Àlex Rovira Unidad de Neurorradiología. Servicio de Radiología

Hospital Vall d’Hebron Barcelona

alex.rovira@idi.gencat.cat

RM en la Esclerosis Múltiple

Barcelona, 6-8 junio 2016 Preceptorship in Multiple sclerosis

T2 and CE T1-WI

• Highly sensitive for detecting MS plaques (white matter)

• Provide quantitative assessment of disease activity and severity

• Characterize disease course over time

• Monitor and predict treatment response

• Most important paraclinical tool for diagnosing and monitoring MS

T2-weighted (FLAIR)

Post-contrast T1-weighted

MR imaging in MS

Bakshi et al. Lancet Neurol 2008;7:615–25

Incidental finding Virchow-Robin spaces normal population (5-10%) migraine (x4) Hipoxic-ischemic vasculopathies

small-vessel disease hyperhomocysthenimia CADASIL Susac’s syndrome

Primary demyelinating diseases multiple sclerosis and variants ADEM neuromyelitis optica

Vasculitis primary systemic: lupus, Behçet, APLAS Miscellaneous

neurosarcoidosis Lyme disease PML metabolic: Fabry, Leber, xantomatosis, adult forms of leukodystrophy effects of radiation therapy or drugs lymphoma metastasic disease

Multifocal WM signal abnormalities: “white spots”

Sánchez Aliaga E, Barkhof F. Handb Clin Neurol 2014;122:291-316; Charil et al. Lancet Neurol 2006;5:841-52

Solomon et al. Neurology 2012;78:1986-91

•Survey: 122 Neurologist (90% from Academic Hospitals) •95% reported having evaluated 1 or more patients who had been diagnosed with MS, but who they strongly felt did not have MS (within the last year) •Mainly due to overuse and improper interpretation of MRI (non specific findings) •>25% under treatment (difficult to take away)

Increase specificity of MRI findings is highly required

Misdiagnosis of Multiple Sclerosis

2013

2015

55 year old female with a diagnosis of multiple sclerosis. Treated with DMDs since 2009

Solomon et al. Curr Neurol Neurosci Rep 2013; Charil et al. Lancet Neurol 2006; 5: 841–52;Solomon et al. Neurology 2012;78:1986-91; Rudick et al. Neurology. 2013;80:777; Kim et al. Mult Scler. 2013;19:1060-7

•Most frequent contemporary reason for misdiagnosis •MR features not considered in the context of careful clinical evaluation •Overdiagnosis of RIS

Misdiagnosis of MS: overreliance on MRI interpretation

•Incidental multifocal WM brain lesions on MRI

normal population aged 18-50 (5-10%) migraine (x4)

Misdiagnosis has significant consequences: • Clinical, psychosocial and scientific • Health care system cost (overtreatment with DMTs; approx 40.000 subjects in US)

Comprehensive checklist for evaluation of WM spots

Brief and precise diagnostic impression that must consider:

Demographics Family history Vascular risk factors Clinical information and question Lab findings

Diagnostic strategy in subjects with incidental multifocal brain T2 lesions of unknown origin

Systematic reading

•Lesion distribution / involvement subcortical/periventricular U-fibers cortical grey matter deep grey matter corpus callosum brainstem spinal cord

•Lesion shape •Enhancement pattern •Features on special sequences: SWI •Ancillary findings

microbleeds vascular abnormalities

Kanecar et al. Radiol Clin N Am 2014; 52:241-61; Sánchez Aliaga E, Barkhof F. Handb Clin Neurol 2014;122:291-316

• Cortical lesions • Spinal cord MRI: diagnostic and prognostic value • Perivenular distribution • Iron imaging

