View
3
Download
0
Category
Preview:
Citation preview
ARTICLE OPEN ACCESS
High serum neurofilament associates withdiffuse white matter damage in MSMaija Saraste PhD Svetlana Bezukladova MSc Markus Matilainen PhD Jouni Tuisku MSc
Eero Rissanen MD PhD Marcus Sucksdorff MD Sini Laaksonen MD Anna Vuorimaa MD
Jens Kuhle MD PhD David Leppert MD PhD and Laura Airas MD PhD
Neurol Neuroimmunol Neuroinflamm 20218e926 doi101212NXI0000000000000926
Correspondence
Dr Saraste
maijasarasteutufi
AbstractObjectiveTo evaluate to which extent serum neurofilament light chain (NfL) increase is related todiffusion tensor imagingndashMRI measurable diffuse normal-appearing white matter (NAWM)damage in MS
MethodsSeventy-nine patients with MS and 10 healthy controls underwent MRI including diffusiontensor sequences and serum NfL determination by single molecule array (Simoa) Fractionalanisotropy and mean axial and radial diffusivities were calculated within the whole and seg-mented (frontal parietal temporal occipital cingulate and deep) NAWM Spearman corre-lations and multiple regression models were used to assess the associations between diffusiontensor imaging volumetric MRI data and NfL
ResultsElevated NfL correlated with decreased fractional anisotropy and increased mean axial andradial diffusivities in the entire and segmented NAWM (for entire NAWM ρ = minus049 p = 0005ρ = 049 p = 0005 ρ = 043 p = 0018 and ρ = 048 p = 0006 respectively) A multipleregression model examining the effect of diffusion tensor indices on NfL showed significantassociations when adjusted for sex age disease type the expanded disability status scaletreatment and presence of relapses In the same model T2 lesion volume was similarlyassociated with NfL
ConclusionsOur findings suggest that elevated serum NfL in MS results from neuroaxonal damage bothwithin the NAWM and focal T2 lesions This pathologic heterogeneity ought to be taken intoaccount when interpreting NfL findings at the individual patient level
From the Turku PET Centre Turku University Hospital and University of Turku (M Saraste SB MM JT ER M Sucksdorff SL AV LA) Division of Clinical Neurosciences (ERM Sucksdorff SL AV LA) Turku University Hospital Finland and Departments of Medicine Biomedicine and Clinical Research Neurologic Clinic and Policlinic (JK DL)University Hospital Basel Switzerland
Go to NeurologyorgNN for full disclosures Funding information is provided at the end of the article
The Article Processing Charge was funded by the authors
This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 40 (CC BY-NC-ND) which permits downloadingand sharing the work provided it is properly cited The work cannot be changed in any way or used commercially without permission from the journal
Copyright copy 2020 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the American Academy of Neurology 1
MS is an inflammatory autoimmune disease of the CNS inwhich both acute and chronic inflammation lead to de-myelination and neuronal damage1 Neurofilament light chain(NfL) is one of the most promising soluble biomarkers forassessing disease activity in MS2 The cause and underlyingpathology of the increased NfL concentration indicative ofneuroaxonal injury may however vary greatly both inter- andintraindividually Significant CSF or blood NfL elevationshave been shown in association with acute focal inflammationand gadolinium-enhancing lesions in relapsing-remitting MS(RRMS)3ndash8 On the other hand moderate elevations in NfLconcentrations have been measured in patients with chronicprogressive MS568ndash10
We hypothesize that the NfL elevation in the context of moreadvanced MS disease where no signs of acute inflammation arepresent is caused by the diffuse pathologic process in the normal-appearing white matter (NAWM) which leads to a diffuseneuroaxonal damage not visible in conventional MRI This dif-fuse axonal injury in the NAWM can however be sensitivelymeasured using diffusion tensor imaging (DTI)11 Of the DTIscalars fractional anisotropy is highly sensitive for microstructuralchanges overall whereas axial and radial diffusivities are morespecific to axonal and myelin damage respectively12 It is notknown whether DTI scalars associate with NfL in MS
The aim of this study was to measure to which extent serumNfL increase is related to the diffuse damage in the NAWMincluding the type and spatial distribution of these changesTo address this we performed correlation analyses of NAWMDTI and serumNfL levels and showed significant associationsbetween these measures
MethodsStandard protocol approvals registrationsand patient consentsThe study was approved by the Ethical Committee of theHospital District of Southwest Finland Written informedconsent was obtained from all participants according to theDeclaration of Helsinki
Study cohortSeventy-nine patients with MS from the Neurology Out-patient Clinic of the Division of Clinical Neurosciences atthe Turku University Hospital Turku Finland and 10healthy age-matched controls were included in the studyMRI and serum sampling were performed ge30 days after aclinical relapse Clinical disease course disease duration
and patient age were reviewed and the Expanded Dis-ability Status Scale (EDSS)13 score was assessed by theinvestigating neurologist
Serum samplesBlood samples were collected and serum was stored at minus40degCwithin 4 hours of sampling Concentration of serum NfL wasmeasured by single molecule array (Simoa) assay technologyas described previously614
MRI and DTIBrain MRI was performed in Turku PET center with a 3 TMRI Phillips Ingenuity scanner (Philips HealthcareCleveland OH) Conventional MRI (3-dimensional T1-weighted MRI T2 and fluid-attenuated inversion recovery[FLAIR] with spatial resolution of 1 times 1 times 1 mm) and DTIsequences were included in the protocol The details of theimaging protocol have been described previously15 ForDTI sequences the following parameters were used bvalue = 1000 smmminus2 repetition timetime to echo =9500120 milliseconds field of view = 256 times 256 mmspatial resolution 2 times 2 times 2 mm acquisition matrix 128 times128 mm flip angle = 90deg and acceleration factor 2 with 33(n = 15) 64 (n = 48) or 67 gradient directions (n = 16)The number of gradient directions did not have a re-markable impact on fractional anisotropy mean diffusivityradial diffusivity and axial diffusivity15
NAWM region of interest was created by excluding alllesions and the cerebellar white matter from the whitematter region of interest (appendix e-1 linkslwwcomNXIA355) The lesions were identified from FLAIR im-ages using Lesion Segmentation Toolbox16 The NAWMregion of interest was further segmented to 6 subregions(frontal parietal temporal occipital cingulate and deepwhite matter which includes left and right unsegmentedwhite matter and insula) using FreeSurfer software17 Theareas included in subregions are shown in detail in ap-pendix e-2 The FreeSurfer software was also used to definethe volumes of cortical gray matter whole cerebral whitematter NAWM and total T1 and T2 lesion volumesaccording to our previously reported methodology18
The DTI data were preprocessed and analyzed withExploreDTI software19 T1 and raw diffusion-weightedimage files were first flip permuted after which the T1 filewas masked The files were then converted to diffusiontensor maps using robust diffusion tensor estimation afterwhich diffusion tensor maps were corrected for motioneddy current and echo planar imagingsusceptibility
GlossaryDTI = diffusion tensor imaging EDSS = Expanded Disability Status Scale FLAIR = fluid-attenuated inversion recovery IQR =interquartile range NAWM = normal-appearing white matter NfL = neurofilament light chain RRMS = relapsing-remittingMS SPMS = secondary progressive MS
2 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
induced distortions using the robust estimation of tensorsby outlier rejection tensor estimation method Fractionalanisotropy mean diffusivity radial diffusivity and axialdiffusivity maps were extracted from the corrected DTImaps Coregistering of maps to corresponding T1-weighted images was performed using SPM8 (The Well-come Centre for Human Neuroimaging University Col-lege London) running on MATLAB (The MathWorks
Natick MA) Finally the mean values for fractional an-isotropy mean diffusivity radial diffusivity and axial dif-fusivity within the whole NAWM and in the segmentedregions of NAWM were calculated in MATLAB
Statistical analysisThe statistical analysis was performed using R statistical soft-ware (version 400) The differences between different MS
Table 1 Demographics clinical characteristics volumetric brain MRI data and cingulate DTI-MRI indices of patients withMS and HCs
MS (n = 79) HC (n = 10)pValue RRMS (n = 56) SPMS (n = 23)
pValue
NfL(low)(n = 45)
NfL(high)(n = 34)
pValue
Femalemale 6019 73 07 4511 158 016 378 2311 018
Age 483(43ndash53)
476(44ndash53)
1 474 (42ndash51) 558 (48ndash61) lt0001 462 (42ndash51) 519 (45ndash58) 0003
Durationa 149 (9ndash19) NA mdash 121 (8ndash14) 196 (14ndash25) lt0001 126 (9ndash15) 145 (10ndash21) 009
EDSS score 3 (2ndash375) NA mdash 25 (2ndash3) 4 (35ndash65) lt0001 25 (2ndash3) 35 (3ndash5) 0005
Relapsesb 4 (2ndash6) 3 (2ndash6) 6 (3ndash8) 002 4 (2ndash6) 3 (2ndash625) 10
DMT (yesno) 5425 NA mdash 4511 914 lt0001 3312 2113 033
First linec 28 NA mdash 27 1 mdash 18 10
Second lined 26 NA mdash 18 8 mdash 15 11
NfL (pgmL) 217(15ndash31)
231(21ndash28)
03 188 (14ndash26) 314 (22ndash44) 0001 16 (12ndash19) 319 (29ndash43) lt0001
Volume (cm3)
WM 4521(412ndash485)
4928(463ndash525)
0009 4642 (423ndash501) 4227 (372ndash463) 0002 4521 (415ndash485) 4542 (406ndash481) 06
NAWM 4445(396ndash479)
NA mdash 4578 (407ndash498) 4107 (339ndash446) 0001 4445 (396ndash477) 4345 (391ndash479) 06
GMctx 4288(409ndash465)
473(450ndash512)
0006 4305 (414ndash470) 4163 (392ndash450) 009 4294 (409ndash464) 4246 (397ndash464) 04
T1 lesion 34 (2ndash8) NA mdash 27 (2ndash6) 65 (4ndash10) 0007 31 (2ndash9) 41 (2ndash7) 03
T2 lesion 69 (3ndash18) NA mdash 49 (2ndash12) 15 (13ndash25) lt0001 56 (2ndash15) 96 (4ndash19) 02
Cingulate DTI
FA 046(043ndash048)
049(047ndash050)
0007 046 (044ndash049) 044 (042ndash048) 01 046 (044ndash048) 047 (042ndash048) 09
MD (mm2s) 091(086ndash096)
083(082ndash087)
0005 090 (084ndash094) 092 (088ndash011) 005 091 (086ndash095) 091 (083ndash10) 07
AD (mm2s) 141(136ndash146)
135(131ndash139)
0028 139 (135ndash144) 144 (137ndash155) 003 141 (137ndash145) 140 (13ndash15) 07
RD (mm2s) 065(059ndash071)
058(056ndash062)
0004 064 (059ndash070) 069 (063ndash081) 005 066 (060ndash071) 064 (058ndash075) 07
Abbreviations AD = axial diffusivity DMT = disease-modifying treatment DTI = diffusion tensor imaging EDSS = Expanded Disability Status Scale FA =fractional anisotropy GMctx = cortical graymatter HC = healthy control MD=meandiffusivity NA = not applicable NAWM=normal-appearingwhitematterNfL = neurofilament light chain RD = radial diffusivity RRMS = relapsing-remitting MS SPMS = secondary progressive MS WM = white matterPatients withMSwere divided into NfL(low) andNfL(high) subgroups based on themedian value of HCs (231 pgmL) Median (interquartile range) values areshown except for sex and DMT where the number of patients is shown The p values are from the Wilcoxon rank-sum test for continuous variables and theFisher exact test for class variables Significant p values are bolded Values of mean axial and radial diffusivities are expressed as e10minus3a Disease duration (years) calculated from the onset of symptomsb Total number of relapses before study onset data of 2 patients are missingc First-line DMT = interferon-beta dimethyl fumarate glatiramer acetate and teriflunomided Second-line DMT = fingolimod natalizumab and rituximab
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 3
groups and between patients and healthy controls wereassessed using theWilcoxon rank-sum (Mann-WhitneyU) testHolm multiple comparison adjustment was used in RRMS vssecondary progressive MS (SPMS) vs healthy control multiplecomparisons The division into NfL(low) and NfL(high)subgroups was based on the median NfL value of the healthycontrols
To test our main hypothesis Spearman correlations were calcu-lated to assess the relationships between NfL and DTI indices(fractional anisotropy mean diffusivity radial diffusivity and axialdiffusivity) in the whole NAWM and in NAWM subregionsSpearman correlation was used instead of Pearson correlation toavoid outlier-induced overestimation of correlations The p valuesof correlation analyses were further adjusted with the false dis-covery rate method (Benjamini-Hochberg procedure) for thenumber of investigated parameters (n = 28)20 The NfL valueswere further modeled by the DTI indices and volumetric MRIdata using multiple regression models The logarithm of NfL wasused as the response because nontransformed values led to non-normality of residuals The models were adjusted by sex agedisease type (relapsing-remittingsecondary progressive) theEDSS score disease-modifying treatment (no treatment first line[dimethyl fumarate glatiramer acetate interferon-beta and teri-flunomide] or second line [fingolimod natalizumab and ritux-imab] treatment) and the presence of relapses within 1 yearbefore sampling (yesno) The normality of the residuals waschecked using the Shapiro-Wilk test Variance inflation factorvalues were used to check that independent variables were nothighly correlated with each other Regression coefficients were
standardized tomake themdirectly comparable to each other Forstandardization the regression results weremultiplied by twice theSD of the DTI or volume variables21 The p values of the multipleregression model were adjusted with the false discovery ratemethod for the number of investigated variables (n = 32)
In addition to verify the quality of the data and increase thegeneralizability and comparability of the results we usedSpearman correlation to assess the relationships betweenfollowing parameters NfL with age disease duration theEDSS score parenchymal fractions of NAWM cortical graymatter and T1 and T2 lesion volumes DTI with the EDSSscore and parenchymal fractions of NAWM cortical graymatter and T1 and T2 lesion volumes (results are shown inappendix e-3 linkslwwcomNXIA355) Spearman cor-relation was used instead of Pearson correlation to avoidoutlier-induced overestimation of correlations and becauseSpearman correlation is preferred in case of ordinal vari-ables ie EDSS score All statistical tests were 2 tailed andp = 005 was used as the threshold for statistical significance
Data availabilityThe anonymized raw data will be shared over the next 3 yearson request from a qualified investigator
ResultsThe demographics clinical characteristics and conventionalimaging data of the 79 patients with MS included in the study
Figure 1 Correlation of serum NfL and DTI indices of the whole NAWM in the NfL(high) subgroup
The serum level of NfL is moderately associatedwith fractional anisotropy (A) mean diffusivity(B) axial diffusivity (C) and radial diffusivity (D)The NfL(high) subgroup is comprised of patientswith serum NfL above the median value ofhealthy controls (231 pgmL) Shown areSpearman correlation coefficients (ρ) and p val-ues The black lines indicate the level and di-rection of the relationship AD = axial diffusivityDTI = diffusion tensor imaging FA = fractionalanisotropy MD = mean diffusivity NAWM =normal-appearing white matter NfL = neuro-filament light chain RD = radial diffusivity
4 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
Table 2 Spearman correlations between serum NfL concentrations and DTI-MRI indices
All MS NfL(low) NfL(high)
ρ p Value ρ p Value ρ p Value
Entire NAWM
FA minus016 02 minus020 02 minus042 0013
MD 012 03 014 04 047 0005
AD 001 09 005 07 041 0017
RD 013 03 015 03 045 0008
Frontal NAWM
FA minus018 01 minus020 02 minus037 0033
MD 017 01 011 05 047 0005
AD 009 04 004 08 044 0010
RD 015 02 013 04 044 0009
Parietal NAWM
FA minus009 04 minus011 05 minus034 0046
MD 005 07 005 08 043 0011
AD minus004 08 minus003 08 034 0053
RD 007 05 008 06 044 0009
Temporal NAWM
FA minus019 01 minus012 04 minus033 0060
MD 014 02 008 06 043 0012
AD minus003 08 004 08 026 0139
RD 014 02 009 07 038 0026
Occipital NAWM
FA minus009 04 minus018 02 minus043 0011
MD 012 03 017 03 053 0001
AD 004 07 014 04 038 0029
RD 010 04 015 03 048 0004
Cingulate NAWM
FA minus010 04 minus006 07 minus048 0004
MD 008 05 002 09 049 0003
AD 009 04 010 05 046 0007
RD 008 05 004 08 050 0003
Deep NAWM
FA minus014 02 minus020 03 minus036 0036
MD 008 05 010 06 033 0053
AD 003 08 001 09 036 0039
RD 01 04 012 05 036 0036
Abbreviations AD = axial diffusivity DTI = diffusion tensor imaging FA = fractional anisotropy MD = mean diffusivity NAWM = normal-appearing whitematter NfL = neurofilament light chain RD = radial diffusivityPatients with MS were divided into NfL(low) and NfL(high) subgroups based on the median value of healthy controls (231 pgmL) Fractional anisotropy andmean axial and radial diffusivities of the entire NAWMand of 6 parcellated subregions of NAWMwere correlated with serumNfL level in all patients with MSand in NfL(low) and NfL(high) subgroups Shown are Spearman correlation coefficients (ρ) and uncorrected p values Significant p values are boldedSignificance of these correlations was sustained after adjustment using the false discovery ratemethod for the number of DTI parameters (n = 28) except forparietal FA
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 5
are shown in table 1 The median (interquartile range [IQR])age of this cohort was 483 (43ndash53) years which is similar tothe age of the healthy individuals in the control group (476[44ndash53] p = 1) Overall the patient cohort was representedwith a quite stable disease as only 13 of the patients (9 RRMSand 1 SPMS) had had a relapse within 1 year before samplingSixty-eight percent of the patients (45 RRMS and 9 SPMS)were on disease-modifying therapy (dimethyl fumarate n = 3fingolimod n = 14 glatiramer acetate n = 5 interferon beta-1a n= 6 natalizumab n = 8 rituximab n = 4 and teriflunomide n =14) In MRI white matter and cortical gray matter volumeswere decreased in patients compared with healthy controlsDTI measures were obtained from both the entire cerebralNAWM and from 6 different NAWM subregions In the cin-gulate area there were significant differences in all 4 DTI pa-rameters in patients with MS compared with healthy controlsfractional anisotropy was decreased whereas mean diffusivityradial diffusivity and axial diffusivity were increased (table 1)
Serum NfL levels in patients with MS andhealthy controlsThe median (IQR) NfL level was higher in patients withSPMS compared with patients with RRMS (314 [22ndash44] vs188 [14ndash26] pgmL p = 0001 table 1) NfL levels in thewhole MS group were not different from healthy controls(217 [15ndash31] vs 231 [21ndash28] pgmL p = 03) regardlesswhether the analysis was performed with age correctionAcute inflammation did not appear to be a factor to affect theNfL level in this cohort with relatively modest acute in-flammatory activity as the median NfL value among thepatients with or without relapse within the previous year wascomparable (223 [16ndash32] vs 186 [15ndash25] pgmL p = 03Wilcoxon rank-sum data not shown)
Characterization of patients with highNfL levelsTo explore associations between increased NfL levels and DTI-measurable diffuse neuroaxonal damage patients with MS weredivided into NfL(high) and NfL(low) subgroups The divisionwas based on the median NfL value measured among healthycontrols (231 [21ndash28] pgmL) In the NfL(high) subgroup theNfL concentration was significantly elevated compared withhealthy controls (p = 0018 table 1) and in the further NfL vsDTI correlation analyses we focused on the NfL(high) subgroupBased on the demographic and clinical data the patients in theNfL(high) subgroupwere at amore advanced stage of the diseaseTheNfL(high) subgroup includedmore patients with SPMS (p =0003) and the patients in the subgroup were also older (p =0003) and had a higher EDSS score (p = 0005 table 1) Inevaluation using conventional MRI the NAWM white matter orcortical gray matter volumes and T1 and T2 lesion loads were notdifferent between NfL(low) and NfL(high) subgroups (table 1)
Associations between NfL and DTIOur results show that DTI metrics of diffuse neuroaxonaldamage within theNAWMassociatewith high serumNfL levelsin MS As a demonstration of this in the NfL(high) subgroup
several (n = 24) significant correlations were found betweenserum NfL levels and NAWM DTI indices both in the entireNAWM and in various brain subregions higher NfL levels wereassociated with lower fractional anisotropy and higher diffusivity(mean axial and radial) in the whole NAWM (figure 1) and inall its subregions except for fractional anisotropy in temporalNAWM axial diffusivity in parietal and temporal NAWM andmean diffusivity in remaining NAWM (table 2) Significance ofthese correlations was sustained after adjustment using the falsediscovery rate method for the number of DTI parameters (n =28) except for parietal fractional anisotropy Results remainedsimilar when the data were analyzed without patients who hadhad a relapse within the previous year before sampling (n = 3data not shown) No correlations between serumNfL levels andDTI metrics were observed in the NfL(low) subgroup or in theoverall MS cohort (table 2)
Multiple regression modeling was performed to further evaluatethe effect of DTI on NfL In the NfL(high) subgroup the DTIindices weremore significantly associated withNfL thanwith theclinical parameters included in the model (sex age disease typeEDSS score treatment and the presence of relapses within 1year before sampling) (table 3) On average 48 of the variancein NfL (SD 44) could be explained by DTI indices and theclinical parameters mentioned above (figure 2) There weremultiple (n = 26) significant associations between NfL and DTIin the whole NAWM and in all its 6 subregions All except 2 ofthe associations remained significant after adjustment with thefalse discovery rate method for the number of investigated var-iables (n = 32) The strongest associations were observed in thecingulate area between higher NfL and higher mean and radialdiffusivity and in deep and frontal NAWM between higher NfLand higher mean diffusivity (figure 2) Analyzing the datawithout the patients with a relapse (n = 3) within the previousyear did not alter the associations (data not shown)
