Detection of Inflammatory White Matter Pathology

Imaging guidelines for conventional brain MR imaging in the diagnosis of inflammatory CNS disease recommend a multisequence protocol consisting of 3 sequences ( Table 1 ). The first is a sagittal (preferably 3-dimensional [3D]) fluid-attenuated inversion recovery (FLAIR) sequence to depict the supratentorial brain, providing the highest sensitivity in detection of lesions close to the cerebrospinal fluid (juxtacortical and periventricular lesions). Second is proton-density (PD)/T2-weighted fast spin echo (FSE) or turbo spin echo (TSE) imaging, being highly sensitive for detection of WM lesions, particularly in the infratentorial WM. Final recommendations in the protocol are precontrast (optional) and postcontrast-enhanced T1-weighted images, which allow visualization of active lesions, that is, those associated with inflammatory activity and blood-brain barrier breakdown. For patients who present with symptoms at spinal cord level, or when brain MR imaging analysis is equivocal, the protocol recommends MR imaging of the spinal cord (postcontrast T1-weighted and FSE/TSE PD/T2-weighted sequences).

Images are evaluated for radiologic findings as seen in inflammatory diseases, concentrating on focal and diffuse WM and GM abnormalities. MR images obtained from patients with suspected MS are analyzed according to MR imaging criteria for DIS (Barkhof, Swanton), and the recently revised IP diagnostic criteria for MS, which are based on magnetic field strengths of 0.5 to 1.5 T. With 3 T MR imaging scanners being used more routinely in the clinical setting, one should question the accuracy of these criteria in determining lesion load in (suspected) MS. When comparing high (3 T/4 T) with lower (1.5 T) field-strength MR imaging in MS patients, a conclusive finding is an improved detection of WM lesions and contrast-enhanced lesions ( Table 2 ). Another important finding is that at high-field, more lesions are detected in anatomic regions important for establishing the diagnosis MS according to diagnostic criteria, such as periventricular, juxtacortical, and infratentorial WM lesions ( Fig. 1 ). The improvement in infratentorial lesion detection is also important in gaining information on prognosis of the disease, because these lesions have important prognostic value in the prediction of long-term disability in patients with CIS suggestive of MS.

Image examples of the higher sensitivity in the detection of inflammatory brain lesions at 3 T in comparison with 1.5 T. ( A , B ) A 23-year-old man presenting with unilateral optic neuritis. An inflammatory lesion in the left hemisphere of the cerebellum ( arrow ) was clearly identified on the T2-weighted turbo spin echo images at 3 T but not on the corresponding 1.5 T examination. ( C , D ) Axial fluid-attenuated inversion recovery (FLAIR) sections of the same patient. A small lesion in the brainstem and a lesion in the right temporal lobe ( arrows ) could be visualized on the 3 T image but not on the corresponding 1.5 T image. ( E , F ) Axial FLAIR sections of the supratentorial brain of a 45-year-old woman presenting with optic neuritis of her left eye. A small juxtacortical lesion ( arrow ) was prospectively identified on the 3 T image but was missed on the 1.5 T examination. Another lesion, which is probably a mixed white matter–gray matter lesion ( arrow ), is sharply delineated on the 3 T image but is more fuzzy and smaller on the 1.5 T image.

( Reproduced from Wattjes MP, Lutterbey GG. Higher sensitivity in the detection of inflammatory brain lesions in patients with clinically isolated syndromes suggestive of multiple sclerosis using high-field MR imaging: an intraindividual comparison of 1.5 T with 3.0 T. Eur Radiol 2006;16:2070; with permission.)

Detection of Gray Matter Pathology

GM pathology (cerebral cortex and deep GM structures) is a key feature of MS. It is already present in the earliest stages of the disease, and accumulates and accelerates more in the later and progressive phases. GM damage can manifest as a mixture of focal demyelinating lesions and diffuse abnormalities. Early pathology studies already acknowledged the extensive involvement of GM in MS.