Update on Diagnosis

Kutzelnigg et al. Brain 2005

RRMS SPMS PPMS

focal demyelinated plaques in the white matter

cortical demyelination

demyelinated lesions in the deep grey matter

Cortical gray matter involvement in MS

• Characterised by: – demyelination1

– microglial activation1

– often meningeal inflammation2,3

• Less often associated with4

– immune cell influx

– complement activation

– BBB leakage

• Difficult to detect by MRI5

• Three types of cortical lesion*6

*Based on post-mortem tissue samples taken from 22 patients with MS. Leukocortical Type I lesions involve neocortex and subcortical white matter; intracortical Type II lesions are confined to the neocortex and often located around a vessel; subpial Type III lesions extend from the pial surface into the neocortex. 1. Peterson JW et al. Ann Neurol 2001; 2. Lucchinetti CF et

al. N Engl J Med 2011; 3. Magliozzi R et al. Ann Neurol 2010; 4. Klaver R et al. Prion 2013; 5. Filippi M et al. Neurology 2010; 6. Wegner C et al. Neurology 2006

38% are Type I (leukocortical)

18% are Type II (intracortical)

44% are Type III (subpial)

Focal lesions in grey matter 90% of MS autopsy cases show cortical demyelination

Type I Type II Type III

Courtesy of Dr. García-Merino

T2-FLAIR

leukocortical lesion Lucchinetti et al. NEJM 2011;365:2188-97

Juxtacortical lesions (type I)

MRI sensitivity depends on lesion type and sequence

Geurts et al. J Neurol 2008; Wattjes et al. Am J Neuroradiol 2006; Geurts et al. Radiology 2005; Roosendaal et al. Mult Scler 2009

• cMRI detects <10% of pure intracortical lesions • Higher detection of mixed lesions • Improved sensitivity by using DIR or heavily 3D T1-weighted sequences

mixed lesion intracortical lesion

Cortical lesions (type II-III)

T2-FLAIR DIR

Type I-III lesions

Present in 44% of CIS and in 70% of MS patients

Absent in NMO, migraine…

Calabrese et al. Neurology 2012; 79:1671-6; Absinta et al. J Neurol 2012; 259:2695-8; Pareto et al. Am J Neuroradiol 2015 Oct 8. [Epub ahead of print]

Cortical lesions in CIS and diagnostic criteria

At least 2 of the following:

•1 enhancing or 1 spinal cord lesion •1 infratentorial lesion •1 cortical lesion

DIS Filippi 2010

DIS Polman 2005

DIS Polman 2010

YES

NO

DIS Polman 2005 DIS Polman 2010 DIS Filippi 2010

Sensitivity

74 % 86% 77%

Specificity

73% 42% 93%

Accuracy

74% 61% 86%

OR

7.9 4.3

47.3

Filippi et al. Neurology 2010; Filippi et al. Lancet Neurol 2016

Filippi M, et al. Lancet Neurol 2016; Fox RJ. Lancet Neurol 2016

Cortical lesions in CIS and diagnostic criteria

• Intracortical lesions typically need advanced imaging (DIR) • Assessment of intracortical lesions has a high variability • Inclusion of intracortical lesions into the diagnostic criteria has little clinical

benefit

• Corpus callosum lesions: • Subependymal lesions: • Type I-III lesions (FLAIR): • Type II-III lesions (DIR): • Type I lesions (DIR):

62 CIS patients (3.0 T) (Vall d’Hebron) Concordance analysis (two observers) Mean of kappa for all coder-pairs

Agreement (Landis and Koch)

0.0-0.2 = slight

0.21-0.40 = fair

0.41-0.60 = moderate

0.61-0.80 = substantial

0.81-1.0 = almost perfect/perfect

0.857 0.821 0.838 0.477 0.430

DIR / T2-FLAIR: intracortical vs leukocortical lesions

DIR T2-FLAIR

Nelson et al. Mult Scler 2008

3D MPRAGE

A significant proportion of lesions classified as cortical GM lesions on DIR appear to contain more WM than expected or do not involve GM

Spinal cord lesions in 30% of subjects with RIS

84% progressed to CIS or PPMS (median time 1.6 years) OR of clinical progression: 75.3

Subclinical lesions in 27-53% of patients with CIS

Spinal cord lesions 83% of patients with early relapsing MS

Spinal cord lesions in 74-92% of patients with MS and in 6% of patients with non-MS white matter diseases

