Lesion size and shape in central vein sign assessment for multiple sclerosis diagnosis: An in vivo and postmortem MRI study

Abstract

Background:

The “central vein sign” (CVS), a linear hypointensity on T2*-weighted imaging corresponding to a central vein/venule, is associated with multiple sclerosis (MS) lesions. The effect of lesion-size exclusion criteria on MS diagnostic accuracy has not been extensively studied.

Objective:

Investigate the optimal lesion-size exclusion criteria for CVS use in MS diagnosis.

Methods:

Cross-sectional study of 163 MS and 51 non-MS, and radiological/histopathological correlation of 5 MS and 1 control autopsy cases. The effects of lesion-size exclusion on MS diagnosis using the CVS, and intralesional vein detection on histopathology were evaluated.

Results:

CVS+ lesions were larger compared to CVS− lesions, with effect modification by MS diagnosis (mean difference +7.7 mm³, p = 0.004). CVS percentage-based criteria with no lesion-size exclusion showed the highest diagnostic accuracy in differentiating MS cases. However, a simple count of three or more CVS+ lesions greater than 3.5 mm is highly accurate and can be rapidly implemented (sensitivity 93%; specificity 88%). On magnetic resonance imaging (MRI)-histopathological correlation, the CVS had high specificity for identifying intralesional veins (0/7 false positives).

Conclusion:

Lesion-size measures add important information when using CVS+ lesion counts for MS diagnosis. The CVS is a specific biomarker corresponding to intralesional veins on histopathology.

Keywords

Biomarkers multiple sclerosis magnetic resonance imaging post mortem T2 lesions central vein sign

Introduction

Multiple sclerosis (MS) is characterized by immune-mediated demyelinating injury to the central nervous system (CNS) associated with dissemination in space and time based on established criteria.¹ However, misdiagnosis of MS remains a prevalent and challenging clinical problem.^2,3 This is in part attributable to the limited accuracy of currently available diagnostic criteria and a number of conditions that can mimic MS on magnetic resonance imaging (MRI). Despite subsequent revisions to the imaging criteria,^4–6 the accuracy of conventional imaging techniques in guiding the diagnosis of MS remains modest.^7,8

The “central vein sign” (CVS), an imaging finding that refers to the presence of a linear hypointensity corresponding to a central vein/venule within lesions on susceptibility-weighted imaging,⁹ has been described as a specific imaging biomarker for MS lesions.¹⁰ Several studies have shown that the CVS can accurately differentiate MS lesions from other mimicking conditions, even in cases where widely used MS imaging diagnostic criteria are met.^10–12 Early reports described a 40% threshold of CVS+ lesions as a specific diagnostic marker for MS,¹⁰ with more recent studies demonstrating that an absolute count of CVS+ lesions can also achieve high accuracy for MS diagnosis while reducing analysis time needed to rate all CVS-eligible lesions.^13,14

To standardize the application of the CVS in practice, the North American Imaging in Multiple Sclerosis Cooperative (NAIMS) put forth recommendations for the radiological evaluation of the CVS for MS diagnosis.¹⁵ Based on consensus expert opinion, a set of criteria were proposed, among which was the exclusion of lesions <3 mm in diameter in any plane. In contrast, guidelines from the Magnetic Resonance Imaging in Multiple Sclerosis (MAGNIMS) group have recommended that lesions included in MS evaluation should be 3 mm or more in at least one plane, regardless of the CVS, thereby retaining more ovoid lesions whose maximum diameter surpasses 3 mm.⁶ Since MRI with isotropic resolution of 1 mm³ or finer is now routinely possible on clinical scanners, it is important to undertake a systematic analysis of the differences between those two paradigms, the effect of lesion-size exclusion on the CVS in MS diagnosis, and its impact on CVS percentage measurements.

In this study, we examine the impact of different lesion-size exclusion thresholds applied in two different paradigms (excluding lesions if they are less than a specific threshold in any or largest dimension) in a large cohort of MS, healthy control (HC), and non-MS subjects with MS-mimicking neurological disorders. We describe the impact of these selection paradigms and associated thresholds on: (1) CVS percentage measurement, (2) sensitivity/specificity of MS diagnosis, using previously described benchmarks and varying thresholds,^10,13 and (3) presence of veins on postmortem MRI using histopathological confirmation in a number of MS and non-MS autopsy cases.

Methods

Participants

Study participants were recruited under the National Institutes of Health (NIH) natural history studies of MS (ClinicalTrials.gov number, NCT00001248) and human T-lymphotropic virus type-I-associated myelopathy/tropical spastic paraparesis (HAM/TSP; ClinicalTrials.gov number, NCT00001778). Both study protocols were approved by the NIH institutional review board, and written, informed consent was obtained from all study participants.

