Imaging Differences between Neuromyelitis Optica Spectrum Disorders and Multiple Sclerosis: A Multi-Institutional Study in Japan

H. Tatekawa; S. Sakamoto; M. Hori; Y. Kaichi; A. Kunimatsu; K. Akazawa; T. Miyasaka; H. Oba; T. Okubo; K. Hasuo; K. Yamada; T. Taoka; S. Doishita; T. Shimono; Y. Miki

doi:10.3174/ajnr.A5663

Article Figures & Data

Figures

Tables

Download figure
Open in new tab
Download powerpoint
Fig 1.
Patient selection flowchart. Consecutive patients, 89 with NMOSD and 89 with MS, are examined. According to the eligibility criteria, 79 patients with NMOSD and 87 with MS are assessed for brain analyses, 57 patients with NMOSD and 55 with MS are assessed for spinal cord analyses, and 42 patients with NMOSD and 14 with MS are assessed for optic nerve analyses.
Download figure
Open in new tab
Download powerpoint
Fig 2.
Bar graphs show the proportion of patients classified by the number of brain lesions in bins of 10 lesions (A), and the distribution of brain lesions categorized by the location (PVWM, DWM, SCWM, DGM, BS, cerebellum) (B) for rater 1. A total of 911 brain lesions in 79 patients with NMOSD and 1659 brain lesions in 87 patients with MS are identified. A, The proportion of patients is significantly different between NMOSD and MS (P = .015; weighted κ value, 0.92). More patients with NMOSD have no brain lesions of ≥3 mm, and a tendency for patients with MS to have more brain lesions than those with NMOSD is found. B, The distribution of lesions categorized by location is significantly different between NMOSD and MS (P < .001). DWM lesions (68%) are more frequent than PVWM lesions (17%) in NMOSD, whereas the difference in the frequencies of lesions in PVWM (41%) and DWM (42%) is small in MS.
Download figure
Open in new tab
Download powerpoint
Fig 3.
Graphs show the proportion of patients classified by the number of spinal cord lesions (A), the distribution and proportion of spinal cord lesions (B), and the length of spinal cord lesions in each location (C) for rater 1. A total of 86 spinal cord lesions in 57 patients with NMOSD and 102 spinal cord lesions in 55 patients with MS are identified. A, No significant difference is found in the number of lesions between NMOSD and MS (P = .76; weighted κ value, 0.94). Forty-seven (82%) patients with NMOSD and 37 (67%) with MS have ≥1 spinal cord lesion. B, Bimodal distributions of spinal cord lesions are present in both NMOSD and MS, but the peak of the distribution in NMOSD is high in thoracic regions, whereas the variation and peaks of the distribution are relatively smaller in MS than in NMOSD. The proportions of lesions categorized into cervical or thoracic regions are significantly different between NMOSD and MS (P = .024). More thoracic lesions (71%) than cervical lesions (29%) are found in NMOSD, whereas the difference in the frequencies of cervical (46%) and thoracic lesions (54%) is small in MS. C, In NMOSD, thoracic lesions are significantly longer than cervical lesions (P = .001), whereas in MS, the length is not significantly different between cervical and thoracic lesions (P = .80).

Tables

Figures

View popup

Table 1:

Patient demographics and characteristics in NMOSD and MS^{^a}

	NMOSD	MS	P Value^{^c}
Demographics (NMOSD, n = 89) (MS, n = 89)
Age (yr) (median) (IQR, range)	51 (39–61, 16–85)	36 (29–43, 18–67)	<.001
No. of females^{^b}	77 (86.5)	68 (76.4)	.12^{^d}
Disease duration (yr) (median) (IQR, range)	4 (0–11, 0–73)	2 (1–7, 0–21)	.40
EDSS (median) (IQR, range)	6 (2–7.5, 1–9)	2 (1–3, 0–6)	<.001
AQP4-IgG^{^b}
Positive	83 (93.3)
Negative	5 (5.6)
Unknown	1 (1.1)
MS type^{^b}
Relapsing-remitting MS		79 (88.8)
Primary-progressive MS		6 (6.7)
Secondary-progressive MS		2 (2.2)
Unknown		2 (2.2)
Brain MRI (NMOSD, n = 79) (MS, n= 87)
Age (yr) (median) (IQR, range)	50 (38–61, 17–83)	36 (29–45, 19–67)	<.001
No. of females^{^b}	68 (86.1)	67 (77)	.16^{^d}
Disease duration (yr) (median) (IQR, range)	4 (0–11, 0–41)	3 (1–7, 0–21)	.59
EDSS (median) (IQR, range)	6 (2–7.5, 1–9)	2 (1–3, 0–6)	<.001
Gadolinium enhancement^{^b}	51 (64.6)	68 (78.2)
Spinal cord MRI (NMOSD, n = 57) (MS, n = 55)
Age (yr) (median) (IQR, range)	53 (39–61, 25–78)	37 (29–43, 18–66)	<.001
No. of females^{^b}	51 (89.5)	36 (65.5)	.003^{^d}
Disease duration (yr) (median) (IQR, range)	4 (0–12, 0–43)	2 (0–7, 0–19)	.18
EDSS (median) (IQR, range)	6 (2–7, 1–9)	2 (2–3.5, 0–6)	<.001
Optic nerve MRI (NMOSD, n = 42) (MS, n = 14)
Age (yr) (median) (IQR, range)	50 (37–61, 17–79)	35 (34–39, 19–67)	.004
No. of females^{^b}	37 (88.1)	9 (64.3)	.10^{^d}
Disease duration (yr) (median) (IQR, range)	5 (0–12, 0–42)	1 (0–6, 0–17)	.15
EDSS (median) (IQR, range)	6 (3–8, 1–9)	3 (2–3.5, 2–4)	.065

Note:—EDSS indicates Expanded Disability Status Scale.
↵a Unless otherwise indicated, data are medians, with interquartile range and total range in parentheses.
↵b Data in parentheses are percentages.
↵c Mann-Whitney U test.
↵d Fisher exact test.