Notwithstanding the recent focus of MR imaging research on GM abnormalities, thus far GM (especially subpial intracortical) lesions are still vastly underdetected by conventional in vivo MR imaging studies when compared with pathology studies. The sensitivity of conventional imaging techniques (T1, PD/T2, FLAIR) in detecting GM damage is poor, because these techniques lack the necessary contrast and resolution to visualize cortical demyelination. The pathophysiology and histopathology of cortical lesions (less inflammatory cell infiltration, no complement activation or blood-brain-barrier damage) and low myelin content of GM, plus partial volume effects from cerebrospinal fluid (CSF) on MR imaging, all contribute to this lack of sensitivity.

Improvement in the sensitivity of GM lesion detection was established with the introduction of the GM-specific double inversion recovery (DIR) sequence, higher resolution imaging of GM by the use of high-field MR imaging and, of course, the combination of these 2 methods. The development of DIR, which is not included in the conventional MR imaging protocol on a regular basis, leads to an improved (gray-white) contrast by depicting only GM. This is managed by using 2 inversion pulses leading to an attenuation of both CSF and WM. Disadvantages are its rather low SNR resulting from the double signal-inversion pulses, and the propensity to (flow and pulsation) artifacts, particularly in the posterior fossa. The development of multislab and, later, single-slab 3D-DIR applications led to a great improvement : a 5-fold increase in cortical lesions in MS patients was detected in comparison with conventional T2-weighted sequences. In addition, an improved distinction was made between mixed GM-WM lesions and purely intracortical lesions.

The introduction of high-field imaging did not immediately lead to an increase in detection of cortical lesions, when applied in postmortem research using conventional PD sequences on a 4.7 T MR imaging system. Most GM lesions were still missed; contrast between GM and GM lesions was found to be very low, independent of sequence or field strength. The combination of higher magnetic field strengths with DIR sequences in vivo seemed to be more successful in visualizing GM pathology. At 3 T, DIR was superior to the standard sequences in the detection of WM, mixed WM/GM, and intracortical GM lesions (see Table 2 ). One of these studies also demonstrated superiority of DIR at 3 T over other sequences for infratentorial WM lesions, which is clinically highly important as stated earlier. Both studies made use of a 2-dimensional (2D)-DIR sequence; implementing 3D-DIR at higher magnetic field (3 T) might result in further improvement of lesion detection by reducing artifacts, but will be accompanied by an increase in acquisition time.

Improved depiction of GM damage was achieved in studies using ultrahigh-field strengths (7 T and higher), the results of which are described in more detail later in this article.

The focus on how to depict GM damage has high clinical relevance, because cortical damage differs between MS disease types and stages, and shows a relationship with physical as well as cognitive disability. Furthermore, when including the presence of intracortical GM lesions in MR imaging diagnostic criteria, an increase in accuracy of these criteria has been reported. However, an official introduction of GM lesions into the diagnostic criteria has not yet been made and needs further multicenter validation. A step in the right direction was recently made by developing consensus recommendations for MS cortical lesion scoring using DIR.

Detection of Active Inflammatory Pathology

Magnetic field strength influences tissue relaxation times. With 3 T MR imaging, T1 (spin lattice/longitudinal) relaxation time increases by 20% to 40%, whereas T2 (spin spin/transverse) relaxation time decreases by about 5% to 10%. When using T1-shortening contrast agents at higher field strengths, for example, paramagnetic gadolinium-based contrast agents, the overall longer high-field T1 relaxation times will create a relatively stronger effect of T1 reduction by the contrast agent. This effect causes a greater postcontrast signal intensity difference at 3 T in comparison with 1.5 T, which increases the detection of contrast-enhancing inflammatory lesions in MS ( Fig. 2 ). It may even allow dosage reduction at higher field strengths. Nonetheless, because of decreased GM-WM contrast with increasing magnetic field, it remains a challenge to develop an SE T1 sequence for high-field MR imaging systems, which is the standard sequence used to detect inflammatory lesions at lower field strengths.