Higher NfL was also associated with lower NAWM andcortical gray matter and higher T2 lesion volumes (figure 2)The volumes and clinical adjustments explained 434494 and 433 of the variance in NfL respectively
DiscussionPresent results demonstrate that the DTI-MRI measures ofNAWM correlate with serumNfL in MS We found that amongpatients with more advanced disease increased serumNfL levelsassociate with DTI measures reflecting diffuse microstructuraldamage ie decreased fractional anisotropy and increasedmeanaxial and radial diffusivities in the NAWM In a multiple re-gression model which was adjusted with sex age disease typeEDSS score treatment and the presence of relapses the DTIindices were more significantly associated with NfL than theabove-mentioned demographic and clinical parameters
The pathophysiology of MS involves acute and chronic mech-anisms that lead to gradual axonal and myelin damage that
6 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
Table 3 Multiple regression analyses evaluating separately the effect of DTI-MRI indices or volumetric brain MRI data onNfL in the NfL(high) subgroup of patients with MS
Variable in model DTI Sex Age Type EDSS score DMT Relapses
Entire NAWM
FA 0018 07 08 10 03 04 02
MD 0003 07 09 10 02 04 04
AD 0010 09 10 08 01 07 05
RD 0005 07 10 10 02 04 03
Frontal NAWM
FA 0045 08 10 10 02 07 02
MD 0001 05 07 10 02 06 04
AD 0004 05 09 08 03 07 04
RD 0003 06 09 09 02 06 03
Parietal NAWM
FA 0023 09 09 09 02 05 01
MD 0004 07 07 08 01 05 02
AD 0017 08 10 10 01 06 03
RD 0006 08 08 09 02 05 02
Temporal NAWM
FA 0116 07 07 10 03 05 03
MD 0007 06 10 10 01 05 05
AD 0009 07 08 08 01 05 05
RD 0020 06 09 10 02 05 04
Occipital NAWM
FA 0011 09 06 09 03 03 01
MD 0009 08 08 09 01 06 02
AD 0086 08 07 08 01 08 03
RD 0011 08 08 08 02 05 02
Cingulate NAWM
FA 0004 05 04 08 06 04 03
MD 0001 06 09 08 06 02 07
AD 0003 08 09 08 03 04 07
RD 0001 06 07 08 07 02 06
Deep NAWM
FA 0047 05 06 09 03 05 03
MD 0001 07 09 09 05 02 07
AD 0005 10 08 10 03 04 08
RD 0002 06 08 10 05 02 08
Volume (pf)
NAWM 0019 08 08 08 03 05 05
Continued
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 7
eventually manifest by brain atrophy and clinically by steadyworsening of disability in almost all patients over time This diffusepathology which in the majority of cases leads to progressiveworsening of the MS-related symptoms is now the focus of in-terest for patients and physicians as modern disease-modifyingtherapies have led to an almost complete suppression of relapseactivity while their impact on progression is at best modest Ourincomplete understanding of the molecular mechanisms that leadto progression is one of the main reasons for the failure in de-velopment ofmore efficacious therapies for secondary progressiveform of MS Another impediment is the relative insensitivity ofcurrent measurement tools to capture features of progression atsubclinical stages and hence to quantitate the effects of anti-neurodegenerative treatments at a time point where the overallbrain structure is still largely intact Specifically conventional MRIcan only quantitate brain volume loss as an end result of thedisease but does not provide insights into the underlying micro-structural changes DTI is a nonconventional MRI technique thatprovides a measure of such diffuse changes beyond volumetry
NfL is a highly sensitive biomarker to detect neuronal damageand is the first of its kind applicable in blood-derived probes22
However the cause and underlying pathology of the increasedNfL concentration may vary greatly both inter- and intra-individually In clinically isolated syndrome and RRMS acuteNfL elevation is particularly associated with focal in-flammation ie relapses and focal lesion formation4ndash823
However the indication for the contribution of diffuse pa-thology to higher NfL levels has so far been only incidentaland consisted of associations of increased NfL with moreadvanced clinical disability and brain atrophy824 Novelfindings presented here provide more direct evidence for theeffect of diffuse pathologic process in the NAWM on in-creased serum NfL We observed that among a subgroup ofolder and more disabled patients NfL associates both withradial and axial diffusivities In the context of MS these DTIscalars are thought to reflect myelin loss and axonal damagerespectively11 NfL also associated with mean diffusivity thatcan also indicate on axonal andmyelin loss11 The associationswere most pronounced in the cingulate deep and frontalNAWM
NfL differs in 2 important aspects from MRI first as asignal of neurodegeneration it reflects ongoing pathologicprocesses in real time while imaging captures morphologicfeatures that are inherently retrospective Second NfLlevels reflect neuronal damage in the CNS comprehen-sively ie in addition to brain pathology it also capturesspinal cord pathology that is not routinely evaluated byimaging in the workup of MS for individual patients As adownside NfL levels cannot differentiate betweenacute inflammatory and chronic neurodegenerative diseaseactivity Our results suggest that DTI-MRI allows us tocategorize the pathogenic source of NfL at least semi-quantitatively We demonstrate that both diffuse axonaldamage within the NAWM measured using DTI and focalinflammatory white matter lesion load measured usingconventional MRI contributed to the elevated NfL serumlevels This highlights that both focal and diffuse pathologicchanges must be taken into account when levels of NfL inindividual patients are interpreted A somewhat largerproportion of NfL variance was explained by DTI indices inaverage than by T2 lesion load This implies that in non-relapsing patients the ongoing diffuse microstructuraldamage in the NAWM outside lesions might predominateover focal axonal damage that has occurred within lesionsas a source of NfL in serum Accordingly in other neuro-logic disease with brain diffuse pathology such as amyo-trophic lateral sclerosis Alzheimer disease traumatic braininjury and age-related white matter pathology elevatedserum NfL levels have been associated with DTI-measurable diffuse damage within the CNS25ndash28 Thesestudies are further evidence for the concept that braindiffuse chronic neuropathologic mechanisms contribute toNfL release and that nonconventional MRI techniques areable to identify the CNS areas of its morphologic source
Current results demonstrate that the cingulate area is thebrain area with the most prominent white matter tractdamage outside focal lesions Periventricular regions arecharacteristically the brain area with a particularly heavylesion load in MS29 We addressed the question whetherthe cingulum-NAWM DTI abnormalities would arise from
Table 3 Multiple regression analyses evaluating separately the effect of DTI-MRI indices or volumetric brain MRI data onNfL in the NfL(high) subgroup of patients with MS (continued)
Variable in model DTI Sex Age Type EDSS score DMT Relapses
GMctx 0004 09 02 04 003 03 06
T2 0019 06 03 07 04 07 03
T1 015 07 04 10 05 07 03
Abbreviations AD = axial diffusivity DMT = disease-modifying treatment DTI = diffusion tensor imaging EDSS = Expanded Disability Status Scale FA =fractional anisotropy GMctx = cortical gray matter MD = mean diffusivity NAWM = normal-appearing white matter NfL = neurofilament light chain pf =parenchymal fraction RD = radial diffusivityPatients withMSwere divided into NfL(low) and NfL(high) subgroups based on themedian NfL value of healthy controls (231 pgmL) Model was adjusted bysex age disease type (relapsing-remittingsecondary progressive) EDSS score treatment (no first- and second-line) and presence of relapses within 1 yearbefore sampling (yesno) Bolded p values indicate that the parameter affected the logarithmofNfL value statistically significantly whenDTI or volume and allthe adjusting variables were in the models
8 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
the dirty-appearing white matter due to close proximity to aheavy lesion load but the relative lesion volume in the cingulatearea was not found to be significantly larger compared withlesion load in other regions or in the whole NAWM Thereforeit seems unlikely that differences in the DTI indices would besolely due to dirty-appearing white matter adjacent to lesionswhereas it is more likely that remote lesions lead to DTI changesin the cingulateNAWM throughWallerian degeneration30 Alsoother studies have observed diffusion changes in the cingulatearea of patients with MS31ndash33 Because the cingulum bundle offibers projects from the cingulate gyrus to the entorhinal cortexin the brain allowing for communication between components
of the limbic system the DTI abnormalities may be the mor-phologic substrate for impaired cognitive function and fatiguesymptoms that are typical in progressive MS3435
Present results are in line with previous observations on therelation between NfL concentration and MS disease sub-type with progressive patients having greater NfL levelscompared with relapsing-remitting patients568ndash102324
However the median NfL value of the entire patient cohortwas relatively low compared with healthy controls which isthe main limitation of our study Because of this we werenot able to use the previously defined cutoff values3 for
Figure 2 The effect of DTI-MRI indices or volumetric MRI data on NfL in the NfL(high) subgroup
The NfL(high) subgroup is comprised of patients with serum NfL above the median value of healthy controls (231 pgmL) Logarithmic serum NfL wasmodeled separately by DTI indices of the entire and parcellated NAWM and volumetric brain MRI data using multiple regression analyses Models wereadjusted by sex age disease type (RRMSSPMS) the EDSS score treatment (no first and second line) and presence of relapseswithin 1 year before sampling(yesno) The results are illustrated using dot and whisker plots in which red dots represent standardized regression coefficients and red lines represent theCIs of the estimates Significant p values of the DTI parameters and the percentage of variance in the response that can be explained by the independentvariables (R2) are also shown The results were ordered according to the R2 value All except 2 of the p values (fractional anisotropy of the frontal and deepNAWM) shown in the figure remained significant after adjustment using the false discovery ratemethod for the number of investigated variables (n = 32) AD =axial diffusivity DTI = diffusion tensor imaging EDSS = Expanded Disability Status Scale FA = fractional anisotropy GMctx = cortical gray matter MD =meandiffusivity NAWM = normal-appearing white matter NfL = neurofilament light chain PF = parenchymal fraction RD = radial diffusivity RRMS = relapsing-remitting MS SPMS = secondary progressive MS WM = white matter
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 9
NfL(low) and NfL(high) subgroups Instead we used themedian value of healthy controls Although the divider usedin our study was based on a small group of individualssimilar values have been observed in previous largerstudies68 One reason for the low median NfL level couldbe that the patients with RRMS are represented with arelatively benign disease course In addition our patientswith RRMS were sampled during a relapse-free time andhad no signs of ongoing focal inflammatory activity in theMRI Moreover most of the patients (68) were using adisease-modifying treatment at the time of evaluationHence similar as observed in other studies35233637 thisresulted in low serum NfL concentration Following thesimilarity in the NfL levels between healthy controls andpatients with MS the main results of our study are based ona relatively low number of patients In addition the patientnumbers under different immunomodulatory treatmentsvaried substantially However the heterogeneity of thecohort was taken account in the multiple regression modelwhere both EDSS score and disease-modifying treatmentwere used as adjustments
Our findings suggest that elevated serum NfL in MS resultsfrom neuroaxonal damage both within the NAWM and infocal T2 lesions The association between DTI-measurablediffuse microstructural white matter damage and serumNfL isfurther conceptual evidence for the latter being a usefulmonitoring tool in assessing the degree of ongoing neuro-degenerative processes in MS
AcknowledgmentThe authors thank all patients with MS and healthycontrols who participated in this study They also thank thepersonnel of Turku PET Centre for their excellenttechnical assistance
Study fundingThis work was supported by the Finnish Academy the SigridJuselius Foundation
DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgNN for full disclosures
Publication historyReceived by Neurology Neuroimmunology amp NeuroinflammationMay 8 2020 Accepted in final form October 21 2020
References1 Reich DS Lucchinetti CF Calabresi PA Multiple sclerosis N Engl J Med 2018378
169ndash1802 Varhaug KN Torkildsen Oslash Myhr KM Vedeler CA Neurofilament light chain as a
biomarker in multiple sclerosis Front Neurol 201910338
Appendix Authors
Author Location Contribution
MaijaSaraste PhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Acquisition of dataanalyzed andinterpreted the data anddrafted and revised themanuscript forintellectual content
Appendix (continued)
Author Location Contribution
SvetlanaBezukladovaMSc
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Acquisition of dataanalyzed and interpretedthe data and revised themanuscript
MarkusMatilainenPhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Statistical analysisdrafted the figures andrevised the manuscript
Jouni TuiskuMSc
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Major role in theacquisition of data andrevised the manuscript
EeroRissanenMDPhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Designed andconceptualized the studyand acquisition of data
MarcusSucksdorffMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Major role in theacquisition of data
SiniLaaksonenMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Acquisition of data
AnnaVuorimaaMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Acquisition of data
Jens KuhleMD PhD
Neurologic Clinic andPoliclinic Departments ofMedicine Biomedicineand Clinical ResearchUniversity Hospital BaselSwitzerland
Major role in theacquisition of data
DavidLeppert MDPhD
Neurologic Clinic andPoliclinic Departments ofMedicine Biomedicineand Clinical ResearchUniversity Hospital BaselSwitzerland
Revised the manuscriptfor intellectual content
Laura AirasMD PhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Designed andconceptualized the studyand revised themanuscript for intellectualcontent
10 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
3 Kuhle J Kropshofer H Haering DA et al Blood neurofilament light chain as abiomarker of MS disease activity and treatment response Neurology 201992e1007ndashe1015
4 Varhaug KN Barro C Bjoslashrnevik K et al Neurofilament light chain predicts diseaseactivity in relapsing-remitting MS Neurol Neuroimmunol Neuroinflamm 20185e422
5 Novakova L Zetterberg H Sundstrom P et al Monitoring disease activity in multiplesclerosis using serum neurofilament light protein Neurology 2017892230ndash2237
6 Disanto G Barro C Benkert P et al Serum neurofilament light a biomarker ofneuronal damage in multiple sclerosis Ann Neurol 201781857ndash870
7 Siller N Kuhle J Muthuraman M et al Serum neurofilament light chain is a bio-marker of acute and chronic neuronal damage in early multiple sclerosis Mult Scler201925678ndash686
8 Barro C Benkert P Disanto G et al Serum neurofilament as a predictor of diseaseworsening and brain and spinal cord atrophy in multiple sclerosis Brain 20181412382ndash2391
9 Ferraro D Guicciardi C De Biasi S et al Plasma neurofilaments correlate withdisability in progressive multiple sclerosis patients Acta Neurol Scand 202014116ndash21
10 Jakimovski D Zivadinov R RamanthanM et al Serum neurofilament light chain levelassociations with clinical and cognitive performance in multiple sclerosis a longitu-dinal retrospective 5-year study Mult Scler 2020261670ndash1681
11 Sbardella E Tona F Petsas N Pantano P DTI measurements in multiple sclerosisevaluation of brain damage and clinical implications Mult Scler Int 20132013671730
12 Tae WS Ham BJ Pyun SB Kang SH Kim BJ Current clinical applications ofdiffusion-tensor imaging in neurological disorders J Clin Neurol 201814129ndash140
13 Kurtzke JF Rating neurologic impairment in multiple sclerosis an expanded disabilitystatus scale (EDSS) Neurology 1983331444ndash1452
14 Kuhle J Barro C Andreasson U et al Comparison of three analytical platforms forquantification of the neurofilament light chain in blood samples ELISA electro-chemiluminescence immunoassay and SimoaClinChemLabMed 2016541655ndash1661
15 Bezukladova S Tuisku J Matilainen M et al Insights into disseminated MS brainpathology with multimodal diffusion tensor and PET imaging Neurol NeuroimmunolNeuroinflamm 20207e691
16 Schmidt P Gaser C Arsic M et al An automated tool for detection of FLAIR-hyperintense white-matter lesions in Multiple Sclerosis Neuroimage 2012593774ndash3783
17 Salat DH Greve DN Pacheco JL et al Regional white matter volume differences innondemented aging and Alzheimerrsquos disease Neuroimage 2009441247ndash1258
18 Rissanen E Tuisku J Vahlberg T et al Microglial activation white matter tractdamage and disability in MS Neurol Neuroimmunol Neuroinflamm 20185e443
19 Leemans A Jeurissen B Sijbers J et al ExploreDTI a graphical toolbox for pro-cessing analysing and visualizing diffusion MR data Presented at the 17th AnnualMeeting of International Society for Magnetic Resonance in Medicine Honolulu HI2009 p 3537
20 Benjamini Y Hochberg Y Controlling the false discovery rate a practical and pow-erful approach to multiple testing J R Stat Soc Ser B 199557289ndash300
21 Gelman A Scaling regression inputs by dividing by two standard deviations Stat Med2008272865ndash2873
22 Khalil M Teunissen CE Otto M et al Neurofilaments as biomarkers in neurologicaldisorders Nat Rev Neurol 201814577ndash589
23 Canto E Barro C Zhao C et al Association between serum neurofilament light chainlevels and long-term disease course among patients with multiple sclerosis followedup for 12 years JAMA Neurol 2019761359ndash1366
24 Hogel H Rissanen E Barro C et al Serum glial fibrillary acidic protein correlates withmultiple sclerosis disease severity Mult Scler 202026210ndash219
25 Menke RA Gray E Lu CH et al CSF neurofilament light chain reflects corticospinaltract degeneration in ALS Ann Clin Transl Neurol 20152748ndash755
26 Moore EE Hohman TJ Badami FS et al Neurofilament relates to white mattermicrostructure in older adults Neurobiol Aging 201870233ndash241
27 Ljungqvist J Zetterberg H Mitsis M Blennow K Skoglund T Serum neurofilamentlight protein as a marker for diffuse axonal injury results from a case series studyJ Neurotrauma 2017341124ndash1127
28 Schultz SA Strain JF Adedokun A et al Serum neurofilament light chain levels areassociated with white matter integrity in autosomal dominant Alzheimerrsquos diseaseNeurobiol Dis 2020142104960
29 Narayanan S Fu L Pioro E et al Imaging of axonal damage in multiple sclerosisspatial distribution of magnetic resonance imaging lesions Ann Neurol 199741385ndash391
30 Ciccarelli O Werring DJ Barker GJ et al A study of the mechanisms of normal-appearing white matter damage in multiple sclerosis using diffusion tensor imagingevidence of Wallerian degeneration J Neurol 2003250287ndash292
31 Pokryszko-Dragan A Banaszek A Nowakowska-Kotas M et al Diffusion tensorimaging findings in the multiple sclerosis patients and their relationships to variousaspects of disability J Neurol Sci 2018391127ndash133
32 Kolasa M Hakulinen U Brander A et al Diffusion tensor imaging and disabilityprogression in multiple sclerosis a 4-year follow-up study Brain Behav 20199e01194
33 Andersen O Hildeman A Longfils M et al Diffusion tensor imaging in multiplesclerosis at different final outcomes Acta Neurol Scand 2018137165ndash173
34 Koenig KA Sakaie KE Lowe MJ et al The relationship between cognitive functionand high-resolution diffusion tensor MRI of the cingulum bundle in multiple sclerosisMult Scler 2015211794ndash1801
35 Pardini M Bonzano L Bergamino M et al Cingulum bundle alterations underliesubjective fatigue in multiple sclerosis Mult Scler 201521442ndash447
36 Sejbaek T Nielsen HH Penner N et al Dimethyl fumarate decreases neurofilamentlight chain in CSF and blood of treatment naıve relapsing MS patients J NeurolNeurosurg Psychiatry 2019901324ndash1330
37 Piehl F Kockum I Khademi M et al Plasma neurofilament light chain levels inpatients with MS switching from injectable therapies to fingolimod Mult Scler 2018241046ndash1054
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 11
DOI 101212NXI000000000000092620218 Neurol Neuroimmunol Neuroinflamm
Maija Saraste Svetlana Bezukladova Markus Matilainen et al High serum neurofilament associates with diffuse white matter damage in MS
This information is current as of December 8 2020
ServicesUpdated Information amp
httpnnneurologyorgcontent81e926fullhtmlincluding high resolution figures can be found at
References httpnnneurologyorgcontent81e926fullhtmlref-list-1
This article cites 36 articles 3 of which you can access for free at
Citations httpnnneurologyorgcontent81e926fullhtmlotherarticles
This article has been cited by 1 HighWire-hosted articles
Subspecialty Collections
