the central vein sign on mri - ectrims congress · recent mri studies in ms patient brains have...
TRANSCRIPT
Pascal Sati, PhDStaff Scientist, Translational Neuroradiology Section
National Institute of Neurological Disorders and StrokeNational Institutes of Health
Bethesda, Maryland, USA
The central vein sign on MRI
ECTRIMS 2017 NAIMS/MAGNIMS ACTRIMS/ECTRIMS Teaching Course
Imaging the non-MS lesion in MS
Goals of this teaching course
To introduce the concept of central vein sign (CVS) and its potential clinical value
To review the literature on the CVS in MS and its MRI mimics
To present (standard and advanced) MRI techniques for imaging CVS
To discuss proposed guidelines and future directions for the clinical evaluation of CVS
Relationship between veins and MS plaques
First pathological observations Rindfleisch, 1863Charcot, 1869Dawson, 1916
CWM Adams. J Neurol Sci. 1975
Perivenular inflammatory infiltration at the onset of the MS lesionH. Lassmann. Phil. Trans. R. Soc. Lond. B 1999
Okudera et al. Neuropathology 1999 (soft X-ray microscopy of postmortem tissues)
Long, fine vessels of uniform caliber
Lumen caliber: 100 - 250 mm
(Hooshmand et al. Neuroradiology 1974)
V
Deep medullary veins of the brain
Deoxyhemoglobin is paramagnetic (high spin state S = 2 of heme iron)
Venous blood (~70% saturated) has shorter T2* relaxation time and faster dephasing than tissue
Gradient-Echo: long TE (T2* contrast) & high spatial resolution to visualize small veins
Reichenbach et al. Radiology 1997
7T MRI (NIH) (250 mm in-plane resolution) In vivo cerebral venography by MRI
First in vivo observation of central veins in MS
Tan et al., AJNR 2000 (1.5T MRI)
T2-weighted MR venogram
High frequency of central veins in MS lesions
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Percentage of CVS positive lesions
% CVS varies with lesion location in the brainTallantyre et al., Neurology 2008 (96% periventricular, 65% peripheral)Kilsdonk et al., J Neurol 2014 (94% PV, 84% DWM, 66% JC, 52% LC , 25% IC)
Possible age effect % CVSKilsdonk et al., J Neurol 2014 (92% when <40y vs 73% when >40y)
No significant differences between relapsing form vs. progressive from Kuchling et al., MSJ 2014 (80.4% in RRMS vs 76.5% in PPMS)
% CVS is unknown in pediatric MS patients
% CVS is unknown in MS spinal cord
Central vein sign (CVS) is observed in all types of MS lesionsperiventricular, deep white matter, subcortical, juxtacortical, leukocortical, intracortical, infratentorial
Question 1:
Recent MRI studies in MS patient brains have demonstrated the presence of central veins in which type(s) of white matter lesions?
A. Periventricular lesions & deep white matter lesions
B. Subcortical lesions & juxtacortical lesions
C. Infratentorial lesions
D. All of the above
Question 1:
Recent MRI studies in MS patient brains have demonstrated the presence of central veins in which type(s) of white matter lesions?
A. Periventricular lesions & deep white matter lesions
B. Subcortical lesions & juxtacortical lesions
C. Infratentorial lesions
D. All of the above
MS non-MS
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Incidental
NMOSD
NMOSD
Headaches
VascularMigraine
Dementia
SVD = small vessel disease
Lower frequency of central veins in non-MS lesions
SVD
SVD
Susac
Percentage of CVS positive lesions
MS non-MS
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Differentiating MS from its MRI mimics with the central vein sign ?
Incidental
Susac
NMOSD
NMOSD
Headaches
VascularMigraine
Dementia
SVD
SVD
SVD = small vessel disease
Percentage of CVS positive lesions
Diagnosing MS with the central vein sign ?