MS patients, HCs, and patients with conditions mimicking MS on MRI were included from the cohort of patients seen at the National Institute of Neurological Disorders and Stroke (NINDS) Neuroimmunology Clinic between 1 January 2012 and 7 January 2019. MS diagnosis was confirmed by the evaluating neurologist based on the 2010 revised McDonald criteria.⁴ Disease subtype was classified as relapsing-remitting multiple sclerosis (RRMS), secondary progressive multiple sclerosis (SPMS), or primary progressive multiple sclerosis (PPMS).¹⁶ Patients were excluded if they had (1) an unclear diagnosis or (2) lack of dedicated T2*-weighted imaging enabling CVS assessment. A control cohort of subjects without MS was used to inform specificity measures, which included a mix of HCs with at least one T2/FLAIR hyperintense lesion on MRI, and patients with MS radiological mimics: neuromyelitis optica spectrum disorders (NMOSDs), Susac syndrome, cerebral small vessel disease (CSVD), and HAM/TSP.

In vivo MRI

Participants included in the in vivo study underwent 3T brain MRI on either a Skyra scanner (Siemens, Erlangen, Germany) equipped with a body transmit and a 32-channel receive coils, or an Achieva scanner (Philips Medical Systems, Best, The Netherlands) equipped with an 8- or 32-channel receive head coil. Optimized T2*-weighted, whole brain, three-dimensional (3D)-segmented echo-planar imaging (repetition time (TR), 64 ms; echo time (TE), 35 ms; flip angle (FA), 10°; echo-train length, 15; 0.65 mm isotropic voxels for Siemens, and 0.55 mm for Philips)¹⁷ was used to perform CVS evaluation. 3D T1-weighted magnetization-prepared rapid acquisition of gradient echoes (T1-MP2RAGE for Siemens and T1-MPRAGE for Philips) and 3D T2-weighted fluid-attenuated inversion recovery (FLAIR) images were acquired as described in detail previously.^18,19 Routine MRI preprocessing was performed and details are provided in the Supplement.

Standardized CVS rating was performed on all lesions by a single rater (O.A.) masked to the clinical status of the cases, as described in detail previously,²⁰ and using a novel specialized workflow displaying each separate segmented lesion for rating while blinding the rater to all other lesions in the scan volume. The implementation details and code for the CVS rating workflow are publicly available: https://github.com/omarallouz/CVS_rater. Confluent lesions and lesions with multiple or eccentric veins were excluded from the analysis.¹⁵

To calculate lesion-size measurements, lesions were first extracted as clusters of connected voxels using MATLAB, version R2019a (The MathWorks Inc., Natick, MA, USA), and their principal axes lengths (PALs) were calculated in all three dimensions (defined as the lengths of the major axes of an ellipsoid with the same normalized second central moments). Manual lesion dimension measurements were obtained on 30 lesions (90 measurements) from a random sample of 10 subjects to assess consistency between manual and automated measurements. Two paradigms of lesion exclusion were compared: (1) excluding lesions if any dimension is less than a particular threshold (similar to the original 2016 NAIMS criteria) or (2) excluding lesions if their largest dimension is less than a particular threshold, which favors the inclusion of more ovoid lesions with one larger dimension.

To measure the relative ovoid extent of lesions regardless of their spatial orientation, we calculated the fractional anisotropy of the lesion dimensions (FALDs), similar to a diffusion ellipsoid in diffusion tensor imaging²¹ and defined as follows

FALD = \sqrt{\frac{3}{2}} \sqrt{\frac{{({PAL}_{1} - {PAL}_{avg})}^{2} + {({PAL}_{2} - {PAL}_{avg})}^{2} + {({PAL}_{3} - {PAL}_{avg})}^{2}}{({PAL}_{1}^{2} + {PAL}_{2}^{2} + {PAL}_{3}^{2})}}

(1)

where PAL_i represents the respective PALs of the lesion as defined previously, and PAL_avg is their average length. FALD values closer to zero describe lesions that are more spherical in shape, whereas values closer to one describe extremely ovoid lesions.

Human autopsies

A total of five MS cases (mean age, 62 years; SD 4; 4 women) who had long-standing, clinically progressive disease at time of death underwent standard autopsies and were included in the analysis to inform CVS imaging-histopathology correlations (Table 2). Human brains were attained at autopsy after obtaining consent from the next of kin. Brain tissue from one additional autopsy case without MS, who died due to small cell lung cancer and had evidence of non-specific T2-FLAIR hyperintense lesions remote from known brain metastatic deposits, was obtained to serve as a matched control (age 48 years; male; Table 2). Details of the postmortem brain MRI acquisition and targeted neuropathological analysis are described in detail in the Supplement.

