, Cappabianca S, Carella C, Cirillo S. Diagnostic accuracy of short-time inversion recovery sequence in Graves' Ophthalmopathy before and after prednisone treatment. Neuroradiology. 2014;56(5):353-61.] [42: Lingam RK, Mundada P, Lee V. 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(5):325-330.] No (qualitative due to variability) Hoh et al., 1994[endnoteRef:43] [43: Laitt RD, Hoh B, Wakeley C, Kabala J, Harrad R, Potts M, Goddard P. The value of the short tau inversion recovery sequence in magnetic resonance imaging of thyroid eye disease. Br J Radiol. 1994;67(795):244-7.] Mayer et al., 2001[endnoteRef:44] [44: Mayer E, Herdman G, Burnett C, Kabala J, Goddard P, Potts MJ. Serial STIR magnetic resonance imaging correlates with clinical score of activity in thyroid disease. Eye (Lond). 2001 Jun;15(Pt 3):313-8.] Tortora et al., 2014[endnoteRef:45] [45: Tortora F, Prudente M, Cirillo M, Elefante A, Belfiore MP, Romano F, Cappabianca S, Carella C, Cirillo S. Diagnostic accuracy of short-time inversion recovery sequence in Graves' Ophthalmopathy before and after prednisone treatment. Neuroradiology. 2014;56(5):353-61.] STIR SIR Signal intensity ratio comparing an unaffected muscle with an affected muscle or other structure. e.g. EOM and temporalis or normal white matter Can help detect activity and differentiate healthy from active disease. Allows tracking change over time Signal correlates with CAS but does not reliably differentiate TED patients from healthy controls. Variability with equipment. Semi-quantitative, using a ratio with an arbitrary denominator. Politi et al., 2014 [endnoteRef:46] [46: Politi LS, Godi C, Cammarata G, Ambrosi A, Iadanza A, Lanzi R, Falini A, Bianchi Marzoli S. 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(5):1118-26.] T2 with fat suppression (eg Dixon method) Aside from STIR, numerous other techniques are available to suppress fat signal. The Dixon method is a specific MRI technique that allows separate ‘water’ and ‘fat’ images to be generated Water fraction correlates with activity scores and can monitor effects of therapy. Exact methods vary between manufacturers. Most are untested for EOMs. Quantitative evaluation of water content and fat fraction in orbital fat. Untested for extra-ocular muscles. Kaichi et al.,2016 [endnoteRef:47] [47: Kaichi Y, Tanitame K, Itakura H, Ohno H, Yoneda M, Takahashi Y, Akiyama Y, Awai K. Orbital Fat Volumetry and Water Fraction Measurements Using T2-Weighted FSE-IDEAL Imaging in Patients with Thyroid-Associated Orbitopathy. AJNR Am J Neuroradiol. 2016;37(11):2123-2128.] Fat Fraction Mapping Technique that allows quantification of fat fraction by exploiting differences in physical properties of protons in water vs. lipid environments Quantitative evaluation of fat content within extra-ocular muscles Allows tracking over time. Sequence may not be readily available Yes. Fat fraction shows demonstrable differences in EOMs between patients with TED and healthy controls Das et al., 2019[endnoteRef:48] [48: Das T, Roos JC, Murthy R. T2-relaxation mapping with fat fraction assessment to objectively quantify clinical activity in Thyroid Eye Disease: a pilot study. Eye. (this report)] DWI Diffusion weighted imaging Scan that is sensitive to diffusivity of protons. Generally, oedema results in greater extra-cellular space and increased diffusivity. Apparent diffusion coefficient (ADC) is a marker of this Differentiates between patients with TED (higher diffusivity in EOMs) and healthy controls Helpful for excluding lymphoma and myositis[endnoteRef:49]. Signal correlates with CAS but values do not differ between patients with active and inactive disease[endnoteRef:50]. Susceptible to artefact and distortions at air-bone interfaces. Requires optimisation for use in orbits [49: Pitz S, Muller-Forell W. Orbital Imaging. In Wiersinga WM, Kahaly GJ (eds): Graves’ Orbitopathy: A Multidisciplinary Approach – Questions and Answers. Basel, Karger, 2017 pp61-73.] [50: Politi LS, Godi C, Cammarata G, Ambrosi A, Iadanza A, Lanzi R, Falini A, Bianchi Marzoli S. 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(5):1118-26.] ADC is a quantitative measure. Reproducibility is not established Politi et al., 2104[endnoteRef:51] [51: Politi LS, Godi C, Cammarata G, Ambrosi A, Iadanza A, Lanzi R, Falini A, Bianchi Marzoli S. 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(5):1118-26.] Lingam et al., 2017[endnoteRef:52] [52: Lingam RK, Mundada P, Lee V. 