Table 3.

MRI sequences applied in GO assessment.

Tissue or organs	Index	Method	MRI sequence	MRI findings	Reference
Orbital fat	Exophthalmos	the perpendicular distance between the interzygomatic line and the surface of the cornea	T1WI	1–2 mm difference between MRI and Hertel ophthalmometry	Cevik et al. (15) Maria et al. (16)
	Volume	ROI outlined and restructured by Mimics	T1WI with thin layers	Orbital fat volume in GO is higher than healthy control	Shen et al. (17)
	Thickness	The maximum distance between the eyeball and medial wall	T1WI	The thickness increased successively among the healthy control, responsive group and unresponsive group	Hu et al. (18) Xu et al. (19)
EOMs	Diameters	Short Diameter: medial and lateral rectus muscles were measured on axial images, others on coronal images	T1WI	Affected by many factors, a possible predictor of glucocorticoid response	Xu et al. (19)
	Volume	ROI outlined and restructured by Mimics	T1WI with thin section	EOMs volume in GO are higher than healthy control	Shen et al. (17)
EOMs	Inflammation	Draw ROI on the maximum EOMs cross-section	T2 mapping	T2RT got from T2 mapping is higher in therapeutic responsive group than unresponsive group	Zhai et al. (20)
		Draw ROI on the muscle with highest signal intensity	STIR-T2WI	SIR is correlate with CAS	Mayer et al. (21, 22)
			Dixon-T2WI	Dixon-T2WI has fewer artifacts and higher efficacy than traditional FS sequences	Ollitrault et al. (23) Chen et al. (24)
			Echo planar DWI, non-EPI DWI	Both sequences can discriminate GO from controls, but non-EPI DWI might have higher efficacy	Politi et al. (25) Feeney et al. (26)
	Fat infiltration	Intramuscular fat quantification by specific calculation	Dixon-T2WI	FF of EOMs in GO is higher than normal	Das et al. (27)
	Fibrosis	Draw ROI of inferior rectus and medial rectus muscles on the maximum cross-section	Non contrast T1 mapping	Although several EOMs show higher signal on FS sequence, decrease in T1 SI predict unresponsible to therapy	Matsuzawa et al. (28)
		Draw ROI of four rectus muscles at muscle belly precontrast and postcontrast	Pre/post contrast T1mapping	ECV is higher and relate to pathological findings in inactive groups	Ma et al. (29)
Lacrimal gland	Herniation	The perpendicular distance between the interzygomatic line and the most anterior tip	T2WI with FS	The herniation value is higher in active and glucocorticoid responsive patients	Gagliardo et al. (30)
	Inflammation	“Hotspot”: ROI which only a little proportion of the whole cross-section placed on the highest SI region	T2WI with FS	SIR is higher in active GO than inactive	Hu et al. (31)
		Draw ROI on the maximum LG cross-section	T2 mapping	T2 value is higher in GO than GD and it’s an independent predictor for the diagnosis of GO	Wu et al. (32)
Optic nerve	DON	Muscle index and T2 value got from four continuous slices and select the most efficacy slice	Dixon-T2WI, T2 mapping	Muscle index and T2 value are higher in DON	Zou et al. (33)
		The optic nerve sheath diameter, optic nerve diameter and optic nerve subarachnoid space got from two continuous slices and select the most efficacy slice	Modified Dixon-T2WI	The optic nerve subarachnoid space is larger in DON than GO and health control	Wu et al. (34)

MRI, magnetic resonance imaging; T1WI, T1 weighted image; ROI, regions of interest; GO, Graves ophthalmopathy; EOMs, extraocular muscles; T2RT, T2 relaxation time; SIR, signal intensity ratio; CAS, clinical activity score; T2WI, T2 weighted images; FS, fat suppressed; DWI, diffusion weighted image; EPI, echo planar imaging; FF, fat fraction; SI, signal intensity; ECV, extracellular volume; LG, lacrimal gland; GD, Graves’ disease; DON, dysthyroid optic neuropathy.