Statistical analysis and comparison metrics

Statistical analyses were performed using Stata (version 17; StataCorp LP, College Station, TX, USA). Comparisons between groups were performed using the Wilcoxon rank-sum test for continuous variables (age, lesion number, and CVS percentage), and chi-square or Fisher’s exact test for categorical variables (sex, race, and comorbidities). Bland–Altman plots were generated to examine agreement between manual and automated lesion-size measurements, and intraclass correlation coefficients (ICCs) were calculated using a two-way mixed-effects model accounting for fixed effects of rating (manual vs automated) and random-effects of lesions.

Comparisons of lesion characteristics were conducted using mixed-effects linear regression with subject-level intercepts accounting for within-subject correlation, and allowing adjustment for subject- and lesion-specific covariates. Interaction terms of CVS status by MS diagnosis were introduced to study whether the relationship between lesion size and CVS status was effect-modified by the underlying diagnostic category. Sensitivity and specificity measures were calculated for each lesion exclusion paradigm at set lesion-size thresholds. Youden’s indices, defined as (sensitivity + specificity − 1) × 100, were calculated for each permutation to examine MS diagnostic accuracy based on absolute CVS+ lesion number or CVS percentage cut-off, and varying lesion-size cut-offs in mm. Statistical significance was defined as p < 0.05.

Results

Participant characteristics

The in vivo cohort comprised 163 MS cases (102 RRMS, 28 SPMS, and 33 PPMS) and 51 non-MS cases (16 HCs, 1 NMOSD, 2 Susac syndrome, 12 CSVD, and 20 HAM/TSP). A total of 11,174 lesions were evaluated in the MS group and 1773 lesions in the non-MS group. There were no significant differences between the MS and non-MS groups with respect to demographic and clinical variables, except for African-American race, which was more frequent in the non-MS group (Table 1). A higher number of lesions were observed per subject in the MS compared to non-MS groups (median 55 vs 16 lesions, respectively, p < 0.001).

Table 1.

Demographic and clinical characteristics of the study population.

	MS cases (n = 163)	Non-MS cases (n = 51)	p-value
Age, years, mean (SD)	48.7 (11)	50.2 (13)	0.39^a
Sex, female, n (%)	37 (73)	101 (62)	0.17^b
Race, n (%):
Caucasian	134 (82)	23 (45)	<0.001 ^c
African American	22 (14)	23 (45)
Hispanic	2 (1)	2 (4)
Asian/Pacific Islander	1 (1)	1 (2)
Other	4 (2)	2 (4)
Diagnosis subtype, n (%):	RRMS: 102 (63) SPMS: 28 (17) PPMS: 33 (20)	Healthy control: 16 (31) NMOSD: 1 (2) Susac syndrome: 2 (4) CSVD: 12 (24) HAM/TSP: 20 (39)	–
Disease duration, years, median (Q1–Q3)^d	13.5 (6–20)	–	–
Total lesion number per subject, median (Q1–Q3)^e	55 (32–95)	16 (8–40)	<0.001 ^a
CVS percentage, median (Q1–Q3)^e	74.8% (64%–84%)	11.5% (0%–22%)	<0.001 ^a
Comorbidities, n (%)
Hypertension	43 (26)	17 (33)	0.34^b
Diabetes	10 (6)	1 (2)	0.47^c
Dyslipidemia	56 (34)	16 (31)	0.69^b
Coronary artery disease	3 (2)	1 (2)	1.00^c
Smoking, active^f	13 (9)	1 (2)	0.12^c
Migraine	24 (15)	7 (14)	0.86^b

CSVD: cerebral small vessel disease; HAM/TSP: HTLV-1 associated myelopathy/tropical spastic paraparesis; NMOSD: neuromyelitis optica spectrum disorder; PPMS: primary progressive multiple sclerosis; RRMS: relapsing-remitting multiple sclerosis; SD: standard deviation; SPMS: secondary progressive multiple sclerosis.

Wilcoxon-rank-sum test.

Pearson’s chi-square test.

Fisher’s exact test.

Not applicable for healthy controls.

Prior to the application of lesion-size exclusion criteria.

Smoking data missing in 29 subjects.

Bold values denote statistical significance at the p < 0.05 level.

Automated lesion size is consistent with manually determined measurements

Comparisons between manually determined and automated lesion dimension measurements were performed on 30 lesions (90 measurements) from 10 cases (7 MS, 3 non-MS) randomly sampled from the overall study population. Excellent agreement was observed between both sets of measurements (intraclass correlation coefficient (ICC) = 0.98; 95% confidence interval (CI) = 0.97–0.99; p < 0.001). Bland–Altman analyses revealed that the 95% limits of agreement of the lesion dimension differences fell between ±0.6 mm, without bias (Supplementary Figure 1, Panel A). As a result, a 0.5-mm increment in successive threshold sizes was utilized throughout the rest of the analyses.