Tallantyre et al. Neurology (2011)
Proposed cutoff to establish MS diagnosis
MS WM lesions
incidental WM lesions (healthy or vascular risk factors)
CVS criteria proposed for MS diagnosis
40% cutoff Mistry et al. JAMA Neurol 201313 MS & 9 with microangiopathic lesions. 100% positive and negative predictive of MS diagnosis in 22 patients.
‘VIL45’ (> 45%) and ‘VIL60’ (>60%) cutoff Campion et al. Eur Radiol 2017 25 MS & 10 SVD. VIL45: 100% sensitivity & 80% specificity, VIL60: 96% sensitivity & 90% specificity.
6 morphologically characteristic lesions Mistry et al. MSJ 2014 13 MS & 7 SVD/migraine. Perfect classification of 20 cases.
Select 3 / Select 3* Solomon et al. MSJ 2017 10 MS & 10 migraine. Select 3: 0.52 sensitivity & 0.98 specificity. Select 3*: 0.83 sensitivity & 0.81 specificity.
Dworkin et al. (P575) Poster Session 1, Thurs 26th
Solomon et al. (P842) Poster Session 2, Friday 27th
Single-center studies indicate that CVS may help with differential diagnosis (discrimination between MS lesions and ischemic lesions)
Lummel et al., Neuroradiology 2011 (80% in MS lesions vs 78% in ischemic lesions)Lamot et al., Mult Scler Relat Disord 2017 (62% in MS lesions vs 71% in ischemic lesions)
Conflicting results on %CV in small vessel diseases
Need for more data on inflammatory MRI mimics of MSBehçet's disease, Sjögren’s syndrome, systemic lupus erythematosus, neurosarcoidosis,…
Massacesi et al., European Academy of Neurology 201689% in MS vs 15% in SAD (Behçet, SLE, APS)
Need for large-scale comparative studies using standardized methodsFilippi et al. Lancet Neurol 2016 (MAGNIMS)Sati et al. Nature Review Neurology 2016 (NAIMS)
Kilsdonk et., Eur Radiol 2014
Lummel et al., Neuroradiology 2011
Different criteria for the central vein sign
Tallantyre et al., Neurology 2011
Solomon et al., ACTN 2015
Reference MRI technique B0 (T) Gad
Tan et al., AJNR 2000 SWI 1.5 Yes
Ge et al., Arch Neurol 2008 T2*w Gradient-Echo 7 No
Tallantyre et al., Neurology 2008 T2*w Gradient-Echo 7 No
Tallantyre et al., Invest Radiol 2009 T2*w Gradient-Echo 7 No
Lummel et al., Neuroradiology 2011 SWAN 3 No
Tallantyre et al., Neurology 2011 T2*w Gradient-Echo 7 No
Wuerfel et al., Mult Scler 2012 T2*w Gradient-Echo 7 No
Sinnecker et al., Neurology 2012 T2*w Gradient-Echo 7 No
Dixon et al., Eur J Radiol 2013 T2*w Gradient-Echo 7 No
Kister et al., Mult. Scler. Int. 2013 T2*w Gradient-Echo 7 No
Kau et al., Eur Radiol 2013 T2*w Gradient-Echo 3 No
Kilsdonk et., J Neurol 2014 FLAIR* 7 No
Kilsdonk et., Eur Radiol 2014 FLAIR* 7 No
Sati et al., Mult Scler 2014 T2*w seg EPI 3 Yes
Kuchling et al, Mult Scler 2014 T2*w Gradient-Echo 7 No
Dal-Bianco et al, Eur Radiol 2015 FLAIR-SWI 7 No
