Intracortical Lesions by 3T Magnetic Resonance Imaging and Correlation with Cognitive Impairment in Multiple Sclerosis

Abstract

Background

Accurate classification of MS lesions in the brain cortex may be important in understanding their impact on cognitive impairment. Improved accuracy in identification/classification of cortical lesions was demonstrated in a study combining two MRI sequences: double inversion recovery (DIR) and T1-weighted phase-sensitive inversion recovery (PSIR).

Objective

To evaluate the role of intracortical lesions (IC) in MS related cognitive impairment (CI) and compare it to the role of mixed (MX), juxtacortical (JX), the sum of IC + MX and with total lesions as detected on DIR/PSIR images. Correlations between CI and brain atrophy, disease severity and disease duration were also sought.

Methods

39 patients underwent extensive neuropsychological testing and were classified into: normal and impaired. Images were obtained on a 3T scanner and cortical lesions were assessed blind to the cognitive status of the subjects.

Results

238 cortical lesions were identified (130 IC, 108 MX) in 82% of the patients, 39 JX lesions were also identified. Correlations between CI and MX lesions alone (p=0.010) and with the sum of IC + MX lesions (p=0.030) were found. A correlation between severity of CI and EDSS was also seen (p=0.009).

Conclusion

Cortical lesions play an important role in CI. However our results suggest that lesions that remain contained within the cortical ribbon do not play a more important role than ones extending into the adjacent white matter; furthermore the size of the cortical lesion, and not the tissue specific location, may better explain their correlation with CI.

Introduction

Cognitive impairment (CI) is prevalent among patients with multiple sclerosis (MS). A controlled long-term natural history study showed an increase in CI from 26 to 56% over a ten-year interval and it also showed that nearly half of all MS patients will develop CI during their lifetime [1], some early in the disease course [2]. The form or degree of CI does not correlate with the disease course [3]. CI is an important cause of disability in many MS patients with minimal or no other physical impairment. Current diagnostic tools for documenting CI (neuropsychological testing) are limited in their availability, expensive and not often included in insurance coverage. This prevents patients from being objectively diagnosed and treated with medication and behavior modification (cognitive rehabilitation, coping skills and counseling) that could potentially prolong their ability to maintain financial and social independence. The absence of documented cognitive abnormalities may also prevent them from qualifying for long term disability. Patients with mild CI are often able to work if adjustments are made. However, as the severity of CI increases they are often encouraged to quit, or terminated with no other option but to apply for long term disability. This has an important economic impact on both the individual and the health care system. Baseline testing on MS patients has shown deficits on tasks of verbal memory, abstract reasoning and linguistic processes and on tasks of attention and short term spatial memory [4]. Cross-sectional studies have also shown relative decline on tasks of recent memory, processing speed, attention, visual-spatial abilities, and executive functions which represent working memory[5].

Conventional and nonconventional MRI measures have been correlated with CI in MS. These include: whole brain atrophy [6–7], cortical volume [8], lesion load [9], and diffusion anisotropy [10]. Rao et al. examined a number of MRI variables including total lesion area, ventricular-brain ratio and size of the corpus callosum. They found total lesion area was predictive of dysfunction in memory, abstract and conceptual reasoning, language and visuospatial problem solving [4]. Another study found a correlation with lesion burden in frontal and non-frontal regions and attention, memory, planning, problem solving, and conceptual reasoning [11]. In a different study the MS functional composite (MSFC) score did not significantly correlate with T2 or T1 lesion load [12]. Volumes of white matter [13] and T2 hyperintense lesions in deep gray matter nuclei [14] also correlated with overall cognitive functioning. In a recent study of 550 mildly disabled MS patients tested at baseline (327 of whom underwent MRI assessments), CI (defined as impaired performance in ≥3 cognitive tests) was present in approximately 20% of all patients. In the subgroup that underwent MRI, T2 and T1 lesion volumes were significantly higher in patients with CI than those without [15]. Although the above studies have provided important information on MRI measures and correlation with CI, there is no consensus if these findings can or should be used as an MRI risk indicator of functional impairment.