MS status effect modifies the relationship between the CVS and lesion size

Rated lesions were divided into three groups according to the NAIMS 2016 guidelines: CVS+, CVS−, and excluded lesions.¹⁵ Discrete lesions eligible for analysis were more likely to be larger in the MS group compared to the non-MS group adjusting for age, sex, and race (Supplementary Table 1, and Figure 1(b); mean volume difference 13.8 mm³; 95% CI = 10–17; p < 0.001). CVS+ lesions were also more likely to be larger in volume compared to their CVS− counterparts after similar adjustment (Supplementary Table 1 and Figure 1(b); mean volume difference 13.1 mm³; 95% CI = 12–15; p < 0.001).

Interestingly, the MS group status effect-modified the relationship between CVS+ lesions and the increase in lesion volume (mean increase in lesion volume for CVS+ lesions in MS vs non-MS 7.8 mm³, p-value for interaction = 0.004), indicating that the observed increase in lesion volume with CVS+ status is more pronounced in the context of MS pathology. Lesions in MS patients similarly had higher FALD values compared to the non-MS group, indicating that MS lesions are more likely to be ovoid rather than spheroid (mean FALD difference = 0.10; 95% CI = 0.09–0.12; p < 0.001). Similarly, CVS+ lesions were also more likely to be ovoid than CVS− lesions (mean FALD difference = 0.04; 95% CI = 0.02–0.06; p < 0.001).

Impact of lesion-size exclusion on CVS percentage and MS diagnostic accuracy

Of the 5503 CVS+ lesions in the MS group, 4580 (83%) met the exclusion criteria if any dimension was less than 3 mm, which was reduced to 1573 (29%) when exclusion criteria were limited to the largest dimension only (p < 0.001). When excluding smaller lesions, a higher cumulative proportion of CVS− lesions were removed, resulting in a relative increase in proportion of retained CVS+ lesions and CVS percentage (Figure 1(a)). This effect was less pronounced when exclusion was restricted to the largest dimension of the lesion, driven by increased retention of more ovoid lesions (Figure 1(b)).

To systematically examine the effects of lesion-size exclusion on MS diagnostic accuracy, lesions were excluded at successive lesion-size thresholds ranging from 0 to 5 mm (Figure 2(a)–(d); Supplementary Tables 2–3). Maximum Youden’s indices were calculated at the intersection of each absolute threshold value with varying CVS+ lesion counts ranging from 1–10 (Figure 2(a) and (c)), or CVS percentage thresholds (Figure 2(b) and (d)). Overall, CVS percentage thresholds had higher maximum Youden’s indices for MS diagnostic accuracy compared to CVS+ lesion counts. Among CVS+ lesion counts, the presence of three or more CVS+ lesions greater than 3.5 mm in the largest dimension per subject attained the maximum Youden’s index for MS diagnosis (Figure 2(c)). When using CVS percentage measurements in MS diagnosis, there was no significant improvement in diagnostic accuracy when excluding lesions based on their size (Figure 2(b) and (d)).

Postmortem MRI-histopathology correlations of the CVS

A total of 26 lesions from six autopsy cases (5 MS and 1 control) were selected based on CVS characteristics observed using postmortem MRI (Table 2). No CVS+ lesions were visualized on postmortem MRI in the control case. Lesions were prospectively rated as either CVS+, CVS+ but with poorly visible or inconspicuous vein (probable CVS+), or CVS− on postmortem 7T MRI. Following histopathological assessment, lesions were categorized into four lesion types based on the combination of imaging and histopathological CVS characteristics:

CVS+ on both MRI and histopathology (9/26; 35%; Figure 3(a)).

Probable CVS+ lesions with inconspicuous veins on MRI, all of which had a central vein observed on histopathology (2/26; 8%; Figure 3(d)).

CVS− on both MRI and histopathology (7/26; 27%; Figure 3(b)).

CVS “converter” (CVSc) lesions, defined as having no clearly visible central vein on MRI, but with one detected on histopathology (8/26; 31%; Figure 3(c)).

Table 2.

Demographic and clinical characteristics of human autopsy patients.