View popup

Table 2:

Number and size of brain lesions for rater 1 and assessment of brain morphologic features and characteristic signs

Quantitative Analyses	NMOSD (n = 79)	MS (n = 87)	P Value	Accuracy	κ Value
Total No. of lesions	911	1659
Per patient^{^a}	5 (1–18, 0–81)	8 (3–28, 0–120)	.004^{^c}
Diameter of lesions (mm)^{^a}	5.7 (4.3–8.5, 3.0–45)	6.1 (4.6–8.3, 3.0–56)	.046^{^c}
In each region (mm)^{^a}
PVWM	9.2 (5.3–14, 3.0–36)	6.8 (5.1–9.5, 3.0–47)	<.001^{^c}
DWM	5.5 (4.2–7.7, 3.0–45)	5.5 (4.3–7.4, 3.0–56)	.73^{^c}
SCWM	5.4 (3.7–7.2, 3.0–28)	5.9 (4.5–7.9, 3.0–25)	.054^{^c}
DGM	5.9 (4.3–8.3, 3.1–30)	6.7 (5.5–9.2, 3.5–48)	.31^{^c}
BS	6.2 (5.3–8.3, 3.3–22)	6.9 (5.2–8.1, 3.6–16)	.96^{^c}
Cerebellum	3.7 (3.7–3.7, 3.7–3.7)	7.3 (5.5–8.9, 3.4–21)	.17^{^c}
Morphologic assessment
Brain atrophy^{^b}	4 (5.1)	5 (5.7)	1^{^d}	0.48	0.52
Ventriculomegaly^{^b}	3 (3.8)	3 (3.4)	1^{^d}	0.48	0.5
Characteristic signs
Ovoid lesions^{^b}	17 (21.5)	55 (63.2)	<.001^{^d}	0.71	0.68
T1 black hole lesions^{^b}	16 (20.3)	47 (54)	<.001^{^d}	0.66	0.65
Callosal-septal-interface lesions^{^b}	23 (29.1)	48 (55.2)	.001^{^d}	0.63	0.75
Isolated U-fiber lesions^{^b}	8 (10.1)	24 (27.6)	.005^{^d}	0.57	0.56
Dirty white matter lesions^{^b}	13 (16.5)	23 (26.4)	.14^{^d}	0.54	0.61
Tumefactive MS lesions^{^b}	1 (1.3)	2 (2.3)	1^{^d}	0.48	0.74
Cloudlike enhancement^{^b}	0 (0)	0 (0)

↵a Data are medians, with interquartile range and total range in parentheses.
↵b Data in parentheses are percentages.
↵c Mann-Whitney U test.
↵d Fisher exact test.

View popup

Table 3:

Number, size, and location of spinal cord lesions for rater 1 and assessment of spinal cord morphologic features^{^a}

Quantitative Analyses	NMOSD (n = 57)	MS (n = 55)	P Value^{^c}	Accuracy	κ Value
Total No. of lesions	86	102
Per patient	1 (1–2, 0–6)	1 (0–3, 0–6)	.77
Longitudinal length (mm)	47 (17–109, 4.5–408)	13 (9.0–20, 4.0–208)	<.001
In cervical region (mm)	23 (9.5–36, 4.5–149)	13 (8.4–21, 4.0–110)	.077
In thoracic region (mm)	63 (25–131, 6.0–408)	13 (9.4–20, 4.1–208)	<.001
Transverse diameter (mm)	4.4 (3.2–6.2, 1.8–13)	4.4 (3.6–5.3, 1.9–11)	.99
In cervical region (mm)	5.8 (3.6–7.5, 1.9–13)	4.7 (4.0–5.8, 1.9–11)	.55
In thoracic region (mm)	4.2 (3.1–5.7, 1.8–10)	4.0 (3.5–4.8, 2.3–7.3)	.76
Intramedullary location^{^b}
Central	66 (76.7)	59 (57.8)	.007^{^d}
Peripheral	11 (12.8)	32 (31.4)
Both	9 (10.5)	11 (10.8)
Morphologic assessment
Atrophy^{^b}	17 (29.8)	5 (9.1)	.008^{^d}	0.40	0.53
Swelling^{^b}	19 (33.3)	8 (14.5)	.027^{^d}	0.41	0.67

↵a Unless otherwise indicated, data are medians, with interquartile range and total range in parentheses.
↵b Data in parentheses are percentages.
↵c Mann-Whitney U test.
↵d Fisher exact test.

View popup

Table 4:

Laterality and location of optic nerve lesions and assessment of optic nerve morphologic features^{^a}

	NMOSD (n = 42)	MS (n = 14)	P Value^{^b}	Accuracy	κ Value
Laterality of lesions
None	16 (38.1)	7 (50)	.67		0.73
Unilateral	18 (42.9)	4 (28.6)
Bilateral	8 (19)	3 (21.4)
Location of lesions
Optic nerve	26 (61.9)	7 (50)	.54		0.7
Optic chiasm	2 (4.8)	2 (14.3)	.26		0.7
Optic tract	0 (0)	1 (7.1)	.25		0.49
Atrophy	5 (11.9)	0 (0)	.32	0.66	0.48
Swelling	10 (23.8)	5 (35.7)	.49	0.66	0.41