The suggested association between cortical/juxtacortical lesions and CI in MS has also been explored but is less than convincing. Lazeron et al, observed a correlation, albeit weak, between juxtacortical lesions on brain MRI and the Cognitive Impairment Index [16]. Charil et al, observed a correlation between CI and lesions at the GM-WM junction of associative, limbic and prefrontal cortices [17]. A recent study found that the impact of neocortical lesions on CI is not independent from white matter disease [18]. Correlations have also been found between specific gray matter volumes in the cerebral cortex and cognitive test results. Portaccio et al, found that neocortical volume loss was significantly associated with worse performance on cognition measures over time; no such correlation was found between total brain or T2 lesion volumes [19].

Direct visualization of intracortical lesions is suboptimal on conventional MRI as demonstrated by combined post-mortem MRI and histopathology studies [20–21], reflecting limitations of imaging methods [22]. The sensitivity of detecting lesions present in the cortex was substantially improved by combining two novel pulse sequences; double inversion recovery (DIR) and T1-weighed phase sensitive inversion recovery (PSIR) [23]. DIR imaging does not allow detection of all lesions potentially present in the cortex, but this sequence undoubtedly provides a more complete picture of the “cortical” disease burden in MS [23–27]. This picture is strengthened by the confirmation of lesion presence on the PSIR images. When obtained using a 3T magnet, the superior GM/WM contrast and clear delineation of lesion borders seen on PSIR also allows for a more accurate classification. Although 3D magnetization prepared rapid acquisition gradient echo (MPRAGE) has been shown to be an excellent sequence that provides important qualitative details for classification of cortical lesions, it does not seem to be the most sensitive for cortical lesion detection. The high spatial resolution of this sequence is helpful to evaluate details (i.e. lesion borders) but when used for detection purposes only, lesion presence is less obvious than on DIR/PSIR. The large number of slices also makes it visually difficult and extremely time consuming for the evaluator [25]. DIR/PSIR remain the best combination for improved detection and satisfactory classification of cortical gray matter lesions in MS at 3T and are presently reasonably accessible by the MS scientific community [23–25]. (Figure 1)

Figure 1.

Examples of Cortical lesions

In this study we proposed that intracortical lesions (IC) defined as total confinement within the cortical ribbon, might play a more important role in cognition when compared with mixed (MX), defined as present mostly in the cortex but with some subcortical extension and juxtacortical (JX), mostly present in the subcortical region with clear extension into the cortex. We have used the advanced MR pulse sequences shown to increase lesion detection and classification accuracy to examine this hypothesis.

Materials and Methods

Forty patients recruited were evaluated for evidence of CI within the year prior to the MRI scan by a neuropsychologist with extensive experience in MS. The subjects were drawn from an existing database at Baylor Neuropsychologists, Department of Neurology, Baylor College of Medicine Houston, TX. Recruitment extended from 2005 to 2008. A full battery of neuropsychological (NP) tests was performed in 40 MS patients (one later had to be excluded due to poor MR image quality). The NP tests results were then correlated with lesion number independently by type (IC, MX and JX). We also evaluated the correlation of CI with the sum of IC + MX which we called cortical lesions (CL). JX lesions were not considered to be “cortical” but their impact on cognition was also evaluated both independently and within the category total lesions (IC + MX + JX). Other correlations included the number of lesions by type with individual neuropsychological test scores and correlations between CI and measures of brain atrophy using Normalized CSF (nCSF), the expanded disability status scale (EDSS) and disease duration (DD).

Neuropsychological assessment

A comprehensive battery of standardized psychological tests that evaluated a broad range of neurocognitive functions was administered. This included (listed by cognitive domains evaluated):

Memory Functions

Wechsler Memory Scale-R (WMS-R) and Rey Auditory Verbal Learning Test (RAVLT).

Intelligence and Cognitive Functions

Wechsler Adult Intelligence Scale III (WAIS-III).

Language Functions

Aphasia Screening Test and Controlled Oral Association (COWA).

Motor Functions

Finger Tapping Test; (FTT), Grooved Pegboard Test (GPV), and Hand Dynamometer Gripped Strength Test (GPT).

Attention and Tracking Functions

Paced Auditory Serial Addition Test (PASAT), Stroop Neuropsychological Screening Test, and Trail Making Test.

Executive Functions

Wisconsin Card Sorting Test (WCST).