Case	Age (years)	Sex	Diagnosis	Post-mortem interval (h)	In vivo MRI	PMI (h)	Cause of death	Total lesions sampled^a	CVS+ lesions sampled^a	CVS− lesions sampled^a	Probable CVS+ lesions sampled^a
MS 1	60	F	SPMS	7.5	N	7–8	Sepsis	5	1	4	0
MS 2	61	F	SPMS	13	N	13	Late-stage MS with significant disability	6	2	3	1
MS 3	60	F	PPMS	8	Y	8	Polymicrobial bronchopneumonia	9	4	4	1
MS 4	59	M	PPMS	9	Y	9	Brainstem stroke	3	2	1	0
MS 5	68	F	PPMS + PML	19	Y	19	Infratentorial PML	1	0	1	0
Control	48	M	Small cell lung carcinoma	6	N	6	Small cell lung carcinoma	2	0	2	0

PMI: post-mortem interval; SPMS: secondary progressive multiple sclerosis; PPMS: primary progressive multiple sclerosis; PML: progressive multifocal leukoencephalopathy.

Lesion ratings displayed were based on prospective ex vivo postmortem MRI characterization of the lesions prior to histopathological examination.

No false positive CVS lesions were observed, that is, lesions with a central vein on MRI but no central vein on histopathology, suggesting excellent biological specificity for the CVS in detection of intralesional veins confirmed on histopathology. Of the 24 lesions assessed in the MS cases, 21/24 (87.5%) were chronic inactive and 3/24 (12.5%) were chronic active (1 CVS+, 1 probable CVS+, and 1 CVSc). Volume analyses revealed that lesions that were CVS+ or probable CVS+ (mean volume 34.7 mm³) on both MRI and histopathology were larger than CVS− (8.8 mm³; p = 0.005) and CVSc lesions (5.0 mm³; p = 0.001) and had a longer maximal dimension than CVSc (mean length 5.6 and 2.4 mm, respectively; p = 0.001) but not CVS− lesions (mean length 4.1 mm; p = 0.12; Supplementary Table 4). All other volume and maximal dimension pairwise comparisons were not statistically significant.

Discussion

The exclusion of lesions based on their size in MS evaluation has been a prevalent aspect of practice that is proposed as part of many international radiological consensus recommendations.^15,22,23 However, a few studies have systematically analyzed the impact and optimal thresholds for lesion-size exclusion in practice. In this study, we sought to evaluate whether the paradigm of lesion exclusion (setting thresholds based on any or the largest dimension), as well as the specific absolute thresholds associated with each paradigm, improve the accuracy of MS diagnosis and its differentiation from other MS-mimicking conditions using the CVS.

Similar to prior studies,^13,24–26 MS patients had a significantly higher number and proportion of CVS+ lesions compared to MS-mimicking conditions. CVS+ lesions are more likely to be larger in volume compared to CVS− lesions and, interestingly, this difference appears to be effect-modified by MS status, suggesting that the perivenular inflammation of MS drives a greater extent of signal change around CVS+ lesions compared to those in controls where CVS+ lesions may have formed stochastically around veins and have sizes similar to CVS− ones. Previous studies demonstrated that MS lesions are more likely to take on an ovoid shape, which is not surprising since the veins at the center of these lesions are linear.^27,28 Our results expand on this work by showing that CVS+ lesions are more ovoid, as defined by the FALD measurements, than their CVS− counterparts, and this difference is greater in MS vs non-MS cases. It would be interesting to explore in future work whether simple information gleaned from lesion shape characteristics could be an independent predictor of their CVS status. This is particularly valuable in retrospective analysis of data sets in which high-resolution T2*-weighted imaging of the CVS is not available.

One intriguing finding in this study is the impact of excluding lesions on the overall CVS percentage measurements in individual subjects (Figure 1(a) and (b)). Due to the smaller size of CVS− lesions, excluding lesions less than 3 mm in any dimension tends to disproportionately remove CVS− lesions more so than CVS+ ones, resulting in an upward drift of the CVS percentage measurements. This can be mitigated by excluding lesions based on their largest dimension instead (Figure 1(b)). This is in line with previous empirical recommendations of measuring lesions in their longest axis when evaluating MS lesions.²³

While previous investigations have described criteria for CVS use in MS diagnosis based on either CVS percentage or absolute CVS+ lesion counts,^10,13,29 very few studies have systematically compared the performance of both at various lesion-size cut-off values. We found that CVS percentage-based criteria, overall, had a higher diagnostic accuracy in differentiating MS cases based on maximum Youden indices (Figure 2). In practice, however, CVS percentage measurements require the rating of all lesions in a given MRI scan and, consequently, are time-consuming in the absence of an automated image-processing algorithm.^30,31 Absolute counts of CVS+ lesions are faster to perform clinically and our work demonstrates that the presence of at least three CVS+ lesions greater than 3.5 mm in largest dimension is associated with the highest accuracy for MS diagnosis when considering various size thresholds (sensitivity 93% and specificity 88%). Importantly, using imaging-histopathological correlation, no false-positive CVS lesions were noted in this study, i.e. lesions that are CVS+ on MRI but without a central vein on histopathology, thereby providing confirmation of the excellent biological specificity of the CVS on neuropathological analysis.