httpnnneurologyorgcgicollectionmultiple_sclerosisMultiple sclerosis
httpnnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpnnneurologyorgmiscaboutxhtmlpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnnneurologyorgmiscaddirxhtmlreprintsusInformation about ordering reprints can be found online
Academy of Neurology All rights reserved Online ISSN 2332-7812Copyright copy 2020 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the AmericanPublished since April 2014 it is an open-access online-only continuous publication journal Copyright
is an official journal of the American Academy of NeurologyNeurol Neuroimmunol Neuroinflamm
MS is an inflammatory autoimmune disease of the CNS inwhich both acute and chronic inflammation lead to de-myelination and neuronal damage1 Neurofilament light chain(NfL) is one of the most promising soluble biomarkers forassessing disease activity in MS2 The cause and underlyingpathology of the increased NfL concentration indicative ofneuroaxonal injury may however vary greatly both inter- andintraindividually Significant CSF or blood NfL elevationshave been shown in association with acute focal inflammationand gadolinium-enhancing lesions in relapsing-remitting MS(RRMS)3ndash8 On the other hand moderate elevations in NfLconcentrations have been measured in patients with chronicprogressive MS568ndash10
We hypothesize that the NfL elevation in the context of moreadvanced MS disease where no signs of acute inflammation arepresent is caused by the diffuse pathologic process in the normal-appearing white matter (NAWM) which leads to a diffuseneuroaxonal damage not visible in conventional MRI This dif-fuse axonal injury in the NAWM can however be sensitivelymeasured using diffusion tensor imaging (DTI)11 Of the DTIscalars fractional anisotropy is highly sensitive for microstructuralchanges overall whereas axial and radial diffusivities are morespecific to axonal and myelin damage respectively12 It is notknown whether DTI scalars associate with NfL in MS
The aim of this study was to measure to which extent serumNfL increase is related to the diffuse damage in the NAWMincluding the type and spatial distribution of these changesTo address this we performed correlation analyses of NAWMDTI and serumNfL levels and showed significant associationsbetween these measures
MethodsStandard protocol approvals registrationsand patient consentsThe study was approved by the Ethical Committee of theHospital District of Southwest Finland Written informedconsent was obtained from all participants according to theDeclaration of Helsinki
Study cohortSeventy-nine patients with MS from the Neurology Out-patient Clinic of the Division of Clinical Neurosciences atthe Turku University Hospital Turku Finland and 10healthy age-matched controls were included in the studyMRI and serum sampling were performed ge30 days after aclinical relapse Clinical disease course disease duration
and patient age were reviewed and the Expanded Dis-ability Status Scale (EDSS)13 score was assessed by theinvestigating neurologist
Serum samplesBlood samples were collected and serum was stored at minus40degCwithin 4 hours of sampling Concentration of serum NfL wasmeasured by single molecule array (Simoa) assay technologyas described previously614
MRI and DTIBrain MRI was performed in Turku PET center with a 3 TMRI Phillips Ingenuity scanner (Philips HealthcareCleveland OH) Conventional MRI (3-dimensional T1-weighted MRI T2 and fluid-attenuated inversion recovery[FLAIR] with spatial resolution of 1 times 1 times 1 mm) and DTIsequences were included in the protocol The details of theimaging protocol have been described previously15 ForDTI sequences the following parameters were used bvalue = 1000 smmminus2 repetition timetime to echo =9500120 milliseconds field of view = 256 times 256 mmspatial resolution 2 times 2 times 2 mm acquisition matrix 128 times128 mm flip angle = 90deg and acceleration factor 2 with 33(n = 15) 64 (n = 48) or 67 gradient directions (n = 16)The number of gradient directions did not have a re-markable impact on fractional anisotropy mean diffusivityradial diffusivity and axial diffusivity15
NAWM region of interest was created by excluding alllesions and the cerebellar white matter from the whitematter region of interest (appendix e-1 linkslwwcomNXIA355) The lesions were identified from FLAIR im-ages using Lesion Segmentation Toolbox16 The NAWMregion of interest was further segmented to 6 subregions(frontal parietal temporal occipital cingulate and deepwhite matter which includes left and right unsegmentedwhite matter and insula) using FreeSurfer software17 Theareas included in subregions are shown in detail in ap-pendix e-2 The FreeSurfer software was also used to definethe volumes of cortical gray matter whole cerebral whitematter NAWM and total T1 and T2 lesion volumesaccording to our previously reported methodology18
The DTI data were preprocessed and analyzed withExploreDTI software19 T1 and raw diffusion-weightedimage files were first flip permuted after which the T1 filewas masked The files were then converted to diffusiontensor maps using robust diffusion tensor estimation afterwhich diffusion tensor maps were corrected for motioneddy current and echo planar imagingsusceptibility
GlossaryDTI = diffusion tensor imaging EDSS = Expanded Disability Status Scale FLAIR = fluid-attenuated inversion recovery IQR =interquartile range NAWM = normal-appearing white matter NfL = neurofilament light chain RRMS = relapsing-remittingMS SPMS = secondary progressive MS
2 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
induced distortions using the robust estimation of tensorsby outlier rejection tensor estimation method Fractionalanisotropy mean diffusivity radial diffusivity and axialdiffusivity maps were extracted from the corrected DTImaps Coregistering of maps to corresponding T1-weighted images was performed using SPM8 (The Well-come Centre for Human Neuroimaging University Col-lege London) running on MATLAB (The MathWorks
Natick MA) Finally the mean values for fractional an-isotropy mean diffusivity radial diffusivity and axial dif-fusivity within the whole NAWM and in the segmentedregions of NAWM were calculated in MATLAB
Statistical analysisThe statistical analysis was performed using R statistical soft-ware (version 400) The differences between different MS
Table 1 Demographics clinical characteristics volumetric brain MRI data and cingulate DTI-MRI indices of patients withMS and HCs
MS (n = 79) HC (n = 10)pValue RRMS (n = 56) SPMS (n = 23)
pValue
NfL(low)(n = 45)
NfL(high)(n = 34)
pValue
Femalemale 6019 73 07 4511 158 016 378 2311 018
Age 483(43ndash53)
476(44ndash53)
1 474 (42ndash51) 558 (48ndash61) lt0001 462 (42ndash51) 519 (45ndash58) 0003
Durationa 149 (9ndash19) NA mdash 121 (8ndash14) 196 (14ndash25) lt0001 126 (9ndash15) 145 (10ndash21) 009
EDSS score 3 (2ndash375) NA mdash 25 (2ndash3) 4 (35ndash65) lt0001 25 (2ndash3) 35 (3ndash5) 0005
Relapsesb 4 (2ndash6) 3 (2ndash6) 6 (3ndash8) 002 4 (2ndash6) 3 (2ndash625) 10
DMT (yesno) 5425 NA mdash 4511 914 lt0001 3312 2113 033
First linec 28 NA mdash 27 1 mdash 18 10
Second lined 26 NA mdash 18 8 mdash 15 11
NfL (pgmL) 217(15ndash31)
231(21ndash28)
03 188 (14ndash26) 314 (22ndash44) 0001 16 (12ndash19) 319 (29ndash43) lt0001
Volume (cm3)
WM 4521(412ndash485)
4928(463ndash525)
0009 4642 (423ndash501) 4227 (372ndash463) 0002 4521 (415ndash485) 4542 (406ndash481) 06
NAWM 4445(396ndash479)
NA mdash 4578 (407ndash498) 4107 (339ndash446) 0001 4445 (396ndash477) 4345 (391ndash479) 06
GMctx 4288(409ndash465)
473(450ndash512)
0006 4305 (414ndash470) 4163 (392ndash450) 009 4294 (409ndash464) 4246 (397ndash464) 04
T1 lesion 34 (2ndash8) NA mdash 27 (2ndash6) 65 (4ndash10) 0007 31 (2ndash9) 41 (2ndash7) 03
T2 lesion 69 (3ndash18) NA mdash 49 (2ndash12) 15 (13ndash25) lt0001 56 (2ndash15) 96 (4ndash19) 02
Cingulate DTI
FA 046(043ndash048)
049(047ndash050)
0007 046 (044ndash049) 044 (042ndash048) 01 046 (044ndash048) 047 (042ndash048) 09
MD (mm2s) 091(086ndash096)
083(082ndash087)
0005 090 (084ndash094) 092 (088ndash011) 005 091 (086ndash095) 091 (083ndash10) 07
AD (mm2s) 141(136ndash146)
135(131ndash139)
0028 139 (135ndash144) 144 (137ndash155) 003 141 (137ndash145) 140 (13ndash15) 07
RD (mm2s) 065(059ndash071)
058(056ndash062)
0004 064 (059ndash070) 069 (063ndash081) 005 066 (060ndash071) 064 (058ndash075) 07
Abbreviations AD = axial diffusivity DMT = disease-modifying treatment DTI = diffusion tensor imaging EDSS = Expanded Disability Status Scale FA =fractional anisotropy GMctx = cortical graymatter HC = healthy control MD=meandiffusivity NA = not applicable NAWM=normal-appearingwhitematterNfL = neurofilament light chain RD = radial diffusivity RRMS = relapsing-remitting MS SPMS = secondary progressive MS WM = white matterPatients withMSwere divided into NfL(low) andNfL(high) subgroups based on themedian value of HCs (231 pgmL) Median (interquartile range) values areshown except for sex and DMT where the number of patients is shown The p values are from the Wilcoxon rank-sum test for continuous variables and theFisher exact test for class variables Significant p values are bolded Values of mean axial and radial diffusivities are expressed as e10minus3a Disease duration (years) calculated from the onset of symptomsb Total number of relapses before study onset data of 2 patients are missingc First-line DMT = interferon-beta dimethyl fumarate glatiramer acetate and teriflunomided Second-line DMT = fingolimod natalizumab and rituximab
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 3
groups and between patients and healthy controls wereassessed using theWilcoxon rank-sum (Mann-WhitneyU) testHolm multiple comparison adjustment was used in RRMS vssecondary progressive MS (SPMS) vs healthy control multiplecomparisons The division into NfL(low) and NfL(high)subgroups was based on the median NfL value of the healthycontrols
To test our main hypothesis Spearman correlations were calcu-lated to assess the relationships between NfL and DTI indices(fractional anisotropy mean diffusivity radial diffusivity and axialdiffusivity) in the whole NAWM and in NAWM subregionsSpearman correlation was used instead of Pearson correlation toavoid outlier-induced overestimation of correlations The p valuesof correlation analyses were further adjusted with the false dis-covery rate method (Benjamini-Hochberg procedure) for thenumber of investigated parameters (n = 28)20 The NfL valueswere further modeled by the DTI indices and volumetric MRIdata using multiple regression models The logarithm of NfL wasused as the response because nontransformed values led to non-normality of residuals The models were adjusted by sex agedisease type (relapsing-remittingsecondary progressive) theEDSS score disease-modifying treatment (no treatment first line[dimethyl fumarate glatiramer acetate interferon-beta and teri-flunomide] or second line [fingolimod natalizumab and ritux-imab] treatment) and the presence of relapses within 1 yearbefore sampling (yesno) The normality of the residuals waschecked using the Shapiro-Wilk test Variance inflation factorvalues were used to check that independent variables were nothighly correlated with each other Regression coefficients were
standardized tomake themdirectly comparable to each other Forstandardization the regression results weremultiplied by twice theSD of the DTI or volume variables21 The p values of the multipleregression model were adjusted with the false discovery ratemethod for the number of investigated variables (n = 32)
In addition to verify the quality of the data and increase thegeneralizability and comparability of the results we usedSpearman correlation to assess the relationships betweenfollowing parameters NfL with age disease duration theEDSS score parenchymal fractions of NAWM cortical graymatter and T1 and T2 lesion volumes DTI with the EDSSscore and parenchymal fractions of NAWM cortical graymatter and T1 and T2 lesion volumes (results are shown inappendix e-3 linkslwwcomNXIA355) Spearman cor-relation was used instead of Pearson correlation to avoidoutlier-induced overestimation of correlations and becauseSpearman correlation is preferred in case of ordinal vari-ables ie EDSS score All statistical tests were 2 tailed andp = 005 was used as the threshold for statistical significance
Data availabilityThe anonymized raw data will be shared over the next 3 yearson request from a qualified investigator
ResultsThe demographics clinical characteristics and conventionalimaging data of the 79 patients with MS included in the study
Figure 1 Correlation of serum NfL and DTI indices of the whole NAWM in the NfL(high) subgroup
The serum level of NfL is moderately associatedwith fractional anisotropy (A) mean diffusivity(B) axial diffusivity (C) and radial diffusivity (D)The NfL(high) subgroup is comprised of patientswith serum NfL above the median value ofhealthy controls (231 pgmL) Shown areSpearman correlation coefficients (ρ) and p val-ues The black lines indicate the level and di-rection of the relationship AD = axial diffusivityDTI = diffusion tensor imaging FA = fractionalanisotropy MD = mean diffusivity NAWM =normal-appearing white matter NfL = neuro-filament light chain RD = radial diffusivity
4 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
Table 2 Spearman correlations between serum NfL concentrations and DTI-MRI indices
All MS NfL(low) NfL(high)
ρ p Value ρ p Value ρ p Value
Entire NAWM
FA minus016 02 minus020 02 minus042 0013
MD 012 03 014 04 047 0005
AD 001 09 005 07 041 0017
RD 013 03 015 03 045 0008
Frontal NAWM
FA minus018 01 minus020 02 minus037 0033
MD 017 01 011 05 047 0005
AD 009 04 004 08 044 0010
RD 015 02 013 04 044 0009
Parietal NAWM
FA minus009 04 minus011 05 minus034 0046
MD 005 07 005 08 043 0011
AD minus004 08 minus003 08 034 0053
RD 007 05 008 06 044 0009
Temporal NAWM
FA minus019 01 minus012 04 minus033 0060
MD 014 02 008 06 043 0012
AD minus003 08 004 08 026 0139
RD 014 02 009 07 038 0026
Occipital NAWM
FA minus009 04 minus018 02 minus043 0011
MD 012 03 017 03 053 0001
AD 004 07 014 04 038 0029
RD 010 04 015 03 048 0004
Cingulate NAWM
FA minus010 04 minus006 07 minus048 0004
MD 008 05 002 09 049 0003
AD 009 04 010 05 046 0007
RD 008 05 004 08 050 0003
Deep NAWM
FA minus014 02 minus020 03 minus036 0036
MD 008 05 010 06 033 0053
AD 003 08 001 09 036 0039
RD 01 04 012 05 036 0036
Abbreviations AD = axial diffusivity DTI = diffusion tensor imaging FA = fractional anisotropy MD = mean diffusivity NAWM = normal-appearing whitematter NfL = neurofilament light chain RD = radial diffusivityPatients with MS were divided into NfL(low) and NfL(high) subgroups based on the median value of healthy controls (231 pgmL) Fractional anisotropy andmean axial and radial diffusivities of the entire NAWMand of 6 parcellated subregions of NAWMwere correlated with serumNfL level in all patients with MSand in NfL(low) and NfL(high) subgroups Shown are Spearman correlation coefficients (ρ) and uncorrected p values Significant p values are boldedSignificance of these correlations was sustained after adjustment using the false discovery ratemethod for the number of DTI parameters (n = 28) except forparietal FA
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 5
are shown in table 1 The median (interquartile range [IQR])age of this cohort was 483 (43ndash53) years which is similar tothe age of the healthy individuals in the control group (476[44ndash53] p = 1) Overall the patient cohort was representedwith a quite stable disease as only 13 of the patients (9 RRMSand 1 SPMS) had had a relapse within 1 year before samplingSixty-eight percent of the patients (45 RRMS and 9 SPMS)were on disease-modifying therapy (dimethyl fumarate n = 3fingolimod n = 14 glatiramer acetate n = 5 interferon beta-1a n= 6 natalizumab n = 8 rituximab n = 4 and teriflunomide n =14) In MRI white matter and cortical gray matter volumeswere decreased in patients compared with healthy controlsDTI measures were obtained from both the entire cerebralNAWM and from 6 different NAWM subregions In the cin-gulate area there were significant differences in all 4 DTI pa-rameters in patients with MS compared with healthy controlsfractional anisotropy was decreased whereas mean diffusivityradial diffusivity and axial diffusivity were increased (table 1)
Serum NfL levels in patients with MS andhealthy controlsThe median (IQR) NfL level was higher in patients withSPMS compared with patients with RRMS (314 [22ndash44] vs188 [14ndash26] pgmL p = 0001 table 1) NfL levels in thewhole MS group were not different from healthy controls(217 [15ndash31] vs 231 [21ndash28] pgmL p = 03) regardlesswhether the analysis was performed with age correctionAcute inflammation did not appear to be a factor to affect theNfL level in this cohort with relatively modest acute in-flammatory activity as the median NfL value among thepatients with or without relapse within the previous year wascomparable (223 [16ndash32] vs 186 [15ndash25] pgmL p = 03Wilcoxon rank-sum data not shown)
Characterization of patients with highNfL levelsTo explore associations between increased NfL levels and DTI-measurable diffuse neuroaxonal damage patients with MS weredivided into NfL(high) and NfL(low) subgroups The divisionwas based on the median NfL value measured among healthycontrols (231 [21ndash28] pgmL) In the NfL(high) subgroup theNfL concentration was significantly elevated compared withhealthy controls (p = 0018 table 1) and in the further NfL vsDTI correlation analyses we focused on the NfL(high) subgroupBased on the demographic and clinical data the patients in theNfL(high) subgroupwere at amore advanced stage of the diseaseTheNfL(high) subgroup includedmore patients with SPMS (p =0003) and the patients in the subgroup were also older (p =0003) and had a higher EDSS score (p = 0005 table 1) Inevaluation using conventional MRI the NAWM white matter orcortical gray matter volumes and T1 and T2 lesion loads were notdifferent between NfL(low) and NfL(high) subgroups (table 1)
Associations between NfL and DTIOur results show that DTI metrics of diffuse neuroaxonaldamage within theNAWMassociatewith high serumNfL levelsin MS As a demonstration of this in the NfL(high) subgroup
several (n = 24) significant correlations were found betweenserum NfL levels and NAWM DTI indices both in the entireNAWM and in various brain subregions higher NfL levels wereassociated with lower fractional anisotropy and higher diffusivity(mean axial and radial) in the whole NAWM (figure 1) and inall its subregions except for fractional anisotropy in temporalNAWM axial diffusivity in parietal and temporal NAWM andmean diffusivity in remaining NAWM (table 2) Significance ofthese correlations was sustained after adjustment using the falsediscovery rate method for the number of DTI parameters (n =28) except for parietal fractional anisotropy Results remainedsimilar when the data were analyzed without patients who hadhad a relapse within the previous year before sampling (n = 3data not shown) No correlations between serumNfL levels andDTI metrics were observed in the NfL(low) subgroup or in theoverall MS cohort (table 2)
Multiple regression modeling was performed to further evaluatethe effect of DTI on NfL In the NfL(high) subgroup the DTIindices weremore significantly associated withNfL thanwith theclinical parameters included in the model (sex age disease typeEDSS score treatment and the presence of relapses within 1year before sampling) (table 3) On average 48 of the variancein NfL (SD 44) could be explained by DTI indices and theclinical parameters mentioned above (figure 2) There weremultiple (n = 26) significant associations between NfL and DTIin the whole NAWM and in all its 6 subregions All except 2 ofthe associations remained significant after adjustment with thefalse discovery rate method for the number of investigated var-iables (n = 32) The strongest associations were observed in thecingulate area between higher NfL and higher mean and radialdiffusivity and in deep and frontal NAWM between higher NfLand higher mean diffusivity (figure 2) Analyzing the datawithout the patients with a relapse (n = 3) within the previousyear did not alter the associations (data not shown)
Higher NfL was also associated with lower NAWM andcortical gray matter and higher T2 lesion volumes (figure 2)The volumes and clinical adjustments explained 434494 and 433 of the variance in NfL respectively
DiscussionPresent results demonstrate that the DTI-MRI measures ofNAWM correlate with serumNfL in MS We found that amongpatients with more advanced disease increased serumNfL levelsassociate with DTI measures reflecting diffuse microstructuraldamage ie decreased fractional anisotropy and increasedmeanaxial and radial diffusivities in the NAWM In a multiple re-gression model which was adjusted with sex age disease typeEDSS score treatment and the presence of relapses the DTIindices were more significantly associated with NfL than theabove-mentioned demographic and clinical parameters
The pathophysiology of MS involves acute and chronic mech-anisms that lead to gradual axonal and myelin damage that
6 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
Table 3 Multiple regression analyses evaluating separately the effect of DTI-MRI indices or volumetric brain MRI data onNfL in the NfL(high) subgroup of patients with MS
Variable in model DTI Sex Age Type EDSS score DMT Relapses
Entire NAWM
FA 0018 07 08 10 03 04 02
MD 0003 07 09 10 02 04 04
AD 0010 09 10 08 01 07 05
RD 0005 07 10 10 02 04 03
Frontal NAWM
FA 0045 08 10 10 02 07 02
MD 0001 05 07 10 02 06 04
AD 0004 05 09 08 03 07 04
RD 0003 06 09 09 02 06 03
Parietal NAWM
FA 0023 09 09 09 02 05 01
MD 0004 07 07 08 01 05 02
AD 0017 08 10 10 01 06 03
RD 0006 08 08 09 02 05 02
Temporal NAWM
FA 0116 07 07 10 03 05 03
MD 0007 06 10 10 01 05 05
AD 0009 07 08 08 01 05 05
RD 0020 06 09 10 02 05 04
Occipital NAWM
