Magnetic Resonance Imaging in Multiple Sclerosis

Carlos A. Pérez

John A. Lincoln

Introduction

Since its clinical introduction in the 1980s, magnetic resonance imaging (MRI) has become an essential tool in supporting the diagnosis, longitudinal monitoring, and evaluation of therapeutic response in multiple sclerosis (MS).¹ Although the diagnosis of MS is mostly based on clinical findings, MRI has become an integral part of the overall diagnostic process because of its ability to sensitively and noninvasively demonstrate the spatial and temporal dissemination of demyelinating plaques in the brain and spinal cord.²,³ In some cases, MRI can be useful for ruling out alternative neurological diseases.⁴ In this chapter, we discuss the underlying principles and clinical utility of MRI with the aim of helping clinicians understand how to better apply this valuable tool in the assessment of MS and related conditions.

How MRI Works

MRI provides exquisite detail of the brain and the spinal cord in the axial, sagittal, and coronal planes.⁵ Unlike computed tomography (CT) scans, it does not require the use of ionizing radiation. Instead, MRI uses a powerful magnetic field that aligns protons (hydrogen atoms) within water
molecules that are normally randomly oriented in the same or opposite direction as the external field.⁶ The alignment is briefly disrupted by the introduction of an external radiofrequency (RF) pulse, and the excited hydrogen atoms emit resonance signals as they return to their previously aligned (equilibrium) state that are then measured by a receiving coil.³ The frequency information contained in the signal from each location in the imaged plane is then converted to corresponding intensity levels that are displayed as shades of gray in a matrix arrangement of pixels.¹ By varying the sequence of the RF pulses applied and collected, different types of images are created.⁷ The contrast between different tissues is determined by the rate at which excited atoms return to their equilibrium state. The amount of time between successive RF pulses is referred to as repetition time, and the time between the delivery of the RF pulse and the receipt of the echo signal is referred to as the time to echo.⁷

T1-Weighted Sequencing

The longitudinal relaxation time, or T1, is the time constant that determines the rate at which the excited protons realign with the external magnetic field.³ The more quickly the protons realign, the greater (and brighter) the signal. The rate at which this occurs is determined by the T1 properties of a tissue. Fat quickly realigns its longitudinal magnetization with the magnetic field, short T1, and it therefore appears bright (i.e., hyperintense) on a T1-weighted image.⁸ Conversely, water has a much slower longitudinal magnetization realignment after an RF pulse, long T1, and therefore it appears dark (i.e., hypointense) on a T1-weighted image.⁸ Therefore, tissues with high fat content (such as white matter) will be bright, and compartments filled with water (such as cerebrospinal fluid [CSF]) will be dark on T1-weighted scans. In this way, T1-weighted images are good for demonstrating anatomy (Figure 4.1).⁸

T1-weighted imaging can also be performed after the administration of gadolinium. Gadolinium is a paramagnetic contrast enhancement agent that facilitates the relaxation of hydrogen atoms (i.e., shortens T1).¹ It preferentially shortens T1 values in tissues where it accumulates, rendering them bright on T1-weighted images. Gadolinium-enhanced images are especially useful in looking at pathological tissues such as tumors, and areas of inflammation or infection, because these will demonstrate accumulation of contrast due to disruption of the blood-brain barrier (BBB), which will make them appear brighter than the surrounding tissue.⁸

T2-Weighted Sequencing

The transverse relaxation time, or T2, is the time constant that determines the rate at which the excited protons lose resonance perpendicular
to the main field and become out of phase with each other after being excited by an RF pulse (i.e., dephasing).⁶ Dephasing occurs because of random and time-dependent field variations induced by spins of neighboring atoms, because not all spins have exactly the same precession frequency.⁶ The precession frequency of an atom refers to the rate of change in orientation of the rotational axis of protons due to an applied external magnetic field.⁷ The addition of an external RF pulse results in augmentation of the angle of precession of the protons, and in doing so, it converts some of the magnetization that exists along the axis of the dominant magnetic field (longitudinal magnetization) into measurable magnetization along a perpendicular axis (transverse magnetization).³,⁷ The rate at which this dephasing occurs is determined by the T2 properties of a tissue. The slower the dephasing, the greater (and brighter) the T2 signal.³ On a T2-weighted scan, compartments filled with water (such as CSF) appear bright because of protons in phase with each other. To illustrate, as water molecules move around in all directions, their local magnetic fields fluctuate, averaging each other out. Without a significant net difference in internal magnetic fields, the protons stay in step with the applied external field for a longer period of time. Conversely, tissues with high fat content (such as white matter) appear dark. T2-weighted scans are good for demonstrating pathology because most, but not all, brain lesions tend to develop edema and are associated with an increase in water content, which will make them appear bright.⁸ In general, T1- and T2-weighted images can be differentiated by looking at the CSF: CSF appears dark on T1-weighted imaging and bright on T2-weighted imaging³ (Table 4.1).