It is worthwhile to note a few limitations of the work presented. The MRI techniques used to visualize the CVS were different between the in vivo and ex vivo portions of this study (3D-segmented echo-planar vs multi-echo gradient recalled echo imaging, respectively).^17,32,33 Furthermore, it is important to highlight the limitations of visualizing the CVS on postmortem MRI given the lack of dynamic blood flow, and wash-out of the deoxygenated hemoglobin signal with formalin fixation. For our in vivo analysis, we evaluated one form of susceptibility-based MRI imaging; therefore, these results are not readily generalizable to other forms of susceptibility-based MRI. Investigating the impact of different susceptibility-based imaging acquisition techniques on MS diagnostic accuracy is an important target for future research.

In summary, in this study, we have systematically examined the impact of lesion-size exclusion paradigms on diagnostic accuracy in a large cohort of MS and non-MS cases in vivo, and on the likelihood of detecting a ground truth central vein on histopathology using postmortem susceptibility-based imaging. Our analysis demonstrates that coupling lesion-size and shape information with the CVS can yield valuable insight into differentiating MS-specific pathology on imaging and optimizing its use in diagnostic evaluations. While CVS percentage measurements remain the gold standard for research studies and automated analysis, simple rules that are quick to implement, based on three or more CVS+ lesions greater than 3.5 mm in largest diameter, are viable alternatives with excellent accuracy for clinical workflow. Ultimately, these findings shed light on important pathobiological aspects of perivenular inflammation in MS and should help guide future modifications of any standardized criteria for implementation of the CVS in practice.

Supplemental Material

sj-docx-1-msj-10.1177_13524585221097560 – Supplemental material for Lesion size and shape in central vein sign assessment for multiple sclerosis diagnosis: An in vivo and postmortem MRI study

Supplemental material, sj-docx-1-msj-10.1177_13524585221097560 for Lesion size and shape in central vein sign assessment for multiple sclerosis diagnosis: An in vivo and postmortem MRI study by Omar Al-Louzi, Sargis Manukyan, Maxime Donadieu, Martina Absinta, Vijay Letchuman, Brent Calabresi, Parth Desai, Erin S Beck, Snehashis Roy, Joan Ohayon, Dzung L Pham, Anish Thomas, Steven Jacobson, Irene Cortese, Pavan K Auluck, Govind Nair, Pascal Sati and Daniel S Reich in Multiple Sclerosis Journal

Footnotes

Acknowledgements

The authors thank all the participants in this study as well as their family members. The authors also acknowledge the contributions of Frances Andrada and Jenifer Dwyer from the National Institute of Neurological Disorders and Stroke (NINDS) Neuroimmunology Clinic to the recruitment, care, and collection of clinical data from the study participants, as well as the NINDS neuroimmunology clinical fellows. The authors also thank Dr John Ostuni and the NINDS information technology department for help with servers and software maintenance, and the staff of the Functional Magnetic Resonance Facility (FMRIF) and radiology department technicians who were instrumental with regards to image acquisition. This study utilized the computational resources of the Biowulf system at the National Institutes of Health, Bethesda, MD, USA (). The authors also wish to thank the Human Brain Collection Core (HBCC) within the National Institute of Mental Health’s (NIMH) Division of Intramural Research Programs (DIRP) for their assistance with the autopsies and pathological specimen collection, as well as the National Cancer Institute Laboratory of Pathology.

Author Contributions

Dr O.A. contributed to the design and conceptualization of the study, data collection, data analysis, and drafting the original manuscript. S.M. contributed to the data collection, data analysis, and revising the manuscript for intellectual content. Dr M.D. contributed to the data collection, data analysis, and revising the manuscript for intellectual content. Dr M.A. contributed to the data collection, data analysis, and revising the manuscript for intellectual content. V.L. contributed to the data collection and revising the manuscript for intellectual content. B.C. contributed to the data collection and revising the manuscript for intellectual content. Dr P.D. contributed to the data collection and revising the manuscript for intellectual content. Dr E.S.B. contributed to the data collection and revising the manuscript for intellectual content. Dr S.R. contributed to the developing software methodology for data analysis and revising the manuscript for intellectual content. J.O. contributed to the data collection and revising the manuscript for intellectual content. Dr D.L.P. contributed to the developing software methodology for data analysis and revising the manuscript for intellectual content. Dr A.T. contributed to the data collection and revising the manuscript for intellectual content. Dr S.J. contributed to the data collection and revising the manuscript for intellectual content. Dr I.C. contributed to the data collection and revising the manuscript for intellectual content. Dr P.A. contributed to the design and interpretation of the histopathological analyses and revising the manuscript for intellectual content. Dr G.N. contributed to the design and conceptualization of the study, data collection, data analysis, and revising the manuscript for intellectual content. Dr P.S. contributed to the design and conceptualization of the study, data collection, data analysis, and revising the manuscript for intellectual content. Dr D.S.R. contributed to the design and conceptualization of the study, data collection, data interpretation, project supervision, and revising the manuscript for intellectual content.