Solomon et al., Ann. Clin. Transl. Neurol. 2015
FLAIR* / T2*w seg EPI 3 Yes
Lane et al., J. Comput. Assist. Tomogr. 2015 SWI 1.5 No
Lamot et al., Mult Scler Relat Disord 2017 SWI 3 No
Mistry et al., Mutl Scler 2016 T2*w Gradient-Echo 7 No
Campion et al., Eur Radiol 2017 FLAIR* / T2*w seg EPI 3 Yes
Different MRI protocols for the central vein sign
T2*-weighted Gradient-Echo
Signal
Echo Time
T2* decayWhite matter
T2* decayVeins
TE
SWM
SVein
Scan time > 8 min
Partial brain coverage
Anisotropic voxel ( ~ 0.5 x 0.5 x 2 mm)
Ge et al., Arch Neurol 2008 (7T)
Siemens: FLASHPhilips: T1 FFEGeneral Electric: SPGR
General Electric: SWAN
Lummel et al., Arch Neurol 2008 (3T)
Scan time ~ 6 min
Partial/Whole brain coverage
Anisotropic voxel (~ 0.5 x 0.5 x 2 mm)
Signal
Echo Time
T2* decay
STE2
STE1
STE3
STE4
TE1 TE2 TE3 TE4
T2* Multi-Echo Gradient-Echo
Magnitude Phase SWI
Siemens: SWI
Philips: SWIp
Haacke et al. AJNR 2009
Scan time = [4 min - 6 min]
Partial/Whole brain coverage
Anisotropic voxel (~ 0.5 x 0.5 x 2 mm)
Susceptibility Weighted Imaging
FLAIR
SWI (minIP)
T2-weighted
Kau et al., Eur Radiol 2013 (3T)
minIP = minimum intensity projection
Susceptibility Weighted Imaging
FLAIR-SWI: a fusion of FLAIR and SWI
Grabner et al. JMRI 2011
FLAIR FLAIR-SWI
T2* segmented 3D echo-planar-imaging (3D EPI)
Scan time = [3 ½ min - 6 min]
Whole brain coverage
Isotropic voxel = [0.5 mm - 0.75 mm]
Prototype sequence (Siemens, Philips)
Sati et al., MSJ (2014)Sati et al., Magnetom FLASH (2017)
3D EPI, 3T Siemens, 0.65 mm iso, ~ 5 ½ min
T2*w
Phase
Absinta et al. (oral #243, Parallel Session 15, Friday 27th)
Optional step: intensity correction (N4)
FLAIR*: a combined MR contrast
3T
Sati et al., Radiology (2012)
Sati et al., Magnetom FLASH (2017)
Effect of static magnetic field strength (B0)
Tallantyre et al., Invest Radiol 2009
7T (0.5 mm) 3T (0.8 mm)
% CVS = 87% % CVS = 45%
Fechner et al. (ePoster) 3DEPI sequence optimized for different field strengths (NIH)
Effect of static magnetic field strength (B0)
Effect of gadolinium-based contrast agent
Pre-injection
Sati et al., Mult Scler. 2014
3T venogram (minIP) with 3DEPI sequence
During injection
single dose gadobutrol (0.1 mmol/kg)
Post-injection (15 min)
Maggi et al., Acta Radiol Open. 2015
1.5T SWI(w/o gd)
1.5T SWI(w/ gd)
Effect of static magnetic field strength (B0)
Which of the following MRI techniques cannot be used to image central veins in brain lesions?
A. T2*-weighted gradient-echoB. T2-weighted spin-echoC. Susceptibility-weighted-imagingD. T2*-weighted segmented 3DEPIE. FLAIR*
Question 2:
Which of the following MRI techniques cannot be used to image central veins in brain lesions?