Image examples of increased detection of contrast-enhanced inflammatory brain lesions at 3 T in comparison with 1.5 T. ( A ) Multiple lesions are seen in the 3 T FLAIR image of a relapsing remitting MS patient. ( B ) On 3 T postcontrast T1-weighted image, a contrast-enhancing lesion with a perivascular location can be visualized ( arrow ). ( C ) This lesion could not be identified on the corresponding 1.5 T postcontrast image.

Paramagnetic contrast agents based on iron oxides such as (ultra) small particles of iron oxide ([U]SPIO), which have shown pluriformity of inflammatory MS pathology complementary to gadolinium-enhanced 1.5 T MR imaging, have not yet been applied at higher magnetic fields.

Detection of Spinal Cord Pathology

MR imaging of the spinal cord has gained more importance in establishing the diagnosis of MS, particularly in the recent IP criteria. As is known from standard field strength studies, conventional MR imaging shows asymptomatic spinal cord lesions in 30% to 40% of CIS patients and in up to 90% of CDMS patients. Besides aiding in diagnosis and differential diagnosis, imaging of spinal cord abnormalities is relevant because these are related to clinical outcome. Advanced MR techniques are sensitive to tissue damage in the spinal cord, and are related to measures of clinical outcome as well.

Postmortem studies using high-field MR imaging showed a better visibility of MS spinal cord pathology including quantitative and GM abnormalities, but the in vivo use of higher magnetic field strengths for imaging spinal cord remains problematic, mainly because of susceptibility, CSF, and pulsation artifacts. In contrast to brain MR imaging studies, in vivo comparison of 1.5 T and 3 T spinal cord MR imaging showed no significant differences in terms of lesion detection and correlations with clinical measures such as the Expanded Disability Status Scale (EDSS). A recent MR imaging study investigating spinal cord volumes at 3 T described a decrease in cervical spinal cord volume in progressive forms of MS and a trend toward increased spinal cord volume in relapsing remitting (RR) MS/CIS patients, which the investigators refer to respectively as atrophy and inflammation/edema–related expansion.

Clinical Value of Conventional High-Field MR imaging in Terms of Diagnostic Criteria

As described, studies investigating the influence of higher magnetic field strengths on lesion load measurement showed an evident improvement in lesion detection. However, the crucial question remains as to whether 3 T MR imaging scanners are of added clinical value in terms of an earlier diagnosis of MS. High-field 3 T MR imaging proved to be able to substantially influence classification of CIS patients according to Barkhof MR imaging criteria: 27.5% of the 40 patients studied fulfilled 1 additional criterion. Diagnostic classification in terms of DIS was mildly influenced: only 1 additional patient had DIS at 3 T on comparison with 1.5 T examinations. During follow-up no additional patients showed DIT at 3 T compared with 1.5 T examination, neither to the revised IP criteria nor to the Swanton criteria. Hence, when using the 2005 IP criteria, 3 T MR imaging does not lead to an earlier diagnosis of MS. When retrospectively applying the data of this cohort to the more liberal 2010 revised IP criteria that are based on MR imaging criteria developed by the MAGNIMS group, again no earlier diagnosis of MS could be established at 3 T when compared with 1.5 T. From these studies it can also be concluded that when using the revised IP criteria, 3 T MR imaging is safe and does not lead to field strength–influenced overdiagnosis of MS resulting from false-positive detection of WM lesions. However, this conclusion, together with the statement that in CIS patients there is no added clinical value of high-field 3 T MR imaging above standard 1.5 T MR imaging, might be too premature, because it was only based on one rather small, single-center, single-vendor data set. Future studies might lead to new and improved criteria for the diagnosis of MS, based on (ultra) high-field MR imaging in combination with novel sequences such as DIR.

Increased MS Lesion Detection Using Ultrahigh-Field MR imaging

Postmortem studies using ultrahigh magnetic field strengths discovered cortical lesions at 8 T that remained invisible at 1.5 T. At histopathologic verification, often these cortical lesions could only be found after observing the 8 T MR images. A recent postmortem study showed that 3D-T2∗ gradient echo sequences and WM-attenuated turbo field echo sequences at 7 T were able to detect most cortical lesions verified by pathological examination. Unfortunately, both postmortem studies used a limited number of patients, and further validation with a larger number of samples is warranted.