Data Availability

The corresponding authors had full access to all data and accept responsibility for the decision to submit for publication.

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Dr O.A., S.M., Dr M.D., V.L., B.C., Dr P.D., Dr E.B., Dr S.R., J.O., Dr D.L.P., Dr A.T., Dr S.J., Dr I.C., Dr P.A., Dr G.N., and Dr P.S. have no disclosures pertaining to the work presented. Dr M.A. received speaker and/or consultancy fees from Celgene and Sanofi-Genzyme, unrelated to the current project. Dr D.S.R. received research funding from Abata, Sanofi-Genzyme, and Vertex, unrelated to the current project.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Intramural Research Program of the NINDS/NIH, the Adelson Medical Research Foundation, and the National Multiple Sclerosis Society (NMSS) (SI-2004-36590). O.A.-L. is the recipient of an NMSS-American Brain Foundation Clinician Scientist Development Award (FAN-1807-32163). D.L.P. is supported by an NMSS grant (RG-1907-34570). S.M. is supported by the NIH Undergraduate Scholars Program. V.L. is supported by the Foundation of the Consortium of Multiple Sclerosis Centers and the NIH Medical Research Scholars Program (a public–private partnership supported jointly by the NIH and contributions to the Foundation for the NIH from the Doris Duke Charitable Foundation, Genentech, the American Association for Dental Research, and the Colgate-Palmolive Company). E.S.B. is supported by a National Multiple Sclerosis Society Career Transition Fellowship (TA-1805-31038). M.A. is supported by the Conrad N. Hilton Foundation (grant no. 17313), the Cariplo Foundation (grant no. 2019-1677), the FRRB Early Career Award (grant no. 1750327), and by the International Progressive MS Alliance (PA-2107-38081).

ORCID iDs

Omar Al-Louzi

Martina Absinta

Erin S Beck

Pascal Sati

Daniel S Reich

Supplemental Material

Supplemental material for this article is available online.

References

McDonald

Compston

Edan

, et al. Recommended diagnostic criteria for multiple sclerosis: Guidelines from the International Panel on the diagnosis of multiple sclerosis. Ann Neurol 2001; 50(1): 121–127, http://www.ncbi.nlm.nih.gov/pubmed/11456302

Kaisey

Solomon

Luu

, et al. Incidence of multiple sclerosis misdiagnosis in referrals to two academic centers. Mult Scler Relat Disord 2019; 30: 51–56.

Solomon

Naismith

Cross

. Misdiagnosis of multiple sclerosis: Impact of the 2017 McDonald criteria on clinical practice. Neurology 2019; 92(1): 26–33.

Polman

Reingold

Banwell

, et al. Diagnostic criteria for multiple sclerosis: 2010 revisions to the McDonald criteria. Ann Neurol 2011; 69(2): 292–302.

Thompson

Banwell

Barkhof

, et al. Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria. Lancet Neurol 2018; 17(2): 162–173.

Filippi

Rocca

Ciccarelli

, et al. MRI criteria for the diagnosis of multiple sclerosis: MAGNIMS consensus guidelines. Lancet Neurol 2016; 15(3): 292–303.

Filippi

Preziosa

Meani

, et al. Prediction of a multiple sclerosis diagnosis in patients with clinically isolated syndrome using the 2016 MAGNIMS and 2010 McDonald criteria: A retrospective study. Lancet Neurol 2018; 17(2): 133–142.

Hyun

Huh

Kim

, et al. Evaluation of 2016 MAGNIMS MRI criteria for dissemination in space in patients with a clinically isolated syndrome. Mult Scler 2018; 24(6): 758–766.

Tan

Van Schijndel

Pouwels

PJW

, et al. MR venography of multiple sclerosis. AJNR Am J Neuroradiol 2000; 21(6): 1039.

10.

Tallantyre

Dixon

Donaldson

, et al. Ultra-high-field imaging distinguishes MS lesions from asymptomatic white matter lesions. Neurology 2011; 76(6): 534–539.

11.

Solomon

Schindler

Howard

, et al. “Central vessel sign” on 3T FLAIR* MRI for the differentiation of multiple sclerosis from migraine. Ann Clin Transl Neurol 2016; 3(2): 82–87.

12.