O`Riordan et al. JNNP 1998; Dalton et al. JNNP 2003 Lycklama à Nijeholt GJ et al. Brain 1998; Bot et al. Neurology 2004

Okuda et al. Neurology 2011

Prevalence of spinal cord lesions

O'Riordan et al. J Neurol Neurosurg Psychiatry. 1998;64:353-7; Dalton et al. J Neurol Neurosurg Psychiatry. 2003;74:1577-80; Nijeholt et al. Brain 1998;121:687-97; Bot et al. Neurology 2004;62:226-33; Okuda et al. Neurology 2011;76:686-92

Selection of T2w MR sequences

Nayak et al. Acta Neurol Scand 2014; Riederer et al. Am J Neuroradiol 2015; Chong et al. Am J Neuroradiol 2016

PD > T2 STIR > T2

T2 STIR T2 PD

DIR > T2

T2 DIR

Cord almost isointense with surrounding CSF

Easy identification of any increase in signal

Higher sensitivity compared to T2 SE

More susceptible to artifacts (false positive)

Use it in combination with T2

Selection of MR sequences

Heavily T1-weighted sequences, such as PSIR (phase-sensitive inversion recovery) or MPRAGE /MP2RAGE (two inversion-contrast magnetization-prepared rapid gradient echo), improve MS lesion detection

Nair et al. Am J Neuroradiol 2013; Alcaide-Leon P, et al. Am J Neuroradiol. 2016

DIR MP2RAGE

T2 sequences: spinal cord

Single echo heavily T2 weighted1-2 : – limited sensitivity in depicting signal abnormalities1-2 – should not be obtained as a stand-alone sequence3

Combination of at least two T2w sequences3: T2, PD, STIR

T2 STIR

1. Philpott et al. Eur J Radiol 2011; 80:780-5; 2. Bot et al. Eur Radiol 2000; 10:753-8; 3.Rovira et al. Nat Rev Neurol 2015;11:471-82.

No cord swelling (unless active)

Unequivocal hyperintense T2 or Gd-enhancing; focal lesions

≥3mm in size; <2 vertebral segments long

Peripheral location, cigar shaped

Occupying only part of cord cross-section (less than 50%)

Typical MR imaging findings: spinal cord

Bot et al. Neurology. 2004; 62:226-33; Weier et al. Mult Scler 2012;18:1560-9; Gass et al. Lancet Neurol 2015;14:443-54

Weier et al. Mult Scler 2012;18:1560-9

Multiple sclerosis: spinal cord MRI

Spinal cord MRI in the diagnosis process and in monitoring

Magnims: Magnetic Resonance Imaging in Multiple Sclerosis network (Europe) CMSC: consortium of Multiple Sclerosis Centers (North America)

1.) Rovira et al. Nat Rev Neurol 2015;;11:471-82; 2.) Wattjes et al. Nat Rev Neurol Advance Pub Online 15 Sept 2015 ; 3.) Traboulsee A, et a. Am J Neuroradiol 2015 (in press)

• Spinal cord MR imaging if:1-3

• Transverse myelitis (PTM vs. LETM)

• Insufficient features on brain MRI to support diagnosis of MS

• Age >40 with non-specific brain MRI findings

• Atypical/unexpected new spinal cord symptoms

• Clinical disease progression and relapses not explained by brain MRI (Serial spinal cord imaging may add little to brain imaging alone in monitoring disease activity and progression)

Brain MRI with equivocal findings

Rovira et al. Nat Rev Neurol 2015;11:471-82.