Figure 4.1. Appearance of tissue on T1, T2, and FLAIR (fluid attenuated inversion recovery) sequences. (A) T1 precontrast sequence showing the caudate nucleus (long arrow) and putamen (arrow head), (B) T2 sequence showing multiple sclerosis (MS) lesions (short arrow), (C) FLAIR sequence showing the same MS lesions (short arrow).

TABLE 4.1 T1- AND T2-WEIGHTED MRI SIGNAL INTENSITIES

	T1	T2
Hyperintense (bright)	Fat, cholesterol, intravascular blood flow, protein-rich fluid, gadolinium, hemorrhage	Water, CSF, vasogenic edema, intravascular slow flow or thrombus, infarction, inflammation, infection
Hypointense (dark)	Water/CSF, air, bone, hemosiderin, edema, infection, gliosis, intravascular flow void	Bone, air, fat, protein-rich fluid, increased cellularity, intravascular flow void
Data from McMahon KL, Cowin G, Galloway G. Magnetic resonance imaging: The underlying principles. J Orthop Sport Phys Ther. 2011;41:806-819. doi:10.2519/jospt.2011.3576.
CSF, cerebrospinal fluid; MRI, magnetic resonance imaging.

FLAIR Sequencing

A third commonly used conventional sequence is the fluid attenuated inversion recovery (FLAIR). The FLAIR sequence is similar to a T2-weighted image, but the high signal of normal CSF fluid is attenuated and made dark.⁵ The CSF signal is nullified by using a long inversion recovery sequence with a long inversion time (TI). A long inversion time suppresses the high CSF signal and improves the visualization of periventricular lesions.⁹ As with the T2-weighted image, this sequence is very sensitive to pathology and makes the differentiation between CSF and brain parenchymal abnormalities much easier to distinguish.⁹ FLAIR is particularly useful in the detection of subtle changes at the periphery of the hemispheres, near sulcal CSF, and in the periventricular region close to CSF (such as those typical of MS) (Figures 4.2A and 4.2B) where the high intensity of the CSF signal itself may attenuate visible contrast when compared with the high intensity of nearby lesions.⁵,¹⁰

Diagnostic Role of MRI in Multiple Sclerosis

Although there is no single diagnostic test for MS, MRI is routinely employed to evaluate a patient clinically suspected with MS. The diagnosis of MS is based on the principle of dissemination in time (DIT) and dissemination in space (DIS) of central nervous system (CNS) demyelination.¹¹,¹²,¹³ Conventional T1- and T2-weighted, as well as contrast-enhanced T1-weighted and FLAIR, images offer the most sensitive way of detecting lesions and are the current standard assessment methods to confirm the clinical diagnosis of MS.⁹ The high conspicuity of MS-related abnormalities seen on MRI provide the best view of tissue injury, lesion activity, and disease accumulation compared with all other imaging modalities, including CT.¹

Figure 4.2. A. Juxtacortical multiple sclerosis (MS) lesions are more readily identified on FLAIR (fluid attenuated inversion recovery) (a) compared with T2 (b) sequences (long arrows). B. As with the prior figure, many periventricular MS lesions are more readily identified on FLAIR (a) compared with T2 (b) sequences (long arrows).

Although the diagnosis of MS can be straightforward in patients with a typical clinical history, when the symptoms are nonspecific or atypical of MS, MRI is the most commonly performed investigation that can support a clinical diagnosis.¹⁴,¹⁵ For a considerable proportion of patients, MRI can replace some of the clinical criteria by revealing brain and spinal cord
changes that are typical of MS4 (Figure 4.3). The evolution of the diagnostic criteria for MS, from solely clinically based to the currently used McDonald criteria, reflects the increasing importance of MRI findings in establishing a timely and accurate diagnosis.¹⁶

Figure 4.3. Characteristic multiple sclerosis (MS) lesions. (A) Axial FLAIR (fluid attenuated inversion recovery) sequence showing MS lesions that are perpendicular to the callosal plane with discrete lesion borders and most measuring more than 3 mm in diameter. (B) Sagittal T2 sequence showing discrete MS lesion at C5 (long arrow).