Kilsdonk

Wattjes

Lopez-Soriano

, et al. Improved differentiation between MS and vascular brain lesions using FLAIR* at 7 Tesla. Eur Radiol 2014; 24(4): 841–849.

13.

Solomon

Watts

Ontaneda

, et al. Diagnostic performance of central vein sign for multiple sclerosis with a simplified three-lesion algorithm. Mult Scler 2018; 24(6): 750–757.

14.

Mistry

Abdel-Fahim

Samaraweera

, et al. Imaging central veins in brain lesions with 3-T T2*-weighted magnetic resonance imaging differentiates multiple sclerosis from microangiopathic brain lesions. Mult Scler 2016; 22(10): 1289–1296.

15.

Sati

Constable

, et al. The central vein sign and its clinical evaluation for the diagnosis of multiple sclerosis: A consensus statement from the North American Imaging in Multiple Sclerosis Cooperative. Nat Rev Neurol 2016; 12(12): 714–722.

16.

Lublin

Reingold

. Defining the clinical course of multiple sclerosis: Results of an international survey. Neurology 1996; 46(4): 907–911.

17.

Sati

Thomasson

, et al. Rapid, high-resolution, whole-brain, susceptibility-based MRI of multiple sclerosis. Mult Scler 2014; 20(11): 1464–1470.

18.

Absinta

Sati

Masuzzo

, et al. Association of Chronic Active Multiple Sclerosis Lesions with Disability in Vivo. JAMA Neurol 2019; 76(12): 1474–1483.

19.

Absinta

Vuolo

Rao

, et al. Gadolinium-based MRI characterization of leptomeningeal inflammation in multiple sclerosis. Neurology 2015; 85(1): 18–28.

20.

Al-Louzi

Letchuman

Manukyan

, et al. Central vein sign profile of newly developing lesions in multiple sclerosis. Neurol Neuroimmunol Neuroinflamm 2022; 9(2): e1120.

21.

Basser

Pierpaoli

Microstructural and physiological features of tissues elucidated by quantitative-diffusion-tensor MRI. J Magn Reson B 1996; 111(3): 209–219, http://www.ncbi.nlm.nih.gov/pubmed/8661285 (accessed 2 November 2014).

22.

Sastre-Garriga

Pareto

Battaglini

, et al. MAGNIMS consensus recommendations on the use of brain and spinal cord atrophy measures in clinical practice. Nat Rev Neurol 2020; 16(3): 171–182.

23.

Filippi

Preziosa

Banwell

, et al. Assessment of lesions on magnetic resonance imaging in multiple sclerosis: Practical guidelines. Brain 2019; 142(7): 1858.

24.

Tallantyre

Brookes

Dixon

, et al. Demonstrating the perivascular distribution of MS lesions in vivo with 7-Tesla MRI. Neurology 2008; 70(22): 2076–2078.

25.

Kaisey

Solomon

Guerrero

, et al. Preventing multiple sclerosis misdiagnosis using the “central vein sign”: A real-world study. Mult Scler Relat Disord 2021; 48: 102671.

26.

Castellaro

Tamanti

Pisani

, et al. The use of the central vein sign in the diagnosis of multiple sclerosis: A systematic review and meta-analysis. Diagnostics 2020; 10(12): 1025.

27.

Okuda

Moog

McCreary

, et al. Utility of shape evolution and displacement in the classification of chronic multiple sclerosis lesions. Sci Rep 2020; 10(1): 1–8.

28.

Horowitz

Kaplan

Grewe

, et al. The ovoid lesion: A new MR observation in patients with multiple sclerosis. AJNR Am J Neuroradiol 1989; 10(2): 303–305.

29.

Maggi

Absinta

Grammatico

, et al. Central vein sign differentiates multiple sclerosis from central nervous system inflammatory vasculopathies. Ann Neurol 2018; 83(2): 283–294.

30.

Dworkin

Sati

Solomon

, et al. Automated integration of multimodal MRI for the probabilistic detection of the central vein sign in white matter lesions. AJNR Am J Neuroradiol 2018; 39(10): 1806–1813.

31.

Maggi

Fartaria

Jorge

, et al. CVSnet: A machine learning approach for automated central vein sign assessment in multiple sclerosis. NMR Biomed 2020; 33(5): e4283.

32.

Absinta

Nair

Filippi

, et al. Postmortem magnetic resonance imaging to guide the pathologic cut individualized, 3-dimensionally printed cutting boxes for fixed brains. J Neuropathol Exp Neurol 2014; 73(8): 780–788.

33.

Chavhan

Babyn

Thomas

, et al. Principles, techniques, and applications of T2*-based MR imaging and its special applications. Radiographics 2009; 29(5): 1433–1449.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.21 MB