FA 0011 09 06 09 03 03 01
MD 0009 08 08 09 01 06 02
AD 0086 08 07 08 01 08 03
RD 0011 08 08 08 02 05 02
Cingulate NAWM
FA 0004 05 04 08 06 04 03
MD 0001 06 09 08 06 02 07
AD 0003 08 09 08 03 04 07
RD 0001 06 07 08 07 02 06
Deep NAWM
FA 0047 05 06 09 03 05 03
MD 0001 07 09 09 05 02 07
AD 0005 10 08 10 03 04 08
RD 0002 06 08 10 05 02 08
Volume (pf)
NAWM 0019 08 08 08 03 05 05
Continued
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 7
eventually manifest by brain atrophy and clinically by steadyworsening of disability in almost all patients over time This diffusepathology which in the majority of cases leads to progressiveworsening of the MS-related symptoms is now the focus of in-terest for patients and physicians as modern disease-modifyingtherapies have led to an almost complete suppression of relapseactivity while their impact on progression is at best modest Ourincomplete understanding of the molecular mechanisms that leadto progression is one of the main reasons for the failure in de-velopment ofmore efficacious therapies for secondary progressiveform of MS Another impediment is the relative insensitivity ofcurrent measurement tools to capture features of progression atsubclinical stages and hence to quantitate the effects of anti-neurodegenerative treatments at a time point where the overallbrain structure is still largely intact Specifically conventional MRIcan only quantitate brain volume loss as an end result of thedisease but does not provide insights into the underlying micro-structural changes DTI is a nonconventional MRI technique thatprovides a measure of such diffuse changes beyond volumetry
NfL is a highly sensitive biomarker to detect neuronal damageand is the first of its kind applicable in blood-derived probes22
However the cause and underlying pathology of the increasedNfL concentration may vary greatly both inter- and intra-individually In clinically isolated syndrome and RRMS acuteNfL elevation is particularly associated with focal in-flammation ie relapses and focal lesion formation4ndash823
However the indication for the contribution of diffuse pa-thology to higher NfL levels has so far been only incidentaland consisted of associations of increased NfL with moreadvanced clinical disability and brain atrophy824 Novelfindings presented here provide more direct evidence for theeffect of diffuse pathologic process in the NAWM on in-creased serum NfL We observed that among a subgroup ofolder and more disabled patients NfL associates both withradial and axial diffusivities In the context of MS these DTIscalars are thought to reflect myelin loss and axonal damagerespectively11 NfL also associated with mean diffusivity thatcan also indicate on axonal andmyelin loss11 The associationswere most pronounced in the cingulate deep and frontalNAWM
NfL differs in 2 important aspects from MRI first as asignal of neurodegeneration it reflects ongoing pathologicprocesses in real time while imaging captures morphologicfeatures that are inherently retrospective Second NfLlevels reflect neuronal damage in the CNS comprehen-sively ie in addition to brain pathology it also capturesspinal cord pathology that is not routinely evaluated byimaging in the workup of MS for individual patients As adownside NfL levels cannot differentiate betweenacute inflammatory and chronic neurodegenerative diseaseactivity Our results suggest that DTI-MRI allows us tocategorize the pathogenic source of NfL at least semi-quantitatively We demonstrate that both diffuse axonaldamage within the NAWM measured using DTI and focalinflammatory white matter lesion load measured usingconventional MRI contributed to the elevated NfL serumlevels This highlights that both focal and diffuse pathologicchanges must be taken into account when levels of NfL inindividual patients are interpreted A somewhat largerproportion of NfL variance was explained by DTI indices inaverage than by T2 lesion load This implies that in non-relapsing patients the ongoing diffuse microstructuraldamage in the NAWM outside lesions might predominateover focal axonal damage that has occurred within lesionsas a source of NfL in serum Accordingly in other neuro-logic disease with brain diffuse pathology such as amyo-trophic lateral sclerosis Alzheimer disease traumatic braininjury and age-related white matter pathology elevatedserum NfL levels have been associated with DTI-measurable diffuse damage within the CNS25ndash28 Thesestudies are further evidence for the concept that braindiffuse chronic neuropathologic mechanisms contribute toNfL release and that nonconventional MRI techniques areable to identify the CNS areas of its morphologic source
Current results demonstrate that the cingulate area is thebrain area with the most prominent white matter tractdamage outside focal lesions Periventricular regions arecharacteristically the brain area with a particularly heavylesion load in MS29 We addressed the question whetherthe cingulum-NAWM DTI abnormalities would arise from
Table 3 Multiple regression analyses evaluating separately the effect of DTI-MRI indices or volumetric brain MRI data onNfL in the NfL(high) subgroup of patients with MS (continued)
Variable in model DTI Sex Age Type EDSS score DMT Relapses
GMctx 0004 09 02 04 003 03 06
T2 0019 06 03 07 04 07 03
T1 015 07 04 10 05 07 03
Abbreviations AD = axial diffusivity DMT = disease-modifying treatment DTI = diffusion tensor imaging EDSS = Expanded Disability Status Scale FA =fractional anisotropy GMctx = cortical gray matter MD = mean diffusivity NAWM = normal-appearing white matter NfL = neurofilament light chain pf =parenchymal fraction RD = radial diffusivityPatients withMSwere divided into NfL(low) and NfL(high) subgroups based on themedian NfL value of healthy controls (231 pgmL) Model was adjusted bysex age disease type (relapsing-remittingsecondary progressive) EDSS score treatment (no first- and second-line) and presence of relapses within 1 yearbefore sampling (yesno) Bolded p values indicate that the parameter affected the logarithmofNfL value statistically significantly whenDTI or volume and allthe adjusting variables were in the models
8 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
the dirty-appearing white matter due to close proximity to aheavy lesion load but the relative lesion volume in the cingulatearea was not found to be significantly larger compared withlesion load in other regions or in the whole NAWM Thereforeit seems unlikely that differences in the DTI indices would besolely due to dirty-appearing white matter adjacent to lesionswhereas it is more likely that remote lesions lead to DTI changesin the cingulateNAWM throughWallerian degeneration30 Alsoother studies have observed diffusion changes in the cingulatearea of patients with MS31ndash33 Because the cingulum bundle offibers projects from the cingulate gyrus to the entorhinal cortexin the brain allowing for communication between components
of the limbic system the DTI abnormalities may be the mor-phologic substrate for impaired cognitive function and fatiguesymptoms that are typical in progressive MS3435
Present results are in line with previous observations on therelation between NfL concentration and MS disease sub-type with progressive patients having greater NfL levelscompared with relapsing-remitting patients568ndash102324
However the median NfL value of the entire patient cohortwas relatively low compared with healthy controls which isthe main limitation of our study Because of this we werenot able to use the previously defined cutoff values3 for
Figure 2 The effect of DTI-MRI indices or volumetric MRI data on NfL in the NfL(high) subgroup
The NfL(high) subgroup is comprised of patients with serum NfL above the median value of healthy controls (231 pgmL) Logarithmic serum NfL wasmodeled separately by DTI indices of the entire and parcellated NAWM and volumetric brain MRI data using multiple regression analyses Models wereadjusted by sex age disease type (RRMSSPMS) the EDSS score treatment (no first and second line) and presence of relapseswithin 1 year before sampling(yesno) The results are illustrated using dot and whisker plots in which red dots represent standardized regression coefficients and red lines represent theCIs of the estimates Significant p values of the DTI parameters and the percentage of variance in the response that can be explained by the independentvariables (R2) are also shown The results were ordered according to the R2 value All except 2 of the p values (fractional anisotropy of the frontal and deepNAWM) shown in the figure remained significant after adjustment using the false discovery ratemethod for the number of investigated variables (n = 32) AD =axial diffusivity DTI = diffusion tensor imaging EDSS = Expanded Disability Status Scale FA = fractional anisotropy GMctx = cortical gray matter MD =meandiffusivity NAWM = normal-appearing white matter NfL = neurofilament light chain PF = parenchymal fraction RD = radial diffusivity RRMS = relapsing-remitting MS SPMS = secondary progressive MS WM = white matter
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 9
NfL(low) and NfL(high) subgroups Instead we used themedian value of healthy controls Although the divider usedin our study was based on a small group of individualssimilar values have been observed in previous largerstudies68 One reason for the low median NfL level couldbe that the patients with RRMS are represented with arelatively benign disease course In addition our patientswith RRMS were sampled during a relapse-free time andhad no signs of ongoing focal inflammatory activity in theMRI Moreover most of the patients (68) were using adisease-modifying treatment at the time of evaluationHence similar as observed in other studies35233637 thisresulted in low serum NfL concentration Following thesimilarity in the NfL levels between healthy controls andpatients with MS the main results of our study are based ona relatively low number of patients In addition the patientnumbers under different immunomodulatory treatmentsvaried substantially However the heterogeneity of thecohort was taken account in the multiple regression modelwhere both EDSS score and disease-modifying treatmentwere used as adjustments
Our findings suggest that elevated serum NfL in MS resultsfrom neuroaxonal damage both within the NAWM and infocal T2 lesions The association between DTI-measurablediffuse microstructural white matter damage and serumNfL isfurther conceptual evidence for the latter being a usefulmonitoring tool in assessing the degree of ongoing neuro-degenerative processes in MS
AcknowledgmentThe authors thank all patients with MS and healthycontrols who participated in this study They also thank thepersonnel of Turku PET Centre for their excellenttechnical assistance
Study fundingThis work was supported by the Finnish Academy the SigridJuselius Foundation
DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgNN for full disclosures
Publication historyReceived by Neurology Neuroimmunology amp NeuroinflammationMay 8 2020 Accepted in final form October 21 2020
References1 Reich DS Lucchinetti CF Calabresi PA Multiple sclerosis N Engl J Med 2018378
169ndash1802 Varhaug KN Torkildsen Oslash Myhr KM Vedeler CA Neurofilament light chain as a
biomarker in multiple sclerosis Front Neurol 201910338
Appendix Authors
Author Location Contribution
MaijaSaraste PhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Acquisition of dataanalyzed andinterpreted the data anddrafted and revised themanuscript forintellectual content
Appendix (continued)
Author Location Contribution
SvetlanaBezukladovaMSc
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Acquisition of dataanalyzed and interpretedthe data and revised themanuscript
MarkusMatilainenPhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Statistical analysisdrafted the figures andrevised the manuscript
Jouni TuiskuMSc
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Major role in theacquisition of data andrevised the manuscript
EeroRissanenMDPhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Designed andconceptualized the studyand acquisition of data
MarcusSucksdorffMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Major role in theacquisition of data
SiniLaaksonenMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Acquisition of data
AnnaVuorimaaMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Acquisition of data
Jens KuhleMD PhD
Neurologic Clinic andPoliclinic Departments ofMedicine Biomedicineand Clinical ResearchUniversity Hospital BaselSwitzerland
Major role in theacquisition of data
DavidLeppert MDPhD
Neurologic Clinic andPoliclinic Departments ofMedicine Biomedicineand Clinical ResearchUniversity Hospital BaselSwitzerland
Revised the manuscriptfor intellectual content
Laura AirasMD PhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Designed andconceptualized the studyand revised themanuscript for intellectualcontent
10 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
3 Kuhle J Kropshofer H Haering DA et al Blood neurofilament light chain as abiomarker of MS disease activity and treatment response Neurology 201992e1007ndashe1015
4 Varhaug KN Barro C Bjoslashrnevik K et al Neurofilament light chain predicts diseaseactivity in relapsing-remitting MS Neurol Neuroimmunol Neuroinflamm 20185e422
5 Novakova L Zetterberg H Sundstrom P et al Monitoring disease activity in multiplesclerosis using serum neurofilament light protein Neurology 2017892230ndash2237
6 Disanto G Barro C Benkert P et al Serum neurofilament light a biomarker ofneuronal damage in multiple sclerosis Ann Neurol 201781857ndash870
7 Siller N Kuhle J Muthuraman M et al Serum neurofilament light chain is a bio-marker of acute and chronic neuronal damage in early multiple sclerosis Mult Scler201925678ndash686
8 Barro C Benkert P Disanto G et al Serum neurofilament as a predictor of diseaseworsening and brain and spinal cord atrophy in multiple sclerosis Brain 20181412382ndash2391
9 Ferraro D Guicciardi C De Biasi S et al Plasma neurofilaments correlate withdisability in progressive multiple sclerosis patients Acta Neurol Scand 202014116ndash21
10 Jakimovski D Zivadinov R RamanthanM et al Serum neurofilament light chain levelassociations with clinical and cognitive performance in multiple sclerosis a longitu-dinal retrospective 5-year study Mult Scler 2020261670ndash1681
11 Sbardella E Tona F Petsas N Pantano P DTI measurements in multiple sclerosisevaluation of brain damage and clinical implications Mult Scler Int 20132013671730
12 Tae WS Ham BJ Pyun SB Kang SH Kim BJ Current clinical applications ofdiffusion-tensor imaging in neurological disorders J Clin Neurol 201814129ndash140
13 Kurtzke JF Rating neurologic impairment in multiple sclerosis an expanded disabilitystatus scale (EDSS) Neurology 1983331444ndash1452
14 Kuhle J Barro C Andreasson U et al Comparison of three analytical platforms forquantification of the neurofilament light chain in blood samples ELISA electro-chemiluminescence immunoassay and SimoaClinChemLabMed 2016541655ndash1661
15 Bezukladova S Tuisku J Matilainen M et al Insights into disseminated MS brainpathology with multimodal diffusion tensor and PET imaging Neurol NeuroimmunolNeuroinflamm 20207e691
16 Schmidt P Gaser C Arsic M et al An automated tool for detection of FLAIR-hyperintense white-matter lesions in Multiple Sclerosis Neuroimage 2012593774ndash3783
17 Salat DH Greve DN Pacheco JL et al Regional white matter volume differences innondemented aging and Alzheimerrsquos disease Neuroimage 2009441247ndash1258
18 Rissanen E Tuisku J Vahlberg T et al Microglial activation white matter tractdamage and disability in MS Neurol Neuroimmunol Neuroinflamm 20185e443
19 Leemans A Jeurissen B Sijbers J et al ExploreDTI a graphical toolbox for pro-cessing analysing and visualizing diffusion MR data Presented at the 17th AnnualMeeting of International Society for Magnetic Resonance in Medicine Honolulu HI2009 p 3537
20 Benjamini Y Hochberg Y Controlling the false discovery rate a practical and pow-erful approach to multiple testing J R Stat Soc Ser B 199557289ndash300
21 Gelman A Scaling regression inputs by dividing by two standard deviations Stat Med2008272865ndash2873
22 Khalil M Teunissen CE Otto M et al Neurofilaments as biomarkers in neurologicaldisorders Nat Rev Neurol 201814577ndash589
23 Canto E Barro C Zhao C et al Association between serum neurofilament light chainlevels and long-term disease course among patients with multiple sclerosis followedup for 12 years JAMA Neurol 2019761359ndash1366
24 Hogel H Rissanen E Barro C et al Serum glial fibrillary acidic protein correlates withmultiple sclerosis disease severity Mult Scler 202026210ndash219
25 Menke RA Gray E Lu CH et al CSF neurofilament light chain reflects corticospinaltract degeneration in ALS Ann Clin Transl Neurol 20152748ndash755
26 Moore EE Hohman TJ Badami FS et al Neurofilament relates to white mattermicrostructure in older adults Neurobiol Aging 201870233ndash241
27 Ljungqvist J Zetterberg H Mitsis M Blennow K Skoglund T Serum neurofilamentlight protein as a marker for diffuse axonal injury results from a case series studyJ Neurotrauma 2017341124ndash1127
28 Schultz SA Strain JF Adedokun A et al Serum neurofilament light chain levels areassociated with white matter integrity in autosomal dominant Alzheimerrsquos diseaseNeurobiol Dis 2020142104960
29 Narayanan S Fu L Pioro E et al Imaging of axonal damage in multiple sclerosisspatial distribution of magnetic resonance imaging lesions Ann Neurol 199741385ndash391
30 Ciccarelli O Werring DJ Barker GJ et al A study of the mechanisms of normal-appearing white matter damage in multiple sclerosis using diffusion tensor imagingevidence of Wallerian degeneration J Neurol 2003250287ndash292
31 Pokryszko-Dragan A Banaszek A Nowakowska-Kotas M et al Diffusion tensorimaging findings in the multiple sclerosis patients and their relationships to variousaspects of disability J Neurol Sci 2018391127ndash133
32 Kolasa M Hakulinen U Brander A et al Diffusion tensor imaging and disabilityprogression in multiple sclerosis a 4-year follow-up study Brain Behav 20199e01194
33 Andersen O Hildeman A Longfils M et al Diffusion tensor imaging in multiplesclerosis at different final outcomes Acta Neurol Scand 2018137165ndash173
34 Koenig KA Sakaie KE Lowe MJ et al The relationship between cognitive functionand high-resolution diffusion tensor MRI of the cingulum bundle in multiple sclerosisMult Scler 2015211794ndash1801
35 Pardini M Bonzano L Bergamino M et al Cingulum bundle alterations underliesubjective fatigue in multiple sclerosis Mult Scler 201521442ndash447
36 Sejbaek T Nielsen HH Penner N et al Dimethyl fumarate decreases neurofilamentlight chain in CSF and blood of treatment naıve relapsing MS patients J NeurolNeurosurg Psychiatry 2019901324ndash1330
37 Piehl F Kockum I Khademi M et al Plasma neurofilament light chain levels inpatients with MS switching from injectable therapies to fingolimod Mult Scler 2018241046ndash1054
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 11
DOI 101212NXI000000000000092620218 Neurol Neuroimmunol Neuroinflamm
Maija Saraste Svetlana Bezukladova Markus Matilainen et al High serum neurofilament associates with diffuse white matter damage in MS
This information is current as of December 8 2020
ServicesUpdated Information amp
httpnnneurologyorgcontent81e926fullhtmlincluding high resolution figures can be found at
References httpnnneurologyorgcontent81e926fullhtmlref-list-1
This article cites 36 articles 3 of which you can access for free at
Citations httpnnneurologyorgcontent81e926fullhtmlotherarticles
This article has been cited by 1 HighWire-hosted articles
Subspecialty Collections
httpnnneurologyorgcgicollectionmultiple_sclerosisMultiple sclerosis
httpnnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpnnneurologyorgmiscaboutxhtmlpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnnneurologyorgmiscaddirxhtmlreprintsusInformation about ordering reprints can be found online
Academy of Neurology All rights reserved Online ISSN 2332-7812Copyright copy 2020 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the AmericanPublished since April 2014 it is an open-access online-only continuous publication journal Copyright
is an official journal of the American Academy of NeurologyNeurol Neuroimmunol Neuroinflamm
induced distortions using the robust estimation of tensorsby outlier rejection tensor estimation method Fractionalanisotropy mean diffusivity radial diffusivity and axialdiffusivity maps were extracted from the corrected DTImaps Coregistering of maps to corresponding T1-weighted images was performed using SPM8 (The Well-come Centre for Human Neuroimaging University Col-lege London) running on MATLAB (The MathWorks
Natick MA) Finally the mean values for fractional an-isotropy mean diffusivity radial diffusivity and axial dif-fusivity within the whole NAWM and in the segmentedregions of NAWM were calculated in MATLAB
Statistical analysisThe statistical analysis was performed using R statistical soft-ware (version 400) The differences between different MS
Table 1 Demographics clinical characteristics volumetric brain MRI data and cingulate DTI-MRI indices of patients withMS and HCs
MS (n = 79) HC (n = 10)pValue RRMS (n = 56) SPMS (n = 23)
pValue
NfL(low)(n = 45)
NfL(high)(n = 34)
pValue
Femalemale 6019 73 07 4511 158 016 378 2311 018
Age 483(43ndash53)
476(44ndash53)
1 474 (42ndash51) 558 (48ndash61) lt0001 462 (42ndash51) 519 (45ndash58) 0003
Durationa 149 (9ndash19) NA mdash 121 (8ndash14) 196 (14ndash25) lt0001 126 (9ndash15) 145 (10ndash21) 009
EDSS score 3 (2ndash375) NA mdash 25 (2ndash3) 4 (35ndash65) lt0001 25 (2ndash3) 35 (3ndash5) 0005
Relapsesb 4 (2ndash6) 3 (2ndash6) 6 (3ndash8) 002 4 (2ndash6) 3 (2ndash625) 10
DMT (yesno) 5425 NA mdash 4511 914 lt0001 3312 2113 033
First linec 28 NA mdash 27 1 mdash 18 10
Second lined 26 NA mdash 18 8 mdash 15 11