Diagnostic Criteria for Multiple Sclerosis

The fundamental concept of DIT and DIS was first introduced by Schumacher et al. in 1965 as a first attempt to standardize diagnostic criteria for MS.¹⁶ Initially, these criteria were based on clinical features alone, as well as the elimination of alternative diagnoses with similar presentations. In 1983, the Poser criteria were proposed, which incorporated paraclinical tests (evoked potentials, neuroimaging, and CSF analysis) to supplement clinical evidence for the diagnosis of MS in situations where clinical criteria were not met.¹³

With the advent of MRI, the need for early diagnosis and treatment prompted revision of the widely used Poser criteria.¹⁷ In 2001, an international panel headed by Ian McDonald published new guidelines for the diagnosis of MS, commonly known as the 2001 McDonald criteria.¹³,¹⁴,¹⁸ For the first time, these guidelines proposed the use of MRI findings as supporting evidence for lesion dissemination in time and space with the potential to enable an earlier diagnosis.¹⁹ The concept of DIS was based
on the Barkhof criteria²⁰ (Table 4.2), which were originally developed to predict conversion to clinically definite MS in patients presenting with an isolated first clinical symptom.²⁰ They require at least 3 of 4 of: (1) one gadolinium-enhancing lesion or nine T2-hyperintense lesions if gadolinium-enhancing lesions are not present; (2) at least one infratentorial lesion; (3) at least one juxtacortical lesion (i.e., involving the subcortical U-fibers); (4) at least three periventricular lesions. DIT was determined by a gadolinium-enhancing or a new T2 lesion detected on repeat MRI performed 3 months (90 d) or more after the baseline scan.¹⁹

TABLE 4.2 BARKHOF CRITERIA FOR PREDICTION OF CIS CONVERSION TO CLINICALLY DEFINITE MS

▪ ≥1 Gadolinium-enhancing lesion or ≥9 T2-hyperintense lesions

▪ ≥1 Infratentorial lesion

▪ ≥1 Juxtacortical lesion

▪ ≥3 Periventricular lesions

CIS, clinically isolated syndrome; MS, multiple sclerosis.

When properly applied, the 2001 McDonald criteria showed high specificity (83%) and sensitivity (83%) for clinically definite MS at 3 years in patients presenting with a clinically isolated syndrome (CIS) suggestive of demyelinating disease.¹³ In 2002, a retrospective analysis reported that, with MRI and the McDonald Criteria, 50% of patients with a first clinical attack would receive a diagnosis of definite MS within a year compared with only 20% when using the Poser criteria.¹⁶ In the light of subsequent studies, the 2001 McDonald criteria were revised in 2005, 2010, and most recently in 2017,¹¹ further clarifying the role of MRI in the diagnosis of MS.

The first revision to the McDonald criteria was published in 2005.18 One key difference from the prior iteration was that DIT could be established on the basis of a gadolinium-enhancing or a new T2 lesion in an MRI scan performed 30 days (rather than 90 d) or more after the baseline scan. In addition, for the first time, spinal cord lesions were incorporated in the total lesion count. These criteria maintained the high specificity of the original 2001 McDonald criteria¹³ and achieved a sensitivity of 77% according to some studies.¹³,²¹,²²

In 2010, new evidence and consensus using the Swanton/MAGNIMS (Magnetic Resonance Imaging in Multiple Sclerosis) criteria led to further revision of the McDonald criteria.¹⁷,²³,²⁴ In the 2010 revision, the definition for DIS was simplified to include one or more T2 lesions in at least two of four key locations: juxtacortical, periventricular, infratentorial, and spinal cord.¹⁴ Gadolinium-enhancing lesions were no longer required for the determination of DIS. The criteria for DIT were modified to include any new
T2 or gadolinium-enhancing lesions on follow-up scan at any time after the baseline scan or the simultaneous presence of asymptomatic enhancing and nonenhancing lesions on the same scan regardless of timing.¹⁴ The sensitivity and specificity of the 2010 McDonald criteria reported by different studies range from 70% to 80% and 48% to 63%, respectively.¹¹,²²,²⁵