Non spinal CIS not fullfilling McDonald brain MRI criteria

Spinal cord MRI in CIS: •All patients with SC presentation •Non SC patients who do not meet McDonald criteria on brain MRI

Sombekke et al. Neurology 2013

SC lesions

No SC lesions

• 7 MRI scans needed to diagnose 1 more patient • Prognostic value: identifies a subgroup that has a very

low risk of developing MS

Prognostic value of spinal cord MRI in CIS Amsterdam cohort

Prognostic value of spinal cord MRI in CIS The Barcelona inception cohort

Arrambide G et al. Poster presentation (P996) ECTRIMS 2015

Presence of SC lesions is an independent risk factor for evolving to MS

Covariables considered:

•Brain T2 lesion number •OCB in CSF •Age •Gender •CIS topography •Treatment with DMT

• Study design: single-center, observational • Sample size: 207 CIS patients (31% with a spinal cord syndrome) • Follow-up: mean 35.7 (15.8) months. • Outcomes: conversion to MS (CDMS, McDonald)

Arrambide G et al. Poster presentation (P996) ECTRIMS 2015

Prognostic value of spinal cord MRI in CIS The Barcelona inception cohort

• Study design: single-center, observational • Sample size: 207 CIS patients (31% with a spinal cord syndrome) • Follow-up: mean 35.7 (15.8) months. • Outcomes: reach significant disability EDSS ≥ 3.0

Proportion of patients with SC lesions and EDSS ≥3.0.

Presence of at least one SC lesion was associated with an EDSS ≥ 3.0: 11.8% (vs 1.8%) p=0.003

The presence of at least one SC lesion at the time of the CIS is associated with short-term disability and further contributes to estimate the risk of disability accumulation, particularly in non-SC CIS.

aHR for reaching an EDSS >=3.0. Prognostic value of SC lesions on reaching an EDSS ≥3.0. A: All patients (n=207). B: SC CIS (n=64). C: Non-SC CIS (n=143).

Indications of spinal cord MRI in Multiple Sclerosis

Situation Objective

Clinically isolated syndrome (all) Detect symptomatic and clinically silent lesions Rule out other diseases Increase specificity and sensitivity at diagnosis Predict disability at diagnosis

Negative brain scan, but strong clinical suspicion of MS

Increase sensitivity of diagnosis Investigate possible absence of spinal cord lesions, which could rule out MS

Nonspecific brain MRI findings (e.g. vascular-related lesions, ageing, incidental findings in migraine and/or chronic headache)

Increase sensitivity of diagnosis Investigate possible presence of spinal cord lesions, which could support the diagnosis of MS

Primary progressive MS Increase sensitivity and specificity of diagnosis Detect additional lesions Rule out other diseases

Radiologically isolated syndrome Increase specificity of diagnosis Predict risk of conversion to MS

Monitoring (disease progression and relapses not explained by brain MRI)

Detect disease activity

Rovira A, De Stefano N. Curr Op Neurol 2016 (in press)

Zecca et al. Mult Scler J 2015

Spinal cord MRI in monitoring MS

103 RRMS patients: clinically stable Median interval between scans: 17 months New asymptomatic lesions

A significant proportion of disease activity only in the SC, a fact that could have important implications in assessing and predicting treatment response

46.6%

15.5% 28.1% 9.8%

brain

spinal cord

•43.7% brain •25.2% spinal cord •9.8% only asymptomatic SC lesions

Perivenular topography of MS plaques “Dawson‘s fingers“

Post-mortem pathology studies show central vein

in > 90% white matter lesions

Jens Wuerfel (Berlin)

Dawson J. Trans Roy Soc Edinb 1916; 50:517-740

Horowitz et al. Am J Neuroradiol 1989;10:303-5

venule plaque

• Novel MR technique described by Haacke et al. (MRM 2004)

• Contrast based on the tissue ‘‘magnetic architecture’’ at the subcellular and cellular levels

• Increases sensitivity and conspicuity of MRI for detecting iron containing tissues and small veins (high concentration of deoxihemoglobin) due to their paramagnetic properties

Susceptibility-weighted iamges

Veins visibility Iron within basal ganglia and thalami

(low signal)

Susceptibility-weighted Imaging

FLAIR* sequence (3D GRE + 2D FLAIR at 3T) Central vein visibility

Tallantyre et al. Neurology 2011;76:534-9

7T

Presence of a central vein could be a marker to discriminate between MS and non-MS WM lesions