NfL (pgmL) 217(15ndash31)
231(21ndash28)
03 188 (14ndash26) 314 (22ndash44) 0001 16 (12ndash19) 319 (29ndash43) lt0001
Volume (cm3)
WM 4521(412ndash485)
4928(463ndash525)
0009 4642 (423ndash501) 4227 (372ndash463) 0002 4521 (415ndash485) 4542 (406ndash481) 06
NAWM 4445(396ndash479)
NA mdash 4578 (407ndash498) 4107 (339ndash446) 0001 4445 (396ndash477) 4345 (391ndash479) 06
GMctx 4288(409ndash465)
473(450ndash512)
0006 4305 (414ndash470) 4163 (392ndash450) 009 4294 (409ndash464) 4246 (397ndash464) 04
T1 lesion 34 (2ndash8) NA mdash 27 (2ndash6) 65 (4ndash10) 0007 31 (2ndash9) 41 (2ndash7) 03
T2 lesion 69 (3ndash18) NA mdash 49 (2ndash12) 15 (13ndash25) lt0001 56 (2ndash15) 96 (4ndash19) 02
Cingulate DTI
FA 046(043ndash048)
049(047ndash050)
0007 046 (044ndash049) 044 (042ndash048) 01 046 (044ndash048) 047 (042ndash048) 09
MD (mm2s) 091(086ndash096)
083(082ndash087)
0005 090 (084ndash094) 092 (088ndash011) 005 091 (086ndash095) 091 (083ndash10) 07
AD (mm2s) 141(136ndash146)
135(131ndash139)
0028 139 (135ndash144) 144 (137ndash155) 003 141 (137ndash145) 140 (13ndash15) 07
RD (mm2s) 065(059ndash071)
058(056ndash062)
0004 064 (059ndash070) 069 (063ndash081) 005 066 (060ndash071) 064 (058ndash075) 07
Abbreviations AD = axial diffusivity DMT = disease-modifying treatment DTI = diffusion tensor imaging EDSS = Expanded Disability Status Scale FA =fractional anisotropy GMctx = cortical graymatter HC = healthy control MD=meandiffusivity NA = not applicable NAWM=normal-appearingwhitematterNfL = neurofilament light chain RD = radial diffusivity RRMS = relapsing-remitting MS SPMS = secondary progressive MS WM = white matterPatients withMSwere divided into NfL(low) andNfL(high) subgroups based on themedian value of HCs (231 pgmL) Median (interquartile range) values areshown except for sex and DMT where the number of patients is shown The p values are from the Wilcoxon rank-sum test for continuous variables and theFisher exact test for class variables Significant p values are bolded Values of mean axial and radial diffusivities are expressed as e10minus3a Disease duration (years) calculated from the onset of symptomsb Total number of relapses before study onset data of 2 patients are missingc First-line DMT = interferon-beta dimethyl fumarate glatiramer acetate and teriflunomided Second-line DMT = fingolimod natalizumab and rituximab
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 3
groups and between patients and healthy controls wereassessed using theWilcoxon rank-sum (Mann-WhitneyU) testHolm multiple comparison adjustment was used in RRMS vssecondary progressive MS (SPMS) vs healthy control multiplecomparisons The division into NfL(low) and NfL(high)subgroups was based on the median NfL value of the healthycontrols
To test our main hypothesis Spearman correlations were calcu-lated to assess the relationships between NfL and DTI indices(fractional anisotropy mean diffusivity radial diffusivity and axialdiffusivity) in the whole NAWM and in NAWM subregionsSpearman correlation was used instead of Pearson correlation toavoid outlier-induced overestimation of correlations The p valuesof correlation analyses were further adjusted with the false dis-covery rate method (Benjamini-Hochberg procedure) for thenumber of investigated parameters (n = 28)20 The NfL valueswere further modeled by the DTI indices and volumetric MRIdata using multiple regression models The logarithm of NfL wasused as the response because nontransformed values led to non-normality of residuals The models were adjusted by sex agedisease type (relapsing-remittingsecondary progressive) theEDSS score disease-modifying treatment (no treatment first line[dimethyl fumarate glatiramer acetate interferon-beta and teri-flunomide] or second line [fingolimod natalizumab and ritux-imab] treatment) and the presence of relapses within 1 yearbefore sampling (yesno) The normality of the residuals waschecked using the Shapiro-Wilk test Variance inflation factorvalues were used to check that independent variables were nothighly correlated with each other Regression coefficients were
standardized tomake themdirectly comparable to each other Forstandardization the regression results weremultiplied by twice theSD of the DTI or volume variables21 The p values of the multipleregression model were adjusted with the false discovery ratemethod for the number of investigated variables (n = 32)
In addition to verify the quality of the data and increase thegeneralizability and comparability of the results we usedSpearman correlation to assess the relationships betweenfollowing parameters NfL with age disease duration theEDSS score parenchymal fractions of NAWM cortical graymatter and T1 and T2 lesion volumes DTI with the EDSSscore and parenchymal fractions of NAWM cortical graymatter and T1 and T2 lesion volumes (results are shown inappendix e-3 linkslwwcomNXIA355) Spearman cor-relation was used instead of Pearson correlation to avoidoutlier-induced overestimation of correlations and becauseSpearman correlation is preferred in case of ordinal vari-ables ie EDSS score All statistical tests were 2 tailed andp = 005 was used as the threshold for statistical significance
Data availabilityThe anonymized raw data will be shared over the next 3 yearson request from a qualified investigator
ResultsThe demographics clinical characteristics and conventionalimaging data of the 79 patients with MS included in the study
Figure 1 Correlation of serum NfL and DTI indices of the whole NAWM in the NfL(high) subgroup
The serum level of NfL is moderately associatedwith fractional anisotropy (A) mean diffusivity(B) axial diffusivity (C) and radial diffusivity (D)The NfL(high) subgroup is comprised of patientswith serum NfL above the median value ofhealthy controls (231 pgmL) Shown areSpearman correlation coefficients (ρ) and p val-ues The black lines indicate the level and di-rection of the relationship AD = axial diffusivityDTI = diffusion tensor imaging FA = fractionalanisotropy MD = mean diffusivity NAWM =normal-appearing white matter NfL = neuro-filament light chain RD = radial diffusivity
4 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
Table 2 Spearman correlations between serum NfL concentrations and DTI-MRI indices
All MS NfL(low) NfL(high)
ρ p Value ρ p Value ρ p Value
Entire NAWM
FA minus016 02 minus020 02 minus042 0013
MD 012 03 014 04 047 0005
AD 001 09 005 07 041 0017
RD 013 03 015 03 045 0008
Frontal NAWM
FA minus018 01 minus020 02 minus037 0033
MD 017 01 011 05 047 0005
AD 009 04 004 08 044 0010
RD 015 02 013 04 044 0009
Parietal NAWM
FA minus009 04 minus011 05 minus034 0046
MD 005 07 005 08 043 0011
AD minus004 08 minus003 08 034 0053
RD 007 05 008 06 044 0009
Temporal NAWM
FA minus019 01 minus012 04 minus033 0060
MD 014 02 008 06 043 0012
AD minus003 08 004 08 026 0139
RD 014 02 009 07 038 0026
Occipital NAWM
FA minus009 04 minus018 02 minus043 0011
MD 012 03 017 03 053 0001
AD 004 07 014 04 038 0029
RD 010 04 015 03 048 0004
Cingulate NAWM
FA minus010 04 minus006 07 minus048 0004
MD 008 05 002 09 049 0003
AD 009 04 010 05 046 0007
RD 008 05 004 08 050 0003
Deep NAWM
FA minus014 02 minus020 03 minus036 0036
MD 008 05 010 06 033 0053
AD 003 08 001 09 036 0039
RD 01 04 012 05 036 0036
Abbreviations AD = axial diffusivity DTI = diffusion tensor imaging FA = fractional anisotropy MD = mean diffusivity NAWM = normal-appearing whitematter NfL = neurofilament light chain RD = radial diffusivityPatients with MS were divided into NfL(low) and NfL(high) subgroups based on the median value of healthy controls (231 pgmL) Fractional anisotropy andmean axial and radial diffusivities of the entire NAWMand of 6 parcellated subregions of NAWMwere correlated with serumNfL level in all patients with MSand in NfL(low) and NfL(high) subgroups Shown are Spearman correlation coefficients (ρ) and uncorrected p values Significant p values are boldedSignificance of these correlations was sustained after adjustment using the false discovery ratemethod for the number of DTI parameters (n = 28) except forparietal FA
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 5
are shown in table 1 The median (interquartile range [IQR])age of this cohort was 483 (43ndash53) years which is similar tothe age of the healthy individuals in the control group (476[44ndash53] p = 1) Overall the patient cohort was representedwith a quite stable disease as only 13 of the patients (9 RRMSand 1 SPMS) had had a relapse within 1 year before samplingSixty-eight percent of the patients (45 RRMS and 9 SPMS)were on disease-modifying therapy (dimethyl fumarate n = 3fingolimod n = 14 glatiramer acetate n = 5 interferon beta-1a n= 6 natalizumab n = 8 rituximab n = 4 and teriflunomide n =14) In MRI white matter and cortical gray matter volumeswere decreased in patients compared with healthy controlsDTI measures were obtained from both the entire cerebralNAWM and from 6 different NAWM subregions In the cin-gulate area there were significant differences in all 4 DTI pa-rameters in patients with MS compared with healthy controlsfractional anisotropy was decreased whereas mean diffusivityradial diffusivity and axial diffusivity were increased (table 1)
Serum NfL levels in patients with MS andhealthy controlsThe median (IQR) NfL level was higher in patients withSPMS compared with patients with RRMS (314 [22ndash44] vs188 [14ndash26] pgmL p = 0001 table 1) NfL levels in thewhole MS group were not different from healthy controls(217 [15ndash31] vs 231 [21ndash28] pgmL p = 03) regardlesswhether the analysis was performed with age correctionAcute inflammation did not appear to be a factor to affect theNfL level in this cohort with relatively modest acute in-flammatory activity as the median NfL value among thepatients with or without relapse within the previous year wascomparable (223 [16ndash32] vs 186 [15ndash25] pgmL p = 03Wilcoxon rank-sum data not shown)
Characterization of patients with highNfL levelsTo explore associations between increased NfL levels and DTI-measurable diffuse neuroaxonal damage patients with MS weredivided into NfL(high) and NfL(low) subgroups The divisionwas based on the median NfL value measured among healthycontrols (231 [21ndash28] pgmL) In the NfL(high) subgroup theNfL concentration was significantly elevated compared withhealthy controls (p = 0018 table 1) and in the further NfL vsDTI correlation analyses we focused on the NfL(high) subgroupBased on the demographic and clinical data the patients in theNfL(high) subgroupwere at amore advanced stage of the diseaseTheNfL(high) subgroup includedmore patients with SPMS (p =0003) and the patients in the subgroup were also older (p =0003) and had a higher EDSS score (p = 0005 table 1) Inevaluation using conventional MRI the NAWM white matter orcortical gray matter volumes and T1 and T2 lesion loads were notdifferent between NfL(low) and NfL(high) subgroups (table 1)
Associations between NfL and DTIOur results show that DTI metrics of diffuse neuroaxonaldamage within theNAWMassociatewith high serumNfL levelsin MS As a demonstration of this in the NfL(high) subgroup
several (n = 24) significant correlations were found betweenserum NfL levels and NAWM DTI indices both in the entireNAWM and in various brain subregions higher NfL levels wereassociated with lower fractional anisotropy and higher diffusivity(mean axial and radial) in the whole NAWM (figure 1) and inall its subregions except for fractional anisotropy in temporalNAWM axial diffusivity in parietal and temporal NAWM andmean diffusivity in remaining NAWM (table 2) Significance ofthese correlations was sustained after adjustment using the falsediscovery rate method for the number of DTI parameters (n =28) except for parietal fractional anisotropy Results remainedsimilar when the data were analyzed without patients who hadhad a relapse within the previous year before sampling (n = 3data not shown) No correlations between serumNfL levels andDTI metrics were observed in the NfL(low) subgroup or in theoverall MS cohort (table 2)
Multiple regression modeling was performed to further evaluatethe effect of DTI on NfL In the NfL(high) subgroup the DTIindices weremore significantly associated withNfL thanwith theclinical parameters included in the model (sex age disease typeEDSS score treatment and the presence of relapses within 1year before sampling) (table 3) On average 48 of the variancein NfL (SD 44) could be explained by DTI indices and theclinical parameters mentioned above (figure 2) There weremultiple (n = 26) significant associations between NfL and DTIin the whole NAWM and in all its 6 subregions All except 2 ofthe associations remained significant after adjustment with thefalse discovery rate method for the number of investigated var-iables (n = 32) The strongest associations were observed in thecingulate area between higher NfL and higher mean and radialdiffusivity and in deep and frontal NAWM between higher NfLand higher mean diffusivity (figure 2) Analyzing the datawithout the patients with a relapse (n = 3) within the previousyear did not alter the associations (data not shown)
Higher NfL was also associated with lower NAWM andcortical gray matter and higher T2 lesion volumes (figure 2)The volumes and clinical adjustments explained 434494 and 433 of the variance in NfL respectively
DiscussionPresent results demonstrate that the DTI-MRI measures ofNAWM correlate with serumNfL in MS We found that amongpatients with more advanced disease increased serumNfL levelsassociate with DTI measures reflecting diffuse microstructuraldamage ie decreased fractional anisotropy and increasedmeanaxial and radial diffusivities in the NAWM In a multiple re-gression model which was adjusted with sex age disease typeEDSS score treatment and the presence of relapses the DTIindices were more significantly associated with NfL than theabove-mentioned demographic and clinical parameters
The pathophysiology of MS involves acute and chronic mech-anisms that lead to gradual axonal and myelin damage that
6 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
Table 3 Multiple regression analyses evaluating separately the effect of DTI-MRI indices or volumetric brain MRI data onNfL in the NfL(high) subgroup of patients with MS
Variable in model DTI Sex Age Type EDSS score DMT Relapses
Entire NAWM
FA 0018 07 08 10 03 04 02
MD 0003 07 09 10 02 04 04
AD 0010 09 10 08 01 07 05
RD 0005 07 10 10 02 04 03
Frontal NAWM
FA 0045 08 10 10 02 07 02
MD 0001 05 07 10 02 06 04
AD 0004 05 09 08 03 07 04
RD 0003 06 09 09 02 06 03
Parietal NAWM
FA 0023 09 09 09 02 05 01
MD 0004 07 07 08 01 05 02
AD 0017 08 10 10 01 06 03
RD 0006 08 08 09 02 05 02
Temporal NAWM
FA 0116 07 07 10 03 05 03
MD 0007 06 10 10 01 05 05
AD 0009 07 08 08 01 05 05
RD 0020 06 09 10 02 05 04
Occipital NAWM
FA 0011 09 06 09 03 03 01
MD 0009 08 08 09 01 06 02
AD 0086 08 07 08 01 08 03
RD 0011 08 08 08 02 05 02
Cingulate NAWM
FA 0004 05 04 08 06 04 03
MD 0001 06 09 08 06 02 07
AD 0003 08 09 08 03 04 07
RD 0001 06 07 08 07 02 06
Deep NAWM
FA 0047 05 06 09 03 05 03
MD 0001 07 09 09 05 02 07
AD 0005 10 08 10 03 04 08
RD 0002 06 08 10 05 02 08
Volume (pf)
NAWM 0019 08 08 08 03 05 05
Continued
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 7
eventually manifest by brain atrophy and clinically by steadyworsening of disability in almost all patients over time This diffusepathology which in the majority of cases leads to progressiveworsening of the MS-related symptoms is now the focus of in-terest for patients and physicians as modern disease-modifyingtherapies have led to an almost complete suppression of relapseactivity while their impact on progression is at best modest Ourincomplete understanding of the molecular mechanisms that leadto progression is one of the main reasons for the failure in de-velopment ofmore efficacious therapies for secondary progressiveform of MS Another impediment is the relative insensitivity ofcurrent measurement tools to capture features of progression atsubclinical stages and hence to quantitate the effects of anti-neurodegenerative treatments at a time point where the overallbrain structure is still largely intact Specifically conventional MRIcan only quantitate brain volume loss as an end result of thedisease but does not provide insights into the underlying micro-structural changes DTI is a nonconventional MRI technique thatprovides a measure of such diffuse changes beyond volumetry
NfL is a highly sensitive biomarker to detect neuronal damageand is the first of its kind applicable in blood-derived probes22
However the cause and underlying pathology of the increasedNfL concentration may vary greatly both inter- and intra-individually In clinically isolated syndrome and RRMS acuteNfL elevation is particularly associated with focal in-flammation ie relapses and focal lesion formation4ndash823
However the indication for the contribution of diffuse pa-thology to higher NfL levels has so far been only incidentaland consisted of associations of increased NfL with moreadvanced clinical disability and brain atrophy824 Novelfindings presented here provide more direct evidence for theeffect of diffuse pathologic process in the NAWM on in-creased serum NfL We observed that among a subgroup ofolder and more disabled patients NfL associates both withradial and axial diffusivities In the context of MS these DTIscalars are thought to reflect myelin loss and axonal damagerespectively11 NfL also associated with mean diffusivity thatcan also indicate on axonal andmyelin loss11 The associationswere most pronounced in the cingulate deep and frontalNAWM
NfL differs in 2 important aspects from MRI first as asignal of neurodegeneration it reflects ongoing pathologicprocesses in real time while imaging captures morphologicfeatures that are inherently retrospective Second NfLlevels reflect neuronal damage in the CNS comprehen-sively ie in addition to brain pathology it also capturesspinal cord pathology that is not routinely evaluated byimaging in the workup of MS for individual patients As adownside NfL levels cannot differentiate betweenacute inflammatory and chronic neurodegenerative diseaseactivity Our results suggest that DTI-MRI allows us tocategorize the pathogenic source of NfL at least semi-quantitatively We demonstrate that both diffuse axonaldamage within the NAWM measured using DTI and focalinflammatory white matter lesion load measured usingconventional MRI contributed to the elevated NfL serumlevels This highlights that both focal and diffuse pathologicchanges must be taken into account when levels of NfL inindividual patients are interpreted A somewhat largerproportion of NfL variance was explained by DTI indices inaverage than by T2 lesion load This implies that in non-relapsing patients the ongoing diffuse microstructuraldamage in the NAWM outside lesions might predominateover focal axonal damage that has occurred within lesionsas a source of NfL in serum Accordingly in other neuro-logic disease with brain diffuse pathology such as amyo-trophic lateral sclerosis Alzheimer disease traumatic braininjury and age-related white matter pathology elevatedserum NfL levels have been associated with DTI-measurable diffuse damage within the CNS25ndash28 Thesestudies are further evidence for the concept that braindiffuse chronic neuropathologic mechanisms contribute toNfL release and that nonconventional MRI techniques areable to identify the CNS areas of its morphologic source
Current results demonstrate that the cingulate area is thebrain area with the most prominent white matter tractdamage outside focal lesions Periventricular regions arecharacteristically the brain area with a particularly heavylesion load in MS29 We addressed the question whetherthe cingulum-NAWM DTI abnormalities would arise from
Table 3 Multiple regression analyses evaluating separately the effect of DTI-MRI indices or volumetric brain MRI data onNfL in the NfL(high) subgroup of patients with MS (continued)
Variable in model DTI Sex Age Type EDSS score DMT Relapses
GMctx 0004 09 02 04 003 03 06
T2 0019 06 03 07 04 07 03
T1 015 07 04 10 05 07 03
Abbreviations AD = axial diffusivity DMT = disease-modifying treatment DTI = diffusion tensor imaging EDSS = Expanded Disability Status Scale FA =fractional anisotropy GMctx = cortical gray matter MD = mean diffusivity NAWM = normal-appearing white matter NfL = neurofilament light chain pf =parenchymal fraction RD = radial diffusivityPatients withMSwere divided into NfL(low) and NfL(high) subgroups based on themedian NfL value of healthy controls (231 pgmL) Model was adjusted bysex age disease type (relapsing-remittingsecondary progressive) EDSS score treatment (no first- and second-line) and presence of relapses within 1 yearbefore sampling (yesno) Bolded p values indicate that the parameter affected the logarithmofNfL value statistically significantly whenDTI or volume and allthe adjusting variables were in the models
8 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
the dirty-appearing white matter due to close proximity to aheavy lesion load but the relative lesion volume in the cingulatearea was not found to be significantly larger compared withlesion load in other regions or in the whole NAWM Thereforeit seems unlikely that differences in the DTI indices would besolely due to dirty-appearing white matter adjacent to lesionswhereas it is more likely that remote lesions lead to DTI changesin the cingulateNAWM throughWallerian degeneration30 Alsoother studies have observed diffusion changes in the cingulatearea of patients with MS31ndash33 Because the cingulum bundle offibers projects from the cingulate gyrus to the entorhinal cortexin the brain allowing for communication between components
of the limbic system the DTI abnormalities may be the mor-phologic substrate for impaired cognitive function and fatiguesymptoms that are typical in progressive MS3435