The McDonald criteria were redefined in 2017 (Table 4.3). As with previous revisions, these newly revised criteria are expected to speed the diagnostic process with increased sensitivity while preserving specificity and to reduce the possibility of misdiagnosis, although this will need to be evaluated prospectively. The core requirement of the diagnosis of MS remains the objective demonstration of dissemination of CNS lesions in both space and time, based on either clinical findings alone or a combination of clinical and MRI findings.¹² DIS is demonstrated with MRI alone by one or more T2 lesions in at least two of the four MS-typical regions of the CNS: periventricular, juxtacortical (and now also cortical), infratentorial, and spinal cord, or by the development of a further clinical attack implicating a different CNS site.²⁶ Both symptomatic and asymptomatic lesions contribute to lesion count. DIT is demonstrated by the simultaneous presence of both symptomatic or asymptomatic gadolinium-enhancing and nonenhancing lesions at any point in time or a new T2 and/or gadolinium-enhancing lesion(s) on follow-up MRI irrespective of its timing with reference to a baseline MRI. Positive findings of oligoclonal bands in the spinal fluid can now substitute for demonstration of DIT in some settings.¹² Table 4.4 shows the definitions of DIS and DIT according to the newly revised 2017 McDonald criteria. The sensitivity and specificity of the complete set of the 2017 McDonald criteria have not been fully evaluated yet.

It should be noted that even with the wide utility of MRI, a diagnosis of MS should follow the exclusion of other possible etiologies that can mimic MS in clinical presentation and/or MRI findings.¹³ MRI, like other clinical features or laboratory tests, is one piece of evidence that must be placed in the appropriate context to arrive at a correct diagnosis.

MRI in Multiple Sclerosis

Based on the chronologic changes to lesion morphology, MS lesion formation and activity can be divided into two phases: an acute phase characterized by contrast enhancement of lesions and a subacute phase characterized by changes in lesion signal intensity and size on unenhanced T1- and T2-weighted images.²,²⁷

In the acute phase, lesions are typically isointense to the normal white matter on T1-weighted imaging and therefore cannot be seen on an unenhanced T1 scan.¹,² However, the formation of new MS lesions is nearly always associated with a focal area of contrast enhancement on T1-weighted scans, which correlates with BBB disruption in the setting of acute perivascular

inflammation.⁸,¹³ Gadolinium enhancement may last up to 2 months in acute lesions, although the average duration of enhancement is 3 weeks.² According to the pattern of contrast uptake, lesions can be classified as nodular or ringlike.¹ New contrast-enhanced lesions are usually associated with a hyperintense lesion in the same location on T2-weighted images but can also be detected before T2 abnormalities develop.²⁷ Clinical relapses often occur when a new lesion involves an eloquent area of the brain or cord. However, many new lesions occur in noneloquent or clinically silent brain regions. Because contrast-enhanced T1-weighted scans can detect disease activity 5 to 10 times more frequently than the clinical evaluation of relapses, it is generally believed that a significant number of these lesions can be clinically silent at any given time.¹,⁸,¹⁴,¹⁵,²⁸

TABLE 4.3 2017 MCDONALD CRITERIA FOR DIAGNOSIS OF MULTIPLE SCLEROSIS (MS)

Clinical Presentation	Additional Data Needed for a Diagnosis of MS
≥2 Attacks and objective clinical evidence of ≥2 lesions ≥2 Attacks and objective clinical evidence of 1 lesion with historical evidence of prior attack involving a lesion in a different location	None. Dissemination in space (DIS) and dissemination in time (DIT) criteria have been met
≥2 Attacks and objective clinical evidence of 1 lesion	One of these criteria: ▪ DIS: additional clinical attack implicating different CNS site ▪ DIS: ≥1 symptomatic or asymptomatic MS-typical T2 lesions in ≥2 areas of the CNS: periventricular, cortical/juxtacortical, infratentorial, or spinal cord
1 Attack and objective clinical evidence of ≥2 lesions	One of these criteria: ▪ DIT: additional clinical attack ▪ DIT: simultaneous presence of both enhancing and nonenhancing symptomatic or asymptomatic MR-typical MRI lesions ▪ DIT: new T2 or enhancing MRI lesion compared with baseline lesion scan (without regard to timing of baseline scan) ▪ CSF-specific oligoclonal bands (not present in serum)
1 Attack and objective clinical evidence of 1 lesion	One of these criteria: ▪ DIS: additional clinical attack implicating different CNS site ▪ DIS: ≥1 symptomatic or asymptomatic MS-typical T2 lesions in ≥2 areas of the CNS: periventricular, cortical/juxtacortical, infratentorial, or spinal cord AND one of these criteria: ▪ DIT: additional clinical attack ▪ DIT: simultaneous presence of both enhancing and nonenhancing symptomatic or asymptomatic MR-typical MRI lesions ▪ DIT: new T2 or enhancing MRI lesion compared with baseline lesion scan (without regard to timing of baseline scan)
CNS, central nervous system; CSF, cerebrospinal fluid.