Emotional Functioning

Minnesota Multiphasic Personality Inventory

From the various standard scores obtained, the data was interpreted using the following standard scores of cognitive functions usually impacted in MS:

• WAIS-III – Working Memory Index and Processing Speed of Information Index.

• WMS – R – General Memory Index and Delayed Recall Index.

• WCST - % errors; 5 perseverative responses; % of perseverative errors; % of non-perseverative errors and % conceptual level responses.

• COWA

• GPT dominant and non-dominant hand performance.

• FTT dominant and non-dominant hand performance.

• MMPI-2 Depression Scale

• PASAT

The clinical interpretation of the neuropsychological data for each patient was based on an integrated test pattern analysis approach that takes into account standard test scores for each patient accounting for their specific demographics and medical data. The results for each patient were assessed relying on normative data and standard scores that facilitate the identification of patterns of scores consistent with deficits in test performance associated with brain dysfunction [28].

Impairment was defined as patterns of impairment in two or more tests in comparison with established normative data. Individual test results were classified as normal if the standard score did not deviate from the normative data. Borderline impairment was classified as below 1 standard deviation (SD) from the normative data. Mild impairment was classified as 1.5 SD below the normative data. Moderate impairment was classified as 2 SD below the mean. Severe impairment was classified as 3 SD from the mean normative data. The specific classification of each patient was made by a board certified neuropsychologist with extensive experience in the evaluation of cognitive functions in patients with MS. Since CI in MS is generally not homogeneous and there is great variability from patient to patient, the neuropsychologist used his clinical judgment in classifying each patient following the criteria established.

Inclusion/exclusion criteria

Patient eligibility was established based on NP test results and criteria as follows:

Inclusion criteria

1) clinically definite relapsing remitting (RR), secondary progressive (SP) or primary progressive (PP) MS by McDonald criteria [29], 2) NP evaluation performed within one year from enrollment, 3) age 18 to 50 (to avoid confounding effects of age related dementia), 4) EDSS score 6.5 or less, 5) brain MRI consistent with the clinical diagnosis of MS.

Exclusion Criteria

1) contraindications for MRI (pace makers, metal implants, claustrophobia etc.), 2) history of clinical relapse or a high dose steroid treatment in the three months prior to study, 3) history of drug or alcohol abuse, 4) acute or uncontrolled depression within three months from testing, and 5) history of cytotoxic or immunosuppressive treatments.

Once eligibility was established, recruitment was done by telephone call or invitation letter. Eligible patients were referred for one imaging session. For quality assurance, patient eligibility (by inclusion/exclusion criteria and standard MRI findings) was confirmed prior to enrollment. Informed consent was obtained prior to imaging session. All personnel aside from the neuropsychologist were blind to the subject’s cognitive status. The neuropsychologist was blind to the MRI findings. Each patient enrolled attended a two hour session that included: 1) collection of demographics and disease history, 2) neurological exam, 3) EDSS score assessment, and 4) MRI of the brain without gadolinium.

MR Imaging

Each patient underwent an MRI of the brain according to a standard protocol that included the optimized sequences DIR/PSIR [30, 31] for detection of CL. MRI scans were performed on a Philips 3T Intera scanner with a Quasar Plus gradient system (maximum gradient amplitude 80 mT/m, slew rate 200 T/m/s) and a six channel SENSE-compatible head coil. Following the tripilot scan for locating the brain region, images were acquired with parameters as listed in Table 1. The entire MRI protocol was completed in less than one hour.

Table 1.

MRI acquisition parameters

Image acquisition parameters for MRI pulse sequences in the MS protocol
Sequence	Plane	TR msec	TE msec	TI msec	Image matrix	FOV mm	Slice mm	SENSE
PSIR	axial	4300	13	400	256×256	240	3	no
DIR	axial	15000	25	3400/325	512×512	240	3	2
T1W	axial	600	9.2	-	256×256	240	3	no
Dual-echo FSE	axial	6800	10/90	-	256×256	240	3
FLAIR	axial	10,000	80	2600	256×256	240	3	2

PSIR= phase sensitive inversion recovery, DIR= double inversion recovery, W= weighted, FSE = fast spin echo, FLAIR= fluid attenuated inversion recovery

Quantitative MRI Analysis

All MRI analysis, including segmentation for tissue quantization was performed using the software developed in-house [32]. In brief, all images acquired in different series (FLAIR, dual echo, T1-images etc.) were retrospectively aligned [33]. Images were stripped of the extrameningeal tissues [34], filtered using the anisotropic diffusion filter [35], corrected for the bias field [36] and intensity normalized [37]. Segmentation of GM, WM, and lesion volume were based on the dual echo and FLAIR, images using the hidden Markov random field – expectation maximization (HMRF-EM) algorithm and Parzen window classification [32]. Normalized CSF (nCSF) volume was used as the measure of whole brain atrophy [38].