Present results are in line with previous observations on therelation between NfL concentration and MS disease sub-type with progressive patients having greater NfL levelscompared with relapsing-remitting patients568ndash102324
However the median NfL value of the entire patient cohortwas relatively low compared with healthy controls which isthe main limitation of our study Because of this we werenot able to use the previously defined cutoff values3 for
Figure 2 The effect of DTI-MRI indices or volumetric MRI data on NfL in the NfL(high) subgroup
The NfL(high) subgroup is comprised of patients with serum NfL above the median value of healthy controls (231 pgmL) Logarithmic serum NfL wasmodeled separately by DTI indices of the entire and parcellated NAWM and volumetric brain MRI data using multiple regression analyses Models wereadjusted by sex age disease type (RRMSSPMS) the EDSS score treatment (no first and second line) and presence of relapseswithin 1 year before sampling(yesno) The results are illustrated using dot and whisker plots in which red dots represent standardized regression coefficients and red lines represent theCIs of the estimates Significant p values of the DTI parameters and the percentage of variance in the response that can be explained by the independentvariables (R2) are also shown The results were ordered according to the R2 value All except 2 of the p values (fractional anisotropy of the frontal and deepNAWM) shown in the figure remained significant after adjustment using the false discovery ratemethod for the number of investigated variables (n = 32) AD =axial diffusivity DTI = diffusion tensor imaging EDSS = Expanded Disability Status Scale FA = fractional anisotropy GMctx = cortical gray matter MD =meandiffusivity NAWM = normal-appearing white matter NfL = neurofilament light chain PF = parenchymal fraction RD = radial diffusivity RRMS = relapsing-remitting MS SPMS = secondary progressive MS WM = white matter
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 9
NfL(low) and NfL(high) subgroups Instead we used themedian value of healthy controls Although the divider usedin our study was based on a small group of individualssimilar values have been observed in previous largerstudies68 One reason for the low median NfL level couldbe that the patients with RRMS are represented with arelatively benign disease course In addition our patientswith RRMS were sampled during a relapse-free time andhad no signs of ongoing focal inflammatory activity in theMRI Moreover most of the patients (68) were using adisease-modifying treatment at the time of evaluationHence similar as observed in other studies35233637 thisresulted in low serum NfL concentration Following thesimilarity in the NfL levels between healthy controls andpatients with MS the main results of our study are based ona relatively low number of patients In addition the patientnumbers under different immunomodulatory treatmentsvaried substantially However the heterogeneity of thecohort was taken account in the multiple regression modelwhere both EDSS score and disease-modifying treatmentwere used as adjustments
Our findings suggest that elevated serum NfL in MS resultsfrom neuroaxonal damage both within the NAWM and infocal T2 lesions The association between DTI-measurablediffuse microstructural white matter damage and serumNfL isfurther conceptual evidence for the latter being a usefulmonitoring tool in assessing the degree of ongoing neuro-degenerative processes in MS
AcknowledgmentThe authors thank all patients with MS and healthycontrols who participated in this study They also thank thepersonnel of Turku PET Centre for their excellenttechnical assistance
Study fundingThis work was supported by the Finnish Academy the SigridJuselius Foundation
DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgNN for full disclosures
Publication historyReceived by Neurology Neuroimmunology amp NeuroinflammationMay 8 2020 Accepted in final form October 21 2020
References1 Reich DS Lucchinetti CF Calabresi PA Multiple sclerosis N Engl J Med 2018378
169ndash1802 Varhaug KN Torkildsen Oslash Myhr KM Vedeler CA Neurofilament light chain as a
biomarker in multiple sclerosis Front Neurol 201910338
Appendix Authors
Author Location Contribution
MaijaSaraste PhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Acquisition of dataanalyzed andinterpreted the data anddrafted and revised themanuscript forintellectual content
Appendix (continued)
Author Location Contribution
SvetlanaBezukladovaMSc
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Acquisition of dataanalyzed and interpretedthe data and revised themanuscript
MarkusMatilainenPhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Statistical analysisdrafted the figures andrevised the manuscript
Jouni TuiskuMSc
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Major role in theacquisition of data andrevised the manuscript
EeroRissanenMDPhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Designed andconceptualized the studyand acquisition of data
MarcusSucksdorffMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Major role in theacquisition of data
SiniLaaksonenMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Acquisition of data
AnnaVuorimaaMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Acquisition of data
Jens KuhleMD PhD
Neurologic Clinic andPoliclinic Departments ofMedicine Biomedicineand Clinical ResearchUniversity Hospital BaselSwitzerland
Major role in theacquisition of data
DavidLeppert MDPhD
Neurologic Clinic andPoliclinic Departments ofMedicine Biomedicineand Clinical ResearchUniversity Hospital BaselSwitzerland
Revised the manuscriptfor intellectual content
Laura AirasMD PhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Designed andconceptualized the studyand revised themanuscript for intellectualcontent
10 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
3 Kuhle J Kropshofer H Haering DA et al Blood neurofilament light chain as abiomarker of MS disease activity and treatment response Neurology 201992e1007ndashe1015
4 Varhaug KN Barro C Bjoslashrnevik K et al Neurofilament light chain predicts diseaseactivity in relapsing-remitting MS Neurol Neuroimmunol Neuroinflamm 20185e422
5 Novakova L Zetterberg H Sundstrom P et al Monitoring disease activity in multiplesclerosis using serum neurofilament light protein Neurology 2017892230ndash2237
6 Disanto G Barro C Benkert P et al Serum neurofilament light a biomarker ofneuronal damage in multiple sclerosis Ann Neurol 201781857ndash870
7 Siller N Kuhle J Muthuraman M et al Serum neurofilament light chain is a bio-marker of acute and chronic neuronal damage in early multiple sclerosis Mult Scler201925678ndash686
8 Barro C Benkert P Disanto G et al Serum neurofilament as a predictor of diseaseworsening and brain and spinal cord atrophy in multiple sclerosis Brain 20181412382ndash2391
9 Ferraro D Guicciardi C De Biasi S et al Plasma neurofilaments correlate withdisability in progressive multiple sclerosis patients Acta Neurol Scand 202014116ndash21
10 Jakimovski D Zivadinov R RamanthanM et al Serum neurofilament light chain levelassociations with clinical and cognitive performance in multiple sclerosis a longitu-dinal retrospective 5-year study Mult Scler 2020261670ndash1681
11 Sbardella E Tona F Petsas N Pantano P DTI measurements in multiple sclerosisevaluation of brain damage and clinical implications Mult Scler Int 20132013671730
12 Tae WS Ham BJ Pyun SB Kang SH Kim BJ Current clinical applications ofdiffusion-tensor imaging in neurological disorders J Clin Neurol 201814129ndash140
13 Kurtzke JF Rating neurologic impairment in multiple sclerosis an expanded disabilitystatus scale (EDSS) Neurology 1983331444ndash1452
14 Kuhle J Barro C Andreasson U et al Comparison of three analytical platforms forquantification of the neurofilament light chain in blood samples ELISA electro-chemiluminescence immunoassay and SimoaClinChemLabMed 2016541655ndash1661
15 Bezukladova S Tuisku J Matilainen M et al Insights into disseminated MS brainpathology with multimodal diffusion tensor and PET imaging Neurol NeuroimmunolNeuroinflamm 20207e691
16 Schmidt P Gaser C Arsic M et al An automated tool for detection of FLAIR-hyperintense white-matter lesions in Multiple Sclerosis Neuroimage 2012593774ndash3783
17 Salat DH Greve DN Pacheco JL et al Regional white matter volume differences innondemented aging and Alzheimerrsquos disease Neuroimage 2009441247ndash1258
18 Rissanen E Tuisku J Vahlberg T et al Microglial activation white matter tractdamage and disability in MS Neurol Neuroimmunol Neuroinflamm 20185e443
19 Leemans A Jeurissen B Sijbers J et al ExploreDTI a graphical toolbox for pro-cessing analysing and visualizing diffusion MR data Presented at the 17th AnnualMeeting of International Society for Magnetic Resonance in Medicine Honolulu HI2009 p 3537
20 Benjamini Y Hochberg Y Controlling the false discovery rate a practical and pow-erful approach to multiple testing J R Stat Soc Ser B 199557289ndash300
21 Gelman A Scaling regression inputs by dividing by two standard deviations Stat Med2008272865ndash2873
22 Khalil M Teunissen CE Otto M et al Neurofilaments as biomarkers in neurologicaldisorders Nat Rev Neurol 201814577ndash589
23 Canto E Barro C Zhao C et al Association between serum neurofilament light chainlevels and long-term disease course among patients with multiple sclerosis followedup for 12 years JAMA Neurol 2019761359ndash1366
24 Hogel H Rissanen E Barro C et al Serum glial fibrillary acidic protein correlates withmultiple sclerosis disease severity Mult Scler 202026210ndash219
25 Menke RA Gray E Lu CH et al CSF neurofilament light chain reflects corticospinaltract degeneration in ALS Ann Clin Transl Neurol 20152748ndash755
26 Moore EE Hohman TJ Badami FS et al Neurofilament relates to white mattermicrostructure in older adults Neurobiol Aging 201870233ndash241
27 Ljungqvist J Zetterberg H Mitsis M Blennow K Skoglund T Serum neurofilamentlight protein as a marker for diffuse axonal injury results from a case series studyJ Neurotrauma 2017341124ndash1127
28 Schultz SA Strain JF Adedokun A et al Serum neurofilament light chain levels areassociated with white matter integrity in autosomal dominant Alzheimerrsquos diseaseNeurobiol Dis 2020142104960
29 Narayanan S Fu L Pioro E et al Imaging of axonal damage in multiple sclerosisspatial distribution of magnetic resonance imaging lesions Ann Neurol 199741385ndash391
30 Ciccarelli O Werring DJ Barker GJ et al A study of the mechanisms of normal-appearing white matter damage in multiple sclerosis using diffusion tensor imagingevidence of Wallerian degeneration J Neurol 2003250287ndash292
31 Pokryszko-Dragan A Banaszek A Nowakowska-Kotas M et al Diffusion tensorimaging findings in the multiple sclerosis patients and their relationships to variousaspects of disability J Neurol Sci 2018391127ndash133
32 Kolasa M Hakulinen U Brander A et al Diffusion tensor imaging and disabilityprogression in multiple sclerosis a 4-year follow-up study Brain Behav 20199e01194
33 Andersen O Hildeman A Longfils M et al Diffusion tensor imaging in multiplesclerosis at different final outcomes Acta Neurol Scand 2018137165ndash173
34 Koenig KA Sakaie KE Lowe MJ et al The relationship between cognitive functionand high-resolution diffusion tensor MRI of the cingulum bundle in multiple sclerosisMult Scler 2015211794ndash1801
35 Pardini M Bonzano L Bergamino M et al Cingulum bundle alterations underliesubjective fatigue in multiple sclerosis Mult Scler 201521442ndash447
36 Sejbaek T Nielsen HH Penner N et al Dimethyl fumarate decreases neurofilamentlight chain in CSF and blood of treatment naıve relapsing MS patients J NeurolNeurosurg Psychiatry 2019901324ndash1330
37 Piehl F Kockum I Khademi M et al Plasma neurofilament light chain levels inpatients with MS switching from injectable therapies to fingolimod Mult Scler 2018241046ndash1054
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 11
DOI 101212NXI000000000000092620218 Neurol Neuroimmunol Neuroinflamm
Maija Saraste Svetlana Bezukladova Markus Matilainen et al High serum neurofilament associates with diffuse white matter damage in MS
This information is current as of December 8 2020
ServicesUpdated Information amp
httpnnneurologyorgcontent81e926fullhtmlincluding high resolution figures can be found at
References httpnnneurologyorgcontent81e926fullhtmlref-list-1
This article cites 36 articles 3 of which you can access for free at
Citations httpnnneurologyorgcontent81e926fullhtmlotherarticles
This article has been cited by 1 HighWire-hosted articles
Subspecialty Collections
httpnnneurologyorgcgicollectionmultiple_sclerosisMultiple sclerosis
httpnnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpnnneurologyorgmiscaboutxhtmlpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnnneurologyorgmiscaddirxhtmlreprintsusInformation about ordering reprints can be found online
Academy of Neurology All rights reserved Online ISSN 2332-7812Copyright copy 2020 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the AmericanPublished since April 2014 it is an open-access online-only continuous publication journal Copyright
is an official journal of the American Academy of NeurologyNeurol Neuroimmunol Neuroinflamm
groups and between patients and healthy controls wereassessed using theWilcoxon rank-sum (Mann-WhitneyU) testHolm multiple comparison adjustment was used in RRMS vssecondary progressive MS (SPMS) vs healthy control multiplecomparisons The division into NfL(low) and NfL(high)subgroups was based on the median NfL value of the healthycontrols
To test our main hypothesis Spearman correlations were calcu-lated to assess the relationships between NfL and DTI indices(fractional anisotropy mean diffusivity radial diffusivity and axialdiffusivity) in the whole NAWM and in NAWM subregionsSpearman correlation was used instead of Pearson correlation toavoid outlier-induced overestimation of correlations The p valuesof correlation analyses were further adjusted with the false dis-covery rate method (Benjamini-Hochberg procedure) for thenumber of investigated parameters (n = 28)20 The NfL valueswere further modeled by the DTI indices and volumetric MRIdata using multiple regression models The logarithm of NfL wasused as the response because nontransformed values led to non-normality of residuals The models were adjusted by sex agedisease type (relapsing-remittingsecondary progressive) theEDSS score disease-modifying treatment (no treatment first line[dimethyl fumarate glatiramer acetate interferon-beta and teri-flunomide] or second line [fingolimod natalizumab and ritux-imab] treatment) and the presence of relapses within 1 yearbefore sampling (yesno) The normality of the residuals waschecked using the Shapiro-Wilk test Variance inflation factorvalues were used to check that independent variables were nothighly correlated with each other Regression coefficients were
standardized tomake themdirectly comparable to each other Forstandardization the regression results weremultiplied by twice theSD of the DTI or volume variables21 The p values of the multipleregression model were adjusted with the false discovery ratemethod for the number of investigated variables (n = 32)
In addition to verify the quality of the data and increase thegeneralizability and comparability of the results we usedSpearman correlation to assess the relationships betweenfollowing parameters NfL with age disease duration theEDSS score parenchymal fractions of NAWM cortical graymatter and T1 and T2 lesion volumes DTI with the EDSSscore and parenchymal fractions of NAWM cortical graymatter and T1 and T2 lesion volumes (results are shown inappendix e-3 linkslwwcomNXIA355) Spearman cor-relation was used instead of Pearson correlation to avoidoutlier-induced overestimation of correlations and becauseSpearman correlation is preferred in case of ordinal vari-ables ie EDSS score All statistical tests were 2 tailed andp = 005 was used as the threshold for statistical significance
Data availabilityThe anonymized raw data will be shared over the next 3 yearson request from a qualified investigator
ResultsThe demographics clinical characteristics and conventionalimaging data of the 79 patients with MS included in the study
Figure 1 Correlation of serum NfL and DTI indices of the whole NAWM in the NfL(high) subgroup
The serum level of NfL is moderately associatedwith fractional anisotropy (A) mean diffusivity(B) axial diffusivity (C) and radial diffusivity (D)The NfL(high) subgroup is comprised of patientswith serum NfL above the median value ofhealthy controls (231 pgmL) Shown areSpearman correlation coefficients (ρ) and p val-ues The black lines indicate the level and di-rection of the relationship AD = axial diffusivityDTI = diffusion tensor imaging FA = fractionalanisotropy MD = mean diffusivity NAWM =normal-appearing white matter NfL = neuro-filament light chain RD = radial diffusivity
4 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
Table 2 Spearman correlations between serum NfL concentrations and DTI-MRI indices
All MS NfL(low) NfL(high)
ρ p Value ρ p Value ρ p Value
Entire NAWM
FA minus016 02 minus020 02 minus042 0013
MD 012 03 014 04 047 0005
AD 001 09 005 07 041 0017
RD 013 03 015 03 045 0008
Frontal NAWM
FA minus018 01 minus020 02 minus037 0033
MD 017 01 011 05 047 0005
AD 009 04 004 08 044 0010
RD 015 02 013 04 044 0009
Parietal NAWM
FA minus009 04 minus011 05 minus034 0046
MD 005 07 005 08 043 0011
AD minus004 08 minus003 08 034 0053
RD 007 05 008 06 044 0009
Temporal NAWM
FA minus019 01 minus012 04 minus033 0060
MD 014 02 008 06 043 0012
AD minus003 08 004 08 026 0139
RD 014 02 009 07 038 0026
Occipital NAWM
FA minus009 04 minus018 02 minus043 0011
MD 012 03 017 03 053 0001
AD 004 07 014 04 038 0029
RD 010 04 015 03 048 0004
Cingulate NAWM
FA minus010 04 minus006 07 minus048 0004
MD 008 05 002 09 049 0003
AD 009 04 010 05 046 0007
RD 008 05 004 08 050 0003
Deep NAWM
FA minus014 02 minus020 03 minus036 0036
MD 008 05 010 06 033 0053
AD 003 08 001 09 036 0039
RD 01 04 012 05 036 0036
Abbreviations AD = axial diffusivity DTI = diffusion tensor imaging FA = fractional anisotropy MD = mean diffusivity NAWM = normal-appearing whitematter NfL = neurofilament light chain RD = radial diffusivityPatients with MS were divided into NfL(low) and NfL(high) subgroups based on the median value of healthy controls (231 pgmL) Fractional anisotropy andmean axial and radial diffusivities of the entire NAWMand of 6 parcellated subregions of NAWMwere correlated with serumNfL level in all patients with MSand in NfL(low) and NfL(high) subgroups Shown are Spearman correlation coefficients (ρ) and uncorrected p values Significant p values are boldedSignificance of these correlations was sustained after adjustment using the false discovery ratemethod for the number of DTI parameters (n = 28) except forparietal FA
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 5
are shown in table 1 The median (interquartile range [IQR])age of this cohort was 483 (43ndash53) years which is similar tothe age of the healthy individuals in the control group (476[44ndash53] p = 1) Overall the patient cohort was representedwith a quite stable disease as only 13 of the patients (9 RRMSand 1 SPMS) had had a relapse within 1 year before samplingSixty-eight percent of the patients (45 RRMS and 9 SPMS)were on disease-modifying therapy (dimethyl fumarate n = 3fingolimod n = 14 glatiramer acetate n = 5 interferon beta-1a n= 6 natalizumab n = 8 rituximab n = 4 and teriflunomide n =14) In MRI white matter and cortical gray matter volumeswere decreased in patients compared with healthy controlsDTI measures were obtained from both the entire cerebralNAWM and from 6 different NAWM subregions In the cin-gulate area there were significant differences in all 4 DTI pa-rameters in patients with MS compared with healthy controlsfractional anisotropy was decreased whereas mean diffusivityradial diffusivity and axial diffusivity were increased (table 1)
Serum NfL levels in patients with MS andhealthy controlsThe median (IQR) NfL level was higher in patients withSPMS compared with patients with RRMS (314 [22ndash44] vs188 [14ndash26] pgmL p = 0001 table 1) NfL levels in thewhole MS group were not different from healthy controls(217 [15ndash31] vs 231 [21ndash28] pgmL p = 03) regardlesswhether the analysis was performed with age correctionAcute inflammation did not appear to be a factor to affect theNfL level in this cohort with relatively modest acute in-flammatory activity as the median NfL value among thepatients with or without relapse within the previous year wascomparable (223 [16ndash32] vs 186 [15ndash25] pgmL p = 03Wilcoxon rank-sum data not shown)
Characterization of patients with highNfL levelsTo explore associations between increased NfL levels and DTI-measurable diffuse neuroaxonal damage patients with MS weredivided into NfL(high) and NfL(low) subgroups The divisionwas based on the median NfL value measured among healthycontrols (231 [21ndash28] pgmL) In the NfL(high) subgroup theNfL concentration was significantly elevated compared withhealthy controls (p = 0018 table 1) and in the further NfL vsDTI correlation analyses we focused on the NfL(high) subgroupBased on the demographic and clinical data the patients in theNfL(high) subgroupwere at amore advanced stage of the diseaseTheNfL(high) subgroup includedmore patients with SPMS (p =0003) and the patients in the subgroup were also older (p =0003) and had a higher EDSS score (p = 0005 table 1) Inevaluation using conventional MRI the NAWM white matter orcortical gray matter volumes and T1 and T2 lesion loads were notdifferent between NfL(low) and NfL(high) subgroups (table 1)
Associations between NfL and DTIOur results show that DTI metrics of diffuse neuroaxonaldamage within theNAWMassociatewith high serumNfL levelsin MS As a demonstration of this in the NfL(high) subgroup