A. T2*-weighted gradient-echoB. T2-weighted spin-echoC. Susceptibility-weighted-imagingD. T2*-weighted segmented 3DEPIE. FLAIR*
Question 2:
Proposed radiological definition of the central vein sign
A central vein exhibits the following properties on T2*‐weighted images:
Appears as a thin hypointense line or small hypointense dot
Can be visualized in at least two perpendicular MRI planes, and appears as a thin line in at least one plane
Has a small apparent diameter (<2 mm)
Runs partially or entirely through the lesion
Is positioned centrally in the lesion regardless of the lesion’s shape
Exclusion criteria for lesions:
Lesion is <3 mm in diameter
Lesion merges with another lesion (confluent lesions)
Lesion has multiple distinct veins
Lesion is poorly visible
Susceptibility-based (T2*) contrast
High-resolution (~ 0.5 mm)
High-field strength (3T and above)
With gadolinium contrast agent injection at 1.5T
SWI, SWAN, SWIp (clinical scanners)
Segmented 3D EPI (research scanners)
Isotropic voxels (multi-planar reconstruction)
Combined FLAIR* for improved lesion detection
Technical recommendations for imaging the central vein sign by MRI
Future directions for the clinical evaluation of CVS
Differential diagnosis for all MRI mimics of MS (both adult and pediatric)Hypoxic-ischemic vasculopathy, inflammatory/inflammatory autoimmune, infectious, toxic and metabolic, traumatic, tumoral, hereditary/unknown. Aliaga and Barkhof, Handb Clin Neurol 2014
Prediction of MS conversion in clinically isolated syndrome (CIS) Miller et al. Lancet Neurol 2012
Prediction of MS conversion in radiologically isolated syndrome (RIS) Okuda, Neuroimaging Clin N Am. 2017Makhani et al., Neurol Neuroimmunol Neuroinflamm 2017
Undiagnosing the misdiagnosedSolomon et al., Neurology 2012Solomon & Corboy, Nat Rev Neurol 2017
The central vein sign in multiple sclerosis: a real-life multi-centre MAGNIMS study
Background: first large-scale multi-centre 3T MRI study to test the usefulness of the central vein sign (CVS)
as an MS specific biomarker in a “real world” setting
479 patients with a variety of conditions (e.g. CIS/RRMS,NMOSD, (cerebral) vasculitis, migraine, small vessel ischaemia,diabetes) from nine MAGNIMS centres were included yet.3784 lesions were analysed. The initial findings areencouraging. A final batch with additional patients is currentlybeing analysed.
Blinding to global lesion loadand distribution: image is splitinto 8 equal-size blocks
Operator views overlay of SWI/FLAIR to decide on thepresence of central vein in accordance to NAIMS CVScriteria
Courtesy of Drs. Tim Sinnecker and Jens Wuerfel (MIAC, University Hospital Basel)
CentrAl Vein Sign in the early diagnosis of Multiple Sclerosis (CAVS-MS)
A pilot study to demonstrate the feasibility of utilizing 3T FLAIR* for central vein imaging in a multi-center setup
Which of the following statements is not part of the radiological definition for the central vein proposed by the NAIMS guidelines?
A. Appears as a thin hypointense line or small hypointense dot on T2* contrastB. Can be visualized in at least two perpendicular MRI planesC. Has a large apparent diameter (>2mm) D. Runs partially or entirely through the lesion E. Is positioned centrally in the lesion
Question 3:
Which of the following statements is not part of the radiological definition for the central vein proposed by the NAIMS guidelines?
A. Appears as a thin hypointense line or small hypointense dot on T2* contrastB. Can be visualized in at least two perpendicular MRI planesC. Has a large apparent diameter (>2mm) D. Runs partially or entirely through the lesion E. Is positioned centrally in the lesion
Question 3:
Does this lesion meet the NAIMS definition for the central vein sign?
A. YesB. NoC. I don’t know…
Axial
Question 4:
Does this lesion meet the NAIMS definition for the central vein sign?
A. YesB. NoC. I don’t know…
Axial
Question 4:
SagittalCoronal
Axial
Question 5:
Does this lesion meet the NAIMS definition for the central vein sign?
A. YesB. NoC. I don’t know…
Axial
Question 5:
Does this lesion meet the NAIMS definition for the central vein sign?
A. YesB. No (no visible vein)C. I don’t know…
SagittalCoronal
Axial
Question 6:
Does this lesion meet the NAIMS definition for the central vein sign?
A. YesB. NoC. I don’t know…
Axial
Question 6:
Does this lesion meet the NAIMS definition for the central vein sign?
A. YesB. No (excluded, confluent & multiple vessels)C. I don’t know…
SagittalCoronal
Central Vein Sign should only be used in a research setting
Thank you for your attention