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(5):325-330.] DCE-MRI Dynamic contrast-enhanced MRI Serial scanning of a region of interest after injecting IV contrast medium to look at dynamics of contrast enhancement Highlights changes in microcirculation in TED. Can correlate with change from active to fibrotic stage Signal correlates with TED activity. Requires contrast and careful timing for image acquisition. Post-processing and analysis is complex Quantitative. However, relationship between contrast dynamics and TED EOM composition is unclear[endnoteRef:53] [53: Taoka T, Sakamoto M, Nakagawa H, Fukusumi A, Iwasaki S, Taoka K, Kichikawa K. Evaluation of extraocular muscles using dynamic contrast enhanced MRI in patients with chronic thyroid orbitopathy. J Comput Assist Tomogr. 2005;29(1):115-20.] Sun et al., 2017[endnoteRef:54] [54: Sun B, Song L, Wang X, Li J, Xian J, Wang F, Tan P. Lymphoma and inflammation in the orbit: Diagnostic performance with diffusion-weighted imaging and dynamic contrast-enhanced MRI. J Magn Reson Imaging. 2017;45(5):1438-1445.] Jiang et al., 2012[endnoteRef:55] [55: Jiang H, Wang Z, Xian J, Li J, Chen Q, Ai L. Evaluation of rectus extraocular muscles using dynamic contrast-enhanced MR imaging in patients with Graves' ophthalmopathy for assessment of disease activity. Acta Radiol. 2012;53(1):87-94.] Guo et al., 2016[endnoteRef:56] [56: Guo Y, Huo L, Wang P, Huang L, Chai C, Sun F, Xia S, Shen W. Evaluating the Microcirculation of Normal Extraocular Muscles Using Quantitative Dynamic Contrast-Enhanced Magnetic Resonance Imaging. J Comput Assist Tomogr. 2016;40(3):419-23] Taoka et al., 2005[endnoteRef:57] [57: Taoka T, Sakamoto M, Nakagawa H, Fukusumi A, Iwasaki S, Taoka K, Kichikawa K. Evaluation of extraocular muscles using dynamic contrast enhanced MRI in patients with chronic thyroid orbitopathy. J Comput Assist Tomogr. 2005;29(1):115-20.] Figure 1 Figure 1 Legend. Images of the orbits of a patient with thyroid eye disease. A - Coronal Short Tau Inversion Recovery (STIR): T2-weighted fat-suppressed image showing enlargement and heterogeneous hyperintensity of extra-ocular muscles (arrowheads). B - Coronal T1-weighted image: areas of T1 hyperintensity within the muscles are consistent with fatty change (arrowheads). C - T2 relaxation map shows areas of longer T2 relaxation (higher intensity) in the right inferior rectus muscle (arrow) compared to the left (arrowhead). D - Fat Fraction map shows higher fat fraction (higher intensity) in the left inferior rectus muscle (arrowhead) compared to the right (arrow). Figure 2 Figure. 2 Legend. A - Scatter plot showing a positive correlation between T2 relaxation times and STIR Signal Intensity Ratio (SIR) in extra-ocular muscles in TED patients and healthy controls (r=0.42 in TED patients). However, the groups are poorly separated across the SIR axis and better separated by T2 relaxation time. B - Comparison of mean(+/-SD) T2 relaxation times for individual extra-ocular muscles across the TED and control groups. T2 relaxation time is higher in the TED patients for each extra-ocular muscle. C - Comparison of mean(+/-SD) SIR for individual extra-ocular muscles across the TED and control groups. There was no statistically significant difference in mean SIR of any of the measured extra-ocular muscles between the two groups. D - Boxplot showing median and range of fat fractions for each extra-ocular muscle compared between TED and control groups. There were significant differences detected in some of the extra-ocular muscles. Mean FF for all muscles as a group was significantly different between TED and control groups (p < 0.001). (R = Right, L= Left, MR = Medial Rectus, IR = Inferior Rectus, SR = Superior Rectus, LR = Lateral Rectus. For p-values: *** - p<0.001, ** - p<0.01, * - p<0.05, ns - non-significant) Figure 3 Figure 3 Legend. The mean T2 relaxation time (A) and SIR (B) in extra-ocular muscles of pre-treatment baseline examinations was higher than on post-treatment examinations (error bars – standard deviation; A: p<0.001; B: p=0.04). (C) shows bubble plots of 12 individual cases with pre- and post-treatment examinations. Serial examinations are represented on the x-axis in arbitrary units of time. Maximal T2 relaxation times are plotted in the y-axis with CAS represented by the area of the bubbles (annotated for each data point). In most cases, T2 relaxation time for the worst affected muscle correlated with CAS, with both reducing following treatment. However, in cases 5, 7 and 9 (highlighted by darker shade), there is initially an increase in T2 relaxation time despite a decrease in recorded CAS. In case 8, there is initially a decrease in T2 relaxation time before a decrease in CAS is detected (see text for further discussion). Acknowledgements The authors should like to acknowledge Dr Paul Meyer, Dr N Antoun, and the MRI Radiographers References image1.png image2.emf image3.png