several (n = 24) significant correlations were found betweenserum NfL levels and NAWM DTI indices both in the entireNAWM and in various brain subregions higher NfL levels wereassociated with lower fractional anisotropy and higher diffusivity(mean axial and radial) in the whole NAWM (figure 1) and inall its subregions except for fractional anisotropy in temporalNAWM axial diffusivity in parietal and temporal NAWM andmean diffusivity in remaining NAWM (table 2) Significance ofthese correlations was sustained after adjustment using the falsediscovery rate method for the number of DTI parameters (n =28) except for parietal fractional anisotropy Results remainedsimilar when the data were analyzed without patients who hadhad a relapse within the previous year before sampling (n = 3data not shown) No correlations between serumNfL levels andDTI metrics were observed in the NfL(low) subgroup or in theoverall MS cohort (table 2)
Multiple regression modeling was performed to further evaluatethe effect of DTI on NfL In the NfL(high) subgroup the DTIindices weremore significantly associated withNfL thanwith theclinical parameters included in the model (sex age disease typeEDSS score treatment and the presence of relapses within 1year before sampling) (table 3) On average 48 of the variancein NfL (SD 44) could be explained by DTI indices and theclinical parameters mentioned above (figure 2) There weremultiple (n = 26) significant associations between NfL and DTIin the whole NAWM and in all its 6 subregions All except 2 ofthe associations remained significant after adjustment with thefalse discovery rate method for the number of investigated var-iables (n = 32) The strongest associations were observed in thecingulate area between higher NfL and higher mean and radialdiffusivity and in deep and frontal NAWM between higher NfLand higher mean diffusivity (figure 2) Analyzing the datawithout the patients with a relapse (n = 3) within the previousyear did not alter the associations (data not shown)
Higher NfL was also associated with lower NAWM andcortical gray matter and higher T2 lesion volumes (figure 2)The volumes and clinical adjustments explained 434494 and 433 of the variance in NfL respectively
DiscussionPresent results demonstrate that the DTI-MRI measures ofNAWM correlate with serumNfL in MS We found that amongpatients with more advanced disease increased serumNfL levelsassociate with DTI measures reflecting diffuse microstructuraldamage ie decreased fractional anisotropy and increasedmeanaxial and radial diffusivities in the NAWM In a multiple re-gression model which was adjusted with sex age disease typeEDSS score treatment and the presence of relapses the DTIindices were more significantly associated with NfL than theabove-mentioned demographic and clinical parameters
The pathophysiology of MS involves acute and chronic mech-anisms that lead to gradual axonal and myelin damage that
6 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
Table 3 Multiple regression analyses evaluating separately the effect of DTI-MRI indices or volumetric brain MRI data onNfL in the NfL(high) subgroup of patients with MS
Variable in model DTI Sex Age Type EDSS score DMT Relapses
Entire NAWM
FA 0018 07 08 10 03 04 02
MD 0003 07 09 10 02 04 04
AD 0010 09 10 08 01 07 05
RD 0005 07 10 10 02 04 03
Frontal NAWM
FA 0045 08 10 10 02 07 02
MD 0001 05 07 10 02 06 04
AD 0004 05 09 08 03 07 04
RD 0003 06 09 09 02 06 03
Parietal NAWM
FA 0023 09 09 09 02 05 01
MD 0004 07 07 08 01 05 02
AD 0017 08 10 10 01 06 03
RD 0006 08 08 09 02 05 02
Temporal NAWM
FA 0116 07 07 10 03 05 03
MD 0007 06 10 10 01 05 05
AD 0009 07 08 08 01 05 05
RD 0020 06 09 10 02 05 04
Occipital NAWM
FA 0011 09 06 09 03 03 01
MD 0009 08 08 09 01 06 02
AD 0086 08 07 08 01 08 03
RD 0011 08 08 08 02 05 02
Cingulate NAWM
FA 0004 05 04 08 06 04 03
MD 0001 06 09 08 06 02 07
AD 0003 08 09 08 03 04 07
RD 0001 06 07 08 07 02 06
Deep NAWM
FA 0047 05 06 09 03 05 03
MD 0001 07 09 09 05 02 07
AD 0005 10 08 10 03 04 08
RD 0002 06 08 10 05 02 08
Volume (pf)
NAWM 0019 08 08 08 03 05 05
Continued
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 7
eventually manifest by brain atrophy and clinically by steadyworsening of disability in almost all patients over time This diffusepathology which in the majority of cases leads to progressiveworsening of the MS-related symptoms is now the focus of in-terest for patients and physicians as modern disease-modifyingtherapies have led to an almost complete suppression of relapseactivity while their impact on progression is at best modest Ourincomplete understanding of the molecular mechanisms that leadto progression is one of the main reasons for the failure in de-velopment ofmore efficacious therapies for secondary progressiveform of MS Another impediment is the relative insensitivity ofcurrent measurement tools to capture features of progression atsubclinical stages and hence to quantitate the effects of anti-neurodegenerative treatments at a time point where the overallbrain structure is still largely intact Specifically conventional MRIcan only quantitate brain volume loss as an end result of thedisease but does not provide insights into the underlying micro-structural changes DTI is a nonconventional MRI technique thatprovides a measure of such diffuse changes beyond volumetry
NfL is a highly sensitive biomarker to detect neuronal damageand is the first of its kind applicable in blood-derived probes22
However the cause and underlying pathology of the increasedNfL concentration may vary greatly both inter- and intra-individually In clinically isolated syndrome and RRMS acuteNfL elevation is particularly associated with focal in-flammation ie relapses and focal lesion formation4ndash823
However the indication for the contribution of diffuse pa-thology to higher NfL levels has so far been only incidentaland consisted of associations of increased NfL with moreadvanced clinical disability and brain atrophy824 Novelfindings presented here provide more direct evidence for theeffect of diffuse pathologic process in the NAWM on in-creased serum NfL We observed that among a subgroup ofolder and more disabled patients NfL associates both withradial and axial diffusivities In the context of MS these DTIscalars are thought to reflect myelin loss and axonal damagerespectively11 NfL also associated with mean diffusivity thatcan also indicate on axonal andmyelin loss11 The associationswere most pronounced in the cingulate deep and frontalNAWM
NfL differs in 2 important aspects from MRI first as asignal of neurodegeneration it reflects ongoing pathologicprocesses in real time while imaging captures morphologicfeatures that are inherently retrospective Second NfLlevels reflect neuronal damage in the CNS comprehen-sively ie in addition to brain pathology it also capturesspinal cord pathology that is not routinely evaluated byimaging in the workup of MS for individual patients As adownside NfL levels cannot differentiate betweenacute inflammatory and chronic neurodegenerative diseaseactivity Our results suggest that DTI-MRI allows us tocategorize the pathogenic source of NfL at least semi-quantitatively We demonstrate that both diffuse axonaldamage within the NAWM measured using DTI and focalinflammatory white matter lesion load measured usingconventional MRI contributed to the elevated NfL serumlevels This highlights that both focal and diffuse pathologicchanges must be taken into account when levels of NfL inindividual patients are interpreted A somewhat largerproportion of NfL variance was explained by DTI indices inaverage than by T2 lesion load This implies that in non-relapsing patients the ongoing diffuse microstructuraldamage in the NAWM outside lesions might predominateover focal axonal damage that has occurred within lesionsas a source of NfL in serum Accordingly in other neuro-logic disease with brain diffuse pathology such as amyo-trophic lateral sclerosis Alzheimer disease traumatic braininjury and age-related white matter pathology elevatedserum NfL levels have been associated with DTI-measurable diffuse damage within the CNS25ndash28 Thesestudies are further evidence for the concept that braindiffuse chronic neuropathologic mechanisms contribute toNfL release and that nonconventional MRI techniques areable to identify the CNS areas of its morphologic source
Current results demonstrate that the cingulate area is thebrain area with the most prominent white matter tractdamage outside focal lesions Periventricular regions arecharacteristically the brain area with a particularly heavylesion load in MS29 We addressed the question whetherthe cingulum-NAWM DTI abnormalities would arise from
Table 3 Multiple regression analyses evaluating separately the effect of DTI-MRI indices or volumetric brain MRI data onNfL in the NfL(high) subgroup of patients with MS (continued)
Variable in model DTI Sex Age Type EDSS score DMT Relapses
GMctx 0004 09 02 04 003 03 06
T2 0019 06 03 07 04 07 03
T1 015 07 04 10 05 07 03
Abbreviations AD = axial diffusivity DMT = disease-modifying treatment DTI = diffusion tensor imaging EDSS = Expanded Disability Status Scale FA =fractional anisotropy GMctx = cortical gray matter MD = mean diffusivity NAWM = normal-appearing white matter NfL = neurofilament light chain pf =parenchymal fraction RD = radial diffusivityPatients withMSwere divided into NfL(low) and NfL(high) subgroups based on themedian NfL value of healthy controls (231 pgmL) Model was adjusted bysex age disease type (relapsing-remittingsecondary progressive) EDSS score treatment (no first- and second-line) and presence of relapses within 1 yearbefore sampling (yesno) Bolded p values indicate that the parameter affected the logarithmofNfL value statistically significantly whenDTI or volume and allthe adjusting variables were in the models
8 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
the dirty-appearing white matter due to close proximity to aheavy lesion load but the relative lesion volume in the cingulatearea was not found to be significantly larger compared withlesion load in other regions or in the whole NAWM Thereforeit seems unlikely that differences in the DTI indices would besolely due to dirty-appearing white matter adjacent to lesionswhereas it is more likely that remote lesions lead to DTI changesin the cingulateNAWM throughWallerian degeneration30 Alsoother studies have observed diffusion changes in the cingulatearea of patients with MS31ndash33 Because the cingulum bundle offibers projects from the cingulate gyrus to the entorhinal cortexin the brain allowing for communication between components
of the limbic system the DTI abnormalities may be the mor-phologic substrate for impaired cognitive function and fatiguesymptoms that are typical in progressive MS3435
Present results are in line with previous observations on therelation between NfL concentration and MS disease sub-type with progressive patients having greater NfL levelscompared with relapsing-remitting patients568ndash102324
However the median NfL value of the entire patient cohortwas relatively low compared with healthy controls which isthe main limitation of our study Because of this we werenot able to use the previously defined cutoff values3 for
Figure 2 The effect of DTI-MRI indices or volumetric MRI data on NfL in the NfL(high) subgroup
The NfL(high) subgroup is comprised of patients with serum NfL above the median value of healthy controls (231 pgmL) Logarithmic serum NfL wasmodeled separately by DTI indices of the entire and parcellated NAWM and volumetric brain MRI data using multiple regression analyses Models wereadjusted by sex age disease type (RRMSSPMS) the EDSS score treatment (no first and second line) and presence of relapseswithin 1 year before sampling(yesno) The results are illustrated using dot and whisker plots in which red dots represent standardized regression coefficients and red lines represent theCIs of the estimates Significant p values of the DTI parameters and the percentage of variance in the response that can be explained by the independentvariables (R2) are also shown The results were ordered according to the R2 value All except 2 of the p values (fractional anisotropy of the frontal and deepNAWM) shown in the figure remained significant after adjustment using the false discovery ratemethod for the number of investigated variables (n = 32) AD =axial diffusivity DTI = diffusion tensor imaging EDSS = Expanded Disability Status Scale FA = fractional anisotropy GMctx = cortical gray matter MD =meandiffusivity NAWM = normal-appearing white matter NfL = neurofilament light chain PF = parenchymal fraction RD = radial diffusivity RRMS = relapsing-remitting MS SPMS = secondary progressive MS WM = white matter
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 9
NfL(low) and NfL(high) subgroups Instead we used themedian value of healthy controls Although the divider usedin our study was based on a small group of individualssimilar values have been observed in previous largerstudies68 One reason for the low median NfL level couldbe that the patients with RRMS are represented with arelatively benign disease course In addition our patientswith RRMS were sampled during a relapse-free time andhad no signs of ongoing focal inflammatory activity in theMRI Moreover most of the patients (68) were using adisease-modifying treatment at the time of evaluationHence similar as observed in other studies35233637 thisresulted in low serum NfL concentration Following thesimilarity in the NfL levels between healthy controls andpatients with MS the main results of our study are based ona relatively low number of patients In addition the patientnumbers under different immunomodulatory treatmentsvaried substantially However the heterogeneity of thecohort was taken account in the multiple regression modelwhere both EDSS score and disease-modifying treatmentwere used as adjustments
Our findings suggest that elevated serum NfL in MS resultsfrom neuroaxonal damage both within the NAWM and infocal T2 lesions The association between DTI-measurablediffuse microstructural white matter damage and serumNfL isfurther conceptual evidence for the latter being a usefulmonitoring tool in assessing the degree of ongoing neuro-degenerative processes in MS
AcknowledgmentThe authors thank all patients with MS and healthycontrols who participated in this study They also thank thepersonnel of Turku PET Centre for their excellenttechnical assistance
Study fundingThis work was supported by the Finnish Academy the SigridJuselius Foundation
DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgNN for full disclosures
Publication historyReceived by Neurology Neuroimmunology amp NeuroinflammationMay 8 2020 Accepted in final form October 21 2020
References1 Reich DS Lucchinetti CF Calabresi PA Multiple sclerosis N Engl J Med 2018378
169ndash1802 Varhaug KN Torkildsen Oslash Myhr KM Vedeler CA Neurofilament light chain as a
biomarker in multiple sclerosis Front Neurol 201910338
Appendix Authors
Author Location Contribution
MaijaSaraste PhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Acquisition of dataanalyzed andinterpreted the data anddrafted and revised themanuscript forintellectual content
Appendix (continued)
Author Location Contribution
SvetlanaBezukladovaMSc
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Acquisition of dataanalyzed and interpretedthe data and revised themanuscript
MarkusMatilainenPhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Statistical analysisdrafted the figures andrevised the manuscript
Jouni TuiskuMSc
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Major role in theacquisition of data andrevised the manuscript
EeroRissanenMDPhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Designed andconceptualized the studyand acquisition of data
MarcusSucksdorffMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Major role in theacquisition of data
SiniLaaksonenMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Acquisition of data
AnnaVuorimaaMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Acquisition of data
Jens KuhleMD PhD
Neurologic Clinic andPoliclinic Departments ofMedicine Biomedicineand Clinical ResearchUniversity Hospital BaselSwitzerland
Major role in theacquisition of data
DavidLeppert MDPhD
Neurologic Clinic andPoliclinic Departments ofMedicine Biomedicineand Clinical ResearchUniversity Hospital BaselSwitzerland
Revised the manuscriptfor intellectual content
Laura AirasMD PhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Designed andconceptualized the studyand revised themanuscript for intellectualcontent
10 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
3 Kuhle J Kropshofer H Haering DA et al Blood neurofilament light chain as abiomarker of MS disease activity and treatment response Neurology 201992e1007ndashe1015
4 Varhaug KN Barro C Bjoslashrnevik K et al Neurofilament light chain predicts diseaseactivity in relapsing-remitting MS Neurol Neuroimmunol Neuroinflamm 20185e422
5 Novakova L Zetterberg H Sundstrom P et al Monitoring disease activity in multiplesclerosis using serum neurofilament light protein Neurology 2017892230ndash2237
6 Disanto G Barro C Benkert P et al Serum neurofilament light a biomarker ofneuronal damage in multiple sclerosis Ann Neurol 201781857ndash870
7 Siller N Kuhle J Muthuraman M et al Serum neurofilament light chain is a bio-marker of acute and chronic neuronal damage in early multiple sclerosis Mult Scler201925678ndash686
8 Barro C Benkert P Disanto G et al Serum neurofilament as a predictor of diseaseworsening and brain and spinal cord atrophy in multiple sclerosis Brain 20181412382ndash2391
9 Ferraro D Guicciardi C De Biasi S et al Plasma neurofilaments correlate withdisability in progressive multiple sclerosis patients Acta Neurol Scand 202014116ndash21
10 Jakimovski D Zivadinov R RamanthanM et al Serum neurofilament light chain levelassociations with clinical and cognitive performance in multiple sclerosis a longitu-dinal retrospective 5-year study Mult Scler 2020261670ndash1681
11 Sbardella E Tona F Petsas N Pantano P DTI measurements in multiple sclerosisevaluation of brain damage and clinical implications Mult Scler Int 20132013671730
12 Tae WS Ham BJ Pyun SB Kang SH Kim BJ Current clinical applications ofdiffusion-tensor imaging in neurological disorders J Clin Neurol 201814129ndash140
13 Kurtzke JF Rating neurologic impairment in multiple sclerosis an expanded disabilitystatus scale (EDSS) Neurology 1983331444ndash1452
14 Kuhle J Barro C Andreasson U et al Comparison of three analytical platforms forquantification of the neurofilament light chain in blood samples ELISA electro-chemiluminescence immunoassay and SimoaClinChemLabMed 2016541655ndash1661
15 Bezukladova S Tuisku J Matilainen M et al Insights into disseminated MS brainpathology with multimodal diffusion tensor and PET imaging Neurol NeuroimmunolNeuroinflamm 20207e691
16 Schmidt P Gaser C Arsic M et al An automated tool for detection of FLAIR-hyperintense white-matter lesions in Multiple Sclerosis Neuroimage 2012593774ndash3783
17 Salat DH Greve DN Pacheco JL et al Regional white matter volume differences innondemented aging and Alzheimerrsquos disease Neuroimage 2009441247ndash1258
18 Rissanen E Tuisku J Vahlberg T et al Microglial activation white matter tractdamage and disability in MS Neurol Neuroimmunol Neuroinflamm 20185e443
19 Leemans A Jeurissen B Sijbers J et al ExploreDTI a graphical toolbox for pro-cessing analysing and visualizing diffusion MR data Presented at the 17th AnnualMeeting of International Society for Magnetic Resonance in Medicine Honolulu HI2009 p 3537
20 Benjamini Y Hochberg Y Controlling the false discovery rate a practical and pow-erful approach to multiple testing J R Stat Soc Ser B 199557289ndash300
21 Gelman A Scaling regression inputs by dividing by two standard deviations Stat Med2008272865ndash2873
22 Khalil M Teunissen CE Otto M et al Neurofilaments as biomarkers in neurologicaldisorders Nat Rev Neurol 201814577ndash589
23 Canto E Barro C Zhao C et al Association between serum neurofilament light chainlevels and long-term disease course among patients with multiple sclerosis followedup for 12 years JAMA Neurol 2019761359ndash1366
24 Hogel H Rissanen E Barro C et al Serum glial fibrillary acidic protein correlates withmultiple sclerosis disease severity Mult Scler 202026210ndash219
25 Menke RA Gray E Lu CH et al CSF neurofilament light chain reflects corticospinaltract degeneration in ALS Ann Clin Transl Neurol 20152748ndash755
26 Moore EE Hohman TJ Badami FS et al Neurofilament relates to white mattermicrostructure in older adults Neurobiol Aging 201870233ndash241
27 Ljungqvist J Zetterberg H Mitsis M Blennow K Skoglund T Serum neurofilamentlight protein as a marker for diffuse axonal injury results from a case series studyJ Neurotrauma 2017341124ndash1127
28 Schultz SA Strain JF Adedokun A et al Serum neurofilament light chain levels areassociated with white matter integrity in autosomal dominant Alzheimerrsquos diseaseNeurobiol Dis 2020142104960
29 Narayanan S Fu L Pioro E et al Imaging of axonal damage in multiple sclerosisspatial distribution of magnetic resonance imaging lesions Ann Neurol 199741385ndash391
30 Ciccarelli O Werring DJ Barker GJ et al A study of the mechanisms of normal-appearing white matter damage in multiple sclerosis using diffusion tensor imagingevidence of Wallerian degeneration J Neurol 2003250287ndash292
31 Pokryszko-Dragan A Banaszek A Nowakowska-Kotas M et al Diffusion tensorimaging findings in the multiple sclerosis patients and their relationships to variousaspects of disability J Neurol Sci 2018391127ndash133
32 Kolasa M Hakulinen U Brander A et al Diffusion tensor imaging and disabilityprogression in multiple sclerosis a 4-year follow-up study Brain Behav 20199e01194
33 Andersen O Hildeman A Longfils M et al Diffusion tensor imaging in multiplesclerosis at different final outcomes Acta Neurol Scand 2018137165ndash173