In vivo studies using ultrahigh magnetic field strength show increased detection of MS lesions in WM as well as in cortical GM. The improvement in detection of cortical GM lesions is important, because the depiction of this type of lesions has been difficult at lower field strengths. Mainero and colleagues found that 7 T MR images were able to differentiate cortical lesions in accordance with histopathologic lesion types (type 1: leukocortical; type 2: intracortical; type 3/4: subpial extending partly/completely through cortical layers). Type 3/4 was found to be the most frequent type of cortical plaques (50.2%), and this type was also related to higher EDSS scores. In vivo studies comparing advanced sequences between 7 T MR imaging and lower field strength do not exist, but the authors’ own data show that 3D-FLAIR and 3D-DIR using magnetization preparation (MP) allows high-quality T2-weighted MR imaging in MS at 7 T, and that 7 T 3D-FLAIR improves cortical lesion detection in comparison with 3 T ( Fig. 3 ).

( A–D ) Image examples of the higher sensitivity in the detection of brain lesions at 7 T in comparison with 3 T. Axial reformatted ( A ) 7 T 3-dimensional magnetization-prepared (3D-MP)-FLAIR and ( B ) 3 T 3D-FLAIR images of a 37-year-old female secondary progressive MS patient. A cortical lesion ( closed arrowhead ) can be identified at 7 T, but not on the corresponding 3 T image. A deep white matter lesion ( open arrowhead ) is visible at both field strengths. Sagittal ( C ) 7 T 3D-MP-FLAIR and ( D ) 3 T 3D-FLAIR images of a 50-year-old male primary progressive MS patient. Arrows indicate examples of lesions visible at 7 T but not on the corresponding 3 T image; the images also show multiple lesions that are visible at both 7 T and 3 T.

Visualizing Heterogeneity of MS Lesions at Ultrahigh-Field Strength

The main focus of ultrahigh-field imaging in MS patients thus far has been on heterogeneity of lesions, which is studied by making use of the increased magnetic susceptibility at higher field strength. Susceptibility-weighted imaging (SWI) identified the relationship between MS lesions and vasculature, and confirmed that MS lesions follow a strict perivascular distribution. That perivenular inflammation plays a role in MS was already identified in histopathology studies, but the possibility to visualize it in vivo provides opportunities to further investigate what determines lesion location (in a longitudinal setting) and might help in differentiating MS from ischemic WM lesions.

Another specific feature discovered by ultrahigh-field SWI in a subset of MS lesions are hypointense rims ( Fig. 4 ). Hammond and colleagues explained these rims to reflect iron-rich macrophages at the periphery of a lesion, which may indicate the site of active inflammation in tissue and might be of help in staging the disease. Postmortem pathology studies have identified this iron accumulation in MS plaques as well. Pitt and colleagues reported the hypointense rims to correspond to increased density of activated microglia. Hypointense rims have not been conclusively identified at 1.5 T.

Example of a deep white matter lesion in a 43-year-old male MS patient, which showed a hypointense ring at ( A ) 7 T 3D-MP-FLAIR ( arrow ) that can also be seen on ( B ) susceptibility-weighted imaging (SWI) ( arrow ), suggestive of iron deposition.

At even higher field strengths (9.4 T), T2-weighted scanning was able to discriminate areas of remyelination and demyelination in postmortem MS lesions. More recently, the same investigators studied 21 tissue samples of MS motor cortex, and found 28 GM cortical lesions that were visible on T2-weighted MR imaging as well as on sections immunostained for myelin basic protein. Furthermore, a correlation between quantitative MR and quantitative histology was made, which suggested that in cortical GM, T1 relaxation-time differences may be a predictor of neuronal density and T2 relaxation-time differences may predict myelin content. When these results can be translated into in vivo studies, for instance at 3 T or 7 T, they might possibly have a great impact on clinical translation of demyelination and neuronal loss.