Lesion detection and classification

CL were identified on DIR and validated on PSIR as previously detailed [23]. For classification purposes an MRI specific lesion classification was created: lesions were defined as (a) IC if total confinement within the cortical ribbon was seen, (b) MX if they were present mostly in the cortex but with some (~25%) subcortical extension and (c) JX if mostly present in the subcortical region with clear extension (~25%) into the cortex. Lesion classification was done without knowledge of the neuropsychological test results (Figure 1).

Statistical Analyses

Poisson regression models were first used to evaluate for evidence of an association between CI and each different type of lesions. Negative Binomial regression was preferred as it explained the variation in the number of lesions better than the Poisson model. Since there were some patients with no lesions the Negative Binomial regression approach allowed for outcomes (lesions) with a large number of zero counts and for adjustment of covariates. Given the small sample size of 39 patients, the CI categories (normal, borderline, mild, moderate and severe) were collapsed into two groups: normal and impaired.

In our first analysis the normal group included only normal subjects, while the subjects in the borderline category were included in the impaired group. Each group was evaluated for correlations between each type of lesion individually (IC, MX, JX), with CL only (CL= IC + MX) and with total lesions (IC+MX+JX). A second analysis was done with the borderline group classified as “normal” for comparison.

Results

A total of 238 CL (130 IC, 108 MX) were visualized on DIR/PSIR in 82% of the evaluable patients; 39 JX lesions were also identified. The proportion of patients with 0 lesions in all of the above categories was 20%. The MS patients were classified into four CI categories as follows: 11 as normal, 9 as borderline, 11 as mild, 7 as moderate and 1 as severe. Total lesion number grouped by degree of impairment showed: normal 32, borderline 38, mild 115, moderate 85 and severe 7 (Table 2).

Table 2.

Number of lesions by type and degree of cognitive impairment

CI. Category	IC lesions	MX lesions	JX lesions	Total lesions
Normal (11)	21	5	6	32
Borderline (9)	22	10	6	38
Mild (11)	58	44	13	115
Moderate (7)	25	48	12	85
Severe (1)	4	1	2	7
Total (39)	130	108	39	277

CI= cognitive impairment, IC= intracortical, MX= mixed; mainly cortical with clear but minimal extension into the WM, JX= juxtacortical; mainly subcortical with clear extension into the cortex.

Based on Negative Binomial regression, correlations between CI and MX, CL (IC + MX), and total lesions (IC + MX + JX) were all significant. There were no important associations seen between IC or JX lesions and CI, independently as seen in table 3.

Table 3.

Cognitive impairment and lesions by type

COGNITIVE IMPAIRMENT (mild, moderate, severe and borderline) vs. normal
Intracortical (IC)	[w=1.80; p=0.179]
Mixed (MX)	[w=6.71; p=0.010]
Juxtacortical (JX)	[w=1.34; p=0.247]
Cortical lesions (IC + MX)	[w=4.70; p=0.030]
Total lesions (IC + MX + JX)	[w=4.48; p=0.034]

w=Walds Chi-Square statistic; p=P-value

We also considered the fact that patients meeting criteria for the borderline category could potentially be misclassified as “impaired” when functionally they were closer to the “normal” group. Therefore, a second analysis was done this time classifying the borderline patients under the “normal” category. When the borderline cases were incorporated in the normal group, as seen in the table below (Table 4), all correlations strengthened and were similar in significance to the first analysis.

Table 4.