TABLE 4.4 2017 McDONALD CRITERIA FOR DEMONSTRATION OF DISSEMINATION IN SPACE AND TIME BY MRI

Dissemination in space

▪ ≥1 T2-hyperintense lesion(s) in two or more areas of the CNS: periventricular, cortical or juxtacortical, infratentorial, and spinal cord

Dissemination in time

▪ Simultaneous presence of gadolinium-enhancing and nonenhancing lesions at any time or by a new T2-hyperintense or gadolinium-enhancing lesion on follow-up MRI, irrespective of the timing of the baseline MRI

CNS, central nervous system; MRI, magnetic resonance imaging.

The subacute phase of MS lesion morphology and activity can be subdivided into early and late periods.³ In the early subacute period, observed within the initial 10 weeks, the T2-hyperintense lesion is a combination of an influx of inflammatory cells resulting in demyelination, axonal transection, and edema. During this time, there is also cessation of lesion contrast enhancement on postgadolinium T1-weighted imaging.¹ In the late subacute period, 3 to 5 months after initial inflammation, the T2-hyperintense lesion often decreases in size, because of not only decreased vasogenic edema but also a combination of degenerative and regenerative processes (gliosis and remyelination), respectively. Over the initial 6-month period, less than 40% of lesions become persistently hypointense on T1-weighted imaging, presumably secondary to permanent demyelination and severe axonal loss, and are referred to as “T1 black holes.”¹ The accumulation of T1 black holes has been shown to correlate with disease progression and disability.⁵,²³ Although not uncommon in the brain, T1 black holes are rarely seen in the spinal cord, although, in part, this is because imaging of the spinal cord is more challenging because of the small cross-sectional size, motion artifacts, and low lesion contrast.²

Although MS lesions can occur anywhere in the CNS, above the tentorium cerebri, they have a predilection for periventricular white matter and tend to have an ovoid configuration with the major axes perpendicular to the ventricular surface (Dawson fingers)¹ (Figure 4.4). During their initial stage, the lesions are typically thin and appear to be linear, which is likely associated with the inflammatory changes around the long axis of the medullary vein that create the dilated perivenular space.¹,²⁹ In addition to the periventricular regions, the corpus callosum, brainstem, U-fibers, optic nerves, and subcortical region are areas where MS lesions are frequently located.³⁰ In addition, lesions occur in the gray matter. Gray matter lesions are more easily detected on FLAIR imaging and other advanced sequencing techniques, including double inversion recovery (DIR)³¹ and phase-sensitive inversion recovery (PSIR) (Figure 4.5).³²,³³,³⁴

FLAIR and T2-weighted sequencing are the mainstays in the diagnostic workup of patients with MS (Figure 4.6).³⁵ The T2 and FLAIR lesion loads reflect the accumulation of gross tissue changes.³⁵ Although newly formed or enlarging T2 lesions might indicate new areas of MS-related tissue damage, T2 hyperintensities are nonspecific with respect to the actual pathological changes within the lesions and can represent areas of inflammation, edema, abnormal myelination, gliosis, or axonal loss.⁴,¹⁵,³⁶,³⁷ Table 4.5 shows the MRI characteristics of brain lesions typical of MS.

The majority of conventional MRI scanners use different magnet strengths, typically 1.5 or 3.0 T. In general, an increase in magnet

strength is expected to lead to an increase in the number of identifiable MS lesions.³⁸ Of note, open-configuration MRI scanners, which may be used when patients have difficulty tolerating a closed MRI machine, are usually less than 1.5 T, and the quality of images they provide is often suboptimal for detecting MS activity.³⁹ Additionally, there may be variability in scanning protocols and voxel size (resolution). Newer protocols use a “3D” image representing an isotropic voxel with dimensions of 1 × 1 × 1 mm cubed. By contrast, conventional “2D” images utilize a nonisotropic voxel with variable dimensions, commonly 2 × 2 × 5 mm. Because facilities may have different MRI scanners with different magnet strengths and imaging protocols, patients should be encouraged to use the same MRI facility for follow-up imaging. When possible, the same scanner should be used for a more accurate comparison of new and old MRI scans.

Only gold members can continue reading. Log In or Register to continue