SWI in T2 lesions (7T)

FLAIR=fluid-attenuated inversion recovery. Kilsdonk ID et al. Eur Radiol. 2014;24:841–849

Axial 7-T FLAIR images of an MS patient and a patient with vascular brain lesions

Solomon et al. ECTRIMS 2015

Migraine-related WMLs vs MS

Central vein sign in MS

Absence of central vein sign in migraine

T2-FLAIR* 3T in RIS

T2-FLAIR* 3T in RIS

Intralesional susceptibility signal (ISS) in MS (3T)

SWI FLAIR

Intralesional susceptibility signal (ISS) 48% of non-enhancing MS lesions 58% of enhancing MS lesions

Susceptibility-weighted MR imaging

Likely represents iron-rich macrophages / microglia Myelin loss also contributes

Rovira et al. Multiple Sclerosis Journal 2015; 21 (S11):209

Hagemeier et al. J Magn Reson Imaging 2012;36:73-83; Bian et al. Mult Scler 2013;19:69-75

Serial analysis with QS mapping at 3T

•Magnetic susceptibility increases rapidly as it changes from enhanced to non-enhanced •High susceptibility values during the first 2-4 years •Then gradually decreases (susceptibility similar to NAWM)

Chen et al. Radiology 2014; 271:183-92

Susceptibility-weighted MR imaging in focal MS lesions

1. Early active

2. Center, chronic-active

Large, myelin-laden macrophages without iron

Small myelin-laden macrophages and occasional iron-containing macrophages

Macrophages with large amounts of iron, but without myelin

Perls’ staining

3. Rim, chronic-active

Myelin debris within macrophages detected with oil red-O

Mehta et al. PLOS one 2013;e57573

MS: Lession categories

1

2

3

No or only small amounts of iron

4. Chronic silent white matter lesions

Susceptibility-weighted MR imaging Determinants in image contrast

Myelin, iron, deoxyhemoglobin, and free radicals—all relevant in MS pathogenesis—influence susceptibility signal

• T2*/phase contrast in acute MS lesions appear to be influenced by the presence of free radicals

• T2*/phase contrast in chronic MS lesions appear to be influenced by the topography of iron-laden macrophages and ferritin within lesions

Absinta et al. Ann Neurol 2013

T2-FLAIR

Migraine-related WMLs vs Multiple Sclerosis (3T)

Rovira et al. Multiple Sclerosis Journal 2015; 21 (S11):209

T2-FLAIR SWI

Areas of intralesional signal loss on SWI increases

diagnostic specificity and accuracy MS

Migraine

Migraine-related WMLs vs Multiple Sclerosis (3T)

Rovira et al. Multiple Sclerosis Journal 2015; 21 (S11):209

Sinneckeret al. Neurology 2012;79:708-14

SWI in T2 lesions (7T) 2D T2*- W FLASH

Neuromyelitis optica vs. multiple sclerosis

Venocentric lesions Hypointense rim

SWI

FLAIR

Ovoid lesions

Juxtacortical lesions

Corpus callosum lesion Juxtacortical lesion

ISS within lesions

Preclinical multiple sclerosis or Radiologically isolated

síndrome (RIS)

Clinical case: young female with migraine

Subclinical spinal cord lesions

Frontal subcortical lesions

No subclinical spinal cord lesions

No juxtacortical, corpus callosum lesions

No ISS

SWI

Incidental findings

FLAIR

Clinical case: young female with migraine

Summary

Wide variety of causes may present with multifocal WM lesions

MRI is the preferred imaging technique for diagnostic workup Radiological interpretation with demographic, clinical history,

and lab findings Standardized brain (spinal cord) MRI protocol Comprehensive checklist for evaluation of WM spots is crucial

Sánchez Aliaga E, Barkhof F. Handb Clin Neurol 2014;122:291-316; Charil et al. Lancet Neurol 2006;5:841-52; Rovira et al. Nat Rev Neurol 2015;11:471-82