Cognitive impairment and lesions by type 2^nd Analysis

COGNITIVE IMPAIRMENT (mild, moderate, severe) vs. (borderline, normal)
Intracortical (IC)	[w=2.71; p=0.100]
Mixed (MX)	[w=8.76; p=0.003]
Juxtacortical (JX)	[w=2.45; p=0.118]
Cortical (IC + MX)	[w=6.67; p=0.010]
Total lesions (IC + MX + JX)	[w=6.47; p=0.011]

Also found was a correlation between severity of CI and EDSS (p = 0.009), but none between CI and DD or CI and nCSF (as a measure of brain atrophy) independently (Table 5).On a different analysis of correlations between CI and each component of the NP evaluation a correlation was seen between number of MX and JX lesions and the PSI component of the WAIS test (p = 0.039, p = 0.026). This test measures information processing speed.

Table 5.

Correlations between cognitive impairment and EDSS, disease duration and nCSF as a measure of atrophy.

	N	r	p
Disease Duration	37	0.01856	0.9132
EDSS	38	0.41783	0.0090
nCSF	35	−0.18503	0.2873

Discussion

This study compared the presence, number and type of cortical and JX lesions in MS subjects with different degrees of CI to those MS subjects with unimpaired or normal cognition based on an extensive NP evaluation. The NP battery of tests was interpreted by the same experienced neuropsychologist over the duration of the study, without any knowledge of the MRI findings. A combination of DIR and PSIR imaging sequences was used for accurate lesion detection/classification. Statistically significant correlations between MX as well as between CL (MX + IC) and CI were found. This association did not change when the JX lesions were added (IC+MX+JX). While our working hypothesis was that careful lesion classification between purely IC, MX and JX lesions using advanced MRI techniques might help isolate the impact of IC lesions on the presence of associated CI, its impact was found to be of less importance than the presence of MX lesions. The potential explanation for this may be that MX lesions tend to be larger in size. Another consideration is the fact that despite the improved detection capabilities of these advanced techniques we are still not capturing all IC lesions present.

Another goal of the study was to determine if correlations existed between the degree of CI and measures of DD, EDSS and brain atrophy. In this study, the only statistically significant correlation that was found was with disease severity as measured by EDSS. This is consistent with the findings from other studies [39].

Strong evidence exists of a relationship between cognitive dysfunction and ventricular enlargement in patients with MS [39]. A relationship between significantly smaller parenchymal brain volumes and brain parenchymal fractions has also been observed in patients with CI compared with those who are cognitively intact [40]. As in other areas of MS research however, results are conflicting. In our study, atrophy alone was shown to be less important as a contributing factor to CI. It is possible that the relatively small subject cohort that was studied may explain this apparent discrepancy in our findings. The statistically significant correlations seen between CL and specific tests like the PSI component of the WAIS test support that information processing speed is more commonly affected by MS related CI.

A possible limitation of this study includes small cohort size and the significant variability in the number of CL found in these patients. The variability of CL presence between MS subjects is quite consistent with previous post mortem histopathological studies by Brownwell and Hughes [41]. In our study there was clearly a significant increase in the total number of IC lesions in the patients with mild and moderate CI (115 and 85 respectively) when compared with the borderline and normal (32 and 38), based on this the low correlation between purely IC lesions and CI was somewhat unexpected. We suspect that the variability in IC lesion number between patients interfered with our ability to establish stronger correlations. The fact that a stronger correlation between CI and MX lesions was seen is not completely unexpected since IC lesions tend to be smaller. The strong correlation found between CL (IC + MX) and CI could also be explained by the fact that it represents a larger area of cortical damage which includes interconnecting neurons present in the gray/white matter junction. Interestingly JX lesions also tend to be larger in size when compared to IC but no clear correlation was observed with these lesions characterized for minimal invasion of the cortical ribbon.

Our findings strongly suggest that CL both IC and MX play a more significant role in cognitive dysfunction than JX lesions and measures of atrophy. More importantly, if our observations can be confirmed in a larger cohort in a prospective study, MRI could potentially be used as a biomarker for diagnosis, and management of CI in MS. Larger studies are needed to confirm these observations as are trials to evaluate the effect of immunomodulators on cortical lesion behavior and CI. Further efforts to advance our understanding of the pathophysiological basis of cognitive dysfunction in MS need continued refinement and future application of combined advanced imaging techniques such as MTR, DTI that may shed more light on the importance of the integrity of neural networks to the preservation cognitive function in MS are currently underway at our institution as are functional MRI (fMRI) studies for CI evaluation.