34 Koenig KA Sakaie KE Lowe MJ et al The relationship between cognitive functionand high-resolution diffusion tensor MRI of the cingulum bundle in multiple sclerosisMult Scler 2015211794ndash1801
35 Pardini M Bonzano L Bergamino M et al Cingulum bundle alterations underliesubjective fatigue in multiple sclerosis Mult Scler 201521442ndash447
36 Sejbaek T Nielsen HH Penner N et al Dimethyl fumarate decreases neurofilamentlight chain in CSF and blood of treatment naıve relapsing MS patients J NeurolNeurosurg Psychiatry 2019901324ndash1330
37 Piehl F Kockum I Khademi M et al Plasma neurofilament light chain levels inpatients with MS switching from injectable therapies to fingolimod Mult Scler 2018241046ndash1054
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 11
DOI 101212NXI000000000000092620218 Neurol Neuroimmunol Neuroinflamm
Maija Saraste Svetlana Bezukladova Markus Matilainen et al High serum neurofilament associates with diffuse white matter damage in MS
This information is current as of December 8 2020
ServicesUpdated Information amp
httpnnneurologyorgcontent81e926fullhtmlincluding high resolution figures can be found at
References httpnnneurologyorgcontent81e926fullhtmlref-list-1
This article cites 36 articles 3 of which you can access for free at
Citations httpnnneurologyorgcontent81e926fullhtmlotherarticles
This article has been cited by 1 HighWire-hosted articles
Subspecialty Collections
httpnnneurologyorgcgicollectionmultiple_sclerosisMultiple sclerosis
httpnnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpnnneurologyorgmiscaboutxhtmlpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnnneurologyorgmiscaddirxhtmlreprintsusInformation about ordering reprints can be found online
Academy of Neurology All rights reserved Online ISSN 2332-7812Copyright copy 2020 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the AmericanPublished since April 2014 it is an open-access online-only continuous publication journal Copyright
is an official journal of the American Academy of NeurologyNeurol Neuroimmunol Neuroinflamm
Table 2 Spearman correlations between serum NfL concentrations and DTI-MRI indices
All MS NfL(low) NfL(high)
ρ p Value ρ p Value ρ p Value
Entire NAWM
FA minus016 02 minus020 02 minus042 0013
MD 012 03 014 04 047 0005
AD 001 09 005 07 041 0017
RD 013 03 015 03 045 0008
Frontal NAWM
FA minus018 01 minus020 02 minus037 0033
MD 017 01 011 05 047 0005
AD 009 04 004 08 044 0010
RD 015 02 013 04 044 0009
Parietal NAWM
FA minus009 04 minus011 05 minus034 0046
MD 005 07 005 08 043 0011
AD minus004 08 minus003 08 034 0053
RD 007 05 008 06 044 0009
Temporal NAWM
FA minus019 01 minus012 04 minus033 0060
MD 014 02 008 06 043 0012
AD minus003 08 004 08 026 0139
RD 014 02 009 07 038 0026
Occipital NAWM
FA minus009 04 minus018 02 minus043 0011
MD 012 03 017 03 053 0001
AD 004 07 014 04 038 0029
RD 010 04 015 03 048 0004
Cingulate NAWM
FA minus010 04 minus006 07 minus048 0004
MD 008 05 002 09 049 0003
AD 009 04 010 05 046 0007
RD 008 05 004 08 050 0003
Deep NAWM
FA minus014 02 minus020 03 minus036 0036
MD 008 05 010 06 033 0053
AD 003 08 001 09 036 0039
RD 01 04 012 05 036 0036
Abbreviations AD = axial diffusivity DTI = diffusion tensor imaging FA = fractional anisotropy MD = mean diffusivity NAWM = normal-appearing whitematter NfL = neurofilament light chain RD = radial diffusivityPatients with MS were divided into NfL(low) and NfL(high) subgroups based on the median value of healthy controls (231 pgmL) Fractional anisotropy andmean axial and radial diffusivities of the entire NAWMand of 6 parcellated subregions of NAWMwere correlated with serumNfL level in all patients with MSand in NfL(low) and NfL(high) subgroups Shown are Spearman correlation coefficients (ρ) and uncorrected p values Significant p values are boldedSignificance of these correlations was sustained after adjustment using the false discovery ratemethod for the number of DTI parameters (n = 28) except forparietal FA
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 5
are shown in table 1 The median (interquartile range [IQR])age of this cohort was 483 (43ndash53) years which is similar tothe age of the healthy individuals in the control group (476[44ndash53] p = 1) Overall the patient cohort was representedwith a quite stable disease as only 13 of the patients (9 RRMSand 1 SPMS) had had a relapse within 1 year before samplingSixty-eight percent of the patients (45 RRMS and 9 SPMS)were on disease-modifying therapy (dimethyl fumarate n = 3fingolimod n = 14 glatiramer acetate n = 5 interferon beta-1a n= 6 natalizumab n = 8 rituximab n = 4 and teriflunomide n =14) In MRI white matter and cortical gray matter volumeswere decreased in patients compared with healthy controlsDTI measures were obtained from both the entire cerebralNAWM and from 6 different NAWM subregions In the cin-gulate area there were significant differences in all 4 DTI pa-rameters in patients with MS compared with healthy controlsfractional anisotropy was decreased whereas mean diffusivityradial diffusivity and axial diffusivity were increased (table 1)
Serum NfL levels in patients with MS andhealthy controlsThe median (IQR) NfL level was higher in patients withSPMS compared with patients with RRMS (314 [22ndash44] vs188 [14ndash26] pgmL p = 0001 table 1) NfL levels in thewhole MS group were not different from healthy controls(217 [15ndash31] vs 231 [21ndash28] pgmL p = 03) regardlesswhether the analysis was performed with age correctionAcute inflammation did not appear to be a factor to affect theNfL level in this cohort with relatively modest acute in-flammatory activity as the median NfL value among thepatients with or without relapse within the previous year wascomparable (223 [16ndash32] vs 186 [15ndash25] pgmL p = 03Wilcoxon rank-sum data not shown)
Characterization of patients with highNfL levelsTo explore associations between increased NfL levels and DTI-measurable diffuse neuroaxonal damage patients with MS weredivided into NfL(high) and NfL(low) subgroups The divisionwas based on the median NfL value measured among healthycontrols (231 [21ndash28] pgmL) In the NfL(high) subgroup theNfL concentration was significantly elevated compared withhealthy controls (p = 0018 table 1) and in the further NfL vsDTI correlation analyses we focused on the NfL(high) subgroupBased on the demographic and clinical data the patients in theNfL(high) subgroupwere at amore advanced stage of the diseaseTheNfL(high) subgroup includedmore patients with SPMS (p =0003) and the patients in the subgroup were also older (p =0003) and had a higher EDSS score (p = 0005 table 1) Inevaluation using conventional MRI the NAWM white matter orcortical gray matter volumes and T1 and T2 lesion loads were notdifferent between NfL(low) and NfL(high) subgroups (table 1)
Associations between NfL and DTIOur results show that DTI metrics of diffuse neuroaxonaldamage within theNAWMassociatewith high serumNfL levelsin MS As a demonstration of this in the NfL(high) subgroup
several (n = 24) significant correlations were found betweenserum NfL levels and NAWM DTI indices both in the entireNAWM and in various brain subregions higher NfL levels wereassociated with lower fractional anisotropy and higher diffusivity(mean axial and radial) in the whole NAWM (figure 1) and inall its subregions except for fractional anisotropy in temporalNAWM axial diffusivity in parietal and temporal NAWM andmean diffusivity in remaining NAWM (table 2) Significance ofthese correlations was sustained after adjustment using the falsediscovery rate method for the number of DTI parameters (n =28) except for parietal fractional anisotropy Results remainedsimilar when the data were analyzed without patients who hadhad a relapse within the previous year before sampling (n = 3data not shown) No correlations between serumNfL levels andDTI metrics were observed in the NfL(low) subgroup or in theoverall MS cohort (table 2)
Multiple regression modeling was performed to further evaluatethe effect of DTI on NfL In the NfL(high) subgroup the DTIindices weremore significantly associated withNfL thanwith theclinical parameters included in the model (sex age disease typeEDSS score treatment and the presence of relapses within 1year before sampling) (table 3) On average 48 of the variancein NfL (SD 44) could be explained by DTI indices and theclinical parameters mentioned above (figure 2) There weremultiple (n = 26) significant associations between NfL and DTIin the whole NAWM and in all its 6 subregions All except 2 ofthe associations remained significant after adjustment with thefalse discovery rate method for the number of investigated var-iables (n = 32) The strongest associations were observed in thecingulate area between higher NfL and higher mean and radialdiffusivity and in deep and frontal NAWM between higher NfLand higher mean diffusivity (figure 2) Analyzing the datawithout the patients with a relapse (n = 3) within the previousyear did not alter the associations (data not shown)
Higher NfL was also associated with lower NAWM andcortical gray matter and higher T2 lesion volumes (figure 2)The volumes and clinical adjustments explained 434494 and 433 of the variance in NfL respectively
DiscussionPresent results demonstrate that the DTI-MRI measures ofNAWM correlate with serumNfL in MS We found that amongpatients with more advanced disease increased serumNfL levelsassociate with DTI measures reflecting diffuse microstructuraldamage ie decreased fractional anisotropy and increasedmeanaxial and radial diffusivities in the NAWM In a multiple re-gression model which was adjusted with sex age disease typeEDSS score treatment and the presence of relapses the DTIindices were more significantly associated with NfL than theabove-mentioned demographic and clinical parameters
The pathophysiology of MS involves acute and chronic mech-anisms that lead to gradual axonal and myelin damage that
6 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
Table 3 Multiple regression analyses evaluating separately the effect of DTI-MRI indices or volumetric brain MRI data onNfL in the NfL(high) subgroup of patients with MS
Variable in model DTI Sex Age Type EDSS score DMT Relapses
Entire NAWM
FA 0018 07 08 10 03 04 02
MD 0003 07 09 10 02 04 04
AD 0010 09 10 08 01 07 05
RD 0005 07 10 10 02 04 03
Frontal NAWM
FA 0045 08 10 10 02 07 02
MD 0001 05 07 10 02 06 04
AD 0004 05 09 08 03 07 04
RD 0003 06 09 09 02 06 03
Parietal NAWM
FA 0023 09 09 09 02 05 01
MD 0004 07 07 08 01 05 02
AD 0017 08 10 10 01 06 03
RD 0006 08 08 09 02 05 02
Temporal NAWM
FA 0116 07 07 10 03 05 03
MD 0007 06 10 10 01 05 05
AD 0009 07 08 08 01 05 05
RD 0020 06 09 10 02 05 04
Occipital NAWM
FA 0011 09 06 09 03 03 01
MD 0009 08 08 09 01 06 02
AD 0086 08 07 08 01 08 03
RD 0011 08 08 08 02 05 02
Cingulate NAWM
FA 0004 05 04 08 06 04 03
MD 0001 06 09 08 06 02 07
AD 0003 08 09 08 03 04 07
RD 0001 06 07 08 07 02 06
Deep NAWM
FA 0047 05 06 09 03 05 03
MD 0001 07 09 09 05 02 07
AD 0005 10 08 10 03 04 08
RD 0002 06 08 10 05 02 08
Volume (pf)
NAWM 0019 08 08 08 03 05 05
Continued
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 7
eventually manifest by brain atrophy and clinically by steadyworsening of disability in almost all patients over time This diffusepathology which in the majority of cases leads to progressiveworsening of the MS-related symptoms is now the focus of in-terest for patients and physicians as modern disease-modifyingtherapies have led to an almost complete suppression of relapseactivity while their impact on progression is at best modest Ourincomplete understanding of the molecular mechanisms that leadto progression is one of the main reasons for the failure in de-velopment ofmore efficacious therapies for secondary progressiveform of MS Another impediment is the relative insensitivity ofcurrent measurement tools to capture features of progression atsubclinical stages and hence to quantitate the effects of anti-neurodegenerative treatments at a time point where the overallbrain structure is still largely intact Specifically conventional MRIcan only quantitate brain volume loss as an end result of thedisease but does not provide insights into the underlying micro-structural changes DTI is a nonconventional MRI technique thatprovides a measure of such diffuse changes beyond volumetry
NfL is a highly sensitive biomarker to detect neuronal damageand is the first of its kind applicable in blood-derived probes22
However the cause and underlying pathology of the increasedNfL concentration may vary greatly both inter- and intra-individually In clinically isolated syndrome and RRMS acuteNfL elevation is particularly associated with focal in-flammation ie relapses and focal lesion formation4ndash823
However the indication for the contribution of diffuse pa-thology to higher NfL levels has so far been only incidentaland consisted of associations of increased NfL with moreadvanced clinical disability and brain atrophy824 Novelfindings presented here provide more direct evidence for theeffect of diffuse pathologic process in the NAWM on in-creased serum NfL We observed that among a subgroup ofolder and more disabled patients NfL associates both withradial and axial diffusivities In the context of MS these DTIscalars are thought to reflect myelin loss and axonal damagerespectively11 NfL also associated with mean diffusivity thatcan also indicate on axonal andmyelin loss11 The associationswere most pronounced in the cingulate deep and frontalNAWM
NfL differs in 2 important aspects from MRI first as asignal of neurodegeneration it reflects ongoing pathologicprocesses in real time while imaging captures morphologicfeatures that are inherently retrospective Second NfLlevels reflect neuronal damage in the CNS comprehen-sively ie in addition to brain pathology it also capturesspinal cord pathology that is not routinely evaluated byimaging in the workup of MS for individual patients As adownside NfL levels cannot differentiate betweenacute inflammatory and chronic neurodegenerative diseaseactivity Our results suggest that DTI-MRI allows us tocategorize the pathogenic source of NfL at least semi-quantitatively We demonstrate that both diffuse axonaldamage within the NAWM measured using DTI and focalinflammatory white matter lesion load measured usingconventional MRI contributed to the elevated NfL serumlevels This highlights that both focal and diffuse pathologicchanges must be taken into account when levels of NfL inindividual patients are interpreted A somewhat largerproportion of NfL variance was explained by DTI indices inaverage than by T2 lesion load This implies that in non-relapsing patients the ongoing diffuse microstructuraldamage in the NAWM outside lesions might predominateover focal axonal damage that has occurred within lesionsas a source of NfL in serum Accordingly in other neuro-logic disease with brain diffuse pathology such as amyo-trophic lateral sclerosis Alzheimer disease traumatic braininjury and age-related white matter pathology elevatedserum NfL levels have been associated with DTI-measurable diffuse damage within the CNS25ndash28 Thesestudies are further evidence for the concept that braindiffuse chronic neuropathologic mechanisms contribute toNfL release and that nonconventional MRI techniques areable to identify the CNS areas of its morphologic source
Current results demonstrate that the cingulate area is thebrain area with the most prominent white matter tractdamage outside focal lesions Periventricular regions arecharacteristically the brain area with a particularly heavylesion load in MS29 We addressed the question whetherthe cingulum-NAWM DTI abnormalities would arise from
Table 3 Multiple regression analyses evaluating separately the effect of DTI-MRI indices or volumetric brain MRI data onNfL in the NfL(high) subgroup of patients with MS (continued)
Variable in model DTI Sex Age Type EDSS score DMT Relapses
GMctx 0004 09 02 04 003 03 06
T2 0019 06 03 07 04 07 03
T1 015 07 04 10 05 07 03
Abbreviations AD = axial diffusivity DMT = disease-modifying treatment DTI = diffusion tensor imaging EDSS = Expanded Disability Status Scale FA =fractional anisotropy GMctx = cortical gray matter MD = mean diffusivity NAWM = normal-appearing white matter NfL = neurofilament light chain pf =parenchymal fraction RD = radial diffusivityPatients withMSwere divided into NfL(low) and NfL(high) subgroups based on themedian NfL value of healthy controls (231 pgmL) Model was adjusted bysex age disease type (relapsing-remittingsecondary progressive) EDSS score treatment (no first- and second-line) and presence of relapses within 1 yearbefore sampling (yesno) Bolded p values indicate that the parameter affected the logarithmofNfL value statistically significantly whenDTI or volume and allthe adjusting variables were in the models
8 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
the dirty-appearing white matter due to close proximity to aheavy lesion load but the relative lesion volume in the cingulatearea was not found to be significantly larger compared withlesion load in other regions or in the whole NAWM Thereforeit seems unlikely that differences in the DTI indices would besolely due to dirty-appearing white matter adjacent to lesionswhereas it is more likely that remote lesions lead to DTI changesin the cingulateNAWM throughWallerian degeneration30 Alsoother studies have observed diffusion changes in the cingulatearea of patients with MS31ndash33 Because the cingulum bundle offibers projects from the cingulate gyrus to the entorhinal cortexin the brain allowing for communication between components
of the limbic system the DTI abnormalities may be the mor-phologic substrate for impaired cognitive function and fatiguesymptoms that are typical in progressive MS3435
Present results are in line with previous observations on therelation between NfL concentration and MS disease sub-type with progressive patients having greater NfL levelscompared with relapsing-remitting patients568ndash102324
However the median NfL value of the entire patient cohortwas relatively low compared with healthy controls which isthe main limitation of our study Because of this we werenot able to use the previously defined cutoff values3 for
Figure 2 The effect of DTI-MRI indices or volumetric MRI data on NfL in the NfL(high) subgroup
The NfL(high) subgroup is comprised of patients with serum NfL above the median value of healthy controls (231 pgmL) Logarithmic serum NfL wasmodeled separately by DTI indices of the entire and parcellated NAWM and volumetric brain MRI data using multiple regression analyses Models wereadjusted by sex age disease type (RRMSSPMS) the EDSS score treatment (no first and second line) and presence of relapseswithin 1 year before sampling(yesno) The results are illustrated using dot and whisker plots in which red dots represent standardized regression coefficients and red lines represent theCIs of the estimates Significant p values of the DTI parameters and the percentage of variance in the response that can be explained by the independentvariables (R2) are also shown The results were ordered according to the R2 value All except 2 of the p values (fractional anisotropy of the frontal and deepNAWM) shown in the figure remained significant after adjustment using the false discovery ratemethod for the number of investigated variables (n = 32) AD =axial diffusivity DTI = diffusion tensor imaging EDSS = Expanded Disability Status Scale FA = fractional anisotropy GMctx = cortical gray matter MD =meandiffusivity NAWM = normal-appearing white matter NfL = neurofilament light chain PF = parenchymal fraction RD = radial diffusivity RRMS = relapsing-remitting MS SPMS = secondary progressive MS WM = white matter
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 9
NfL(low) and NfL(high) subgroups Instead we used themedian value of healthy controls Although the divider usedin our study was based on a small group of individualssimilar values have been observed in previous largerstudies68 One reason for the low median NfL level couldbe that the patients with RRMS are represented with arelatively benign disease course In addition our patientswith RRMS were sampled during a relapse-free time andhad no signs of ongoing focal inflammatory activity in theMRI Moreover most of the patients (68) were using adisease-modifying treatment at the time of evaluationHence similar as observed in other studies35233637 thisresulted in low serum NfL concentration Following thesimilarity in the NfL levels between healthy controls andpatients with MS the main results of our study are based ona relatively low number of patients In addition the patientnumbers under different immunomodulatory treatmentsvaried substantially However the heterogeneity of thecohort was taken account in the multiple regression modelwhere both EDSS score and disease-modifying treatmentwere used as adjustments
Our findings suggest that elevated serum NfL in MS resultsfrom neuroaxonal damage both within the NAWM and infocal T2 lesions The association between DTI-measurablediffuse microstructural white matter damage and serumNfL isfurther conceptual evidence for the latter being a usefulmonitoring tool in assessing the degree of ongoing neuro-degenerative processes in MS
AcknowledgmentThe authors thank all patients with MS and healthycontrols who participated in this study They also thank thepersonnel of Turku PET Centre for their excellenttechnical assistance
Study fundingThis work was supported by the Finnish Academy the SigridJuselius Foundation
DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgNN for full disclosures
Publication historyReceived by Neurology Neuroimmunology amp NeuroinflammationMay 8 2020 Accepted in final form October 21 2020
References1 Reich DS Lucchinetti CF Calabresi PA Multiple sclerosis N Engl J Med 2018378
169ndash1802 Varhaug KN Torkildsen Oslash Myhr KM Vedeler CA Neurofilament light chain as a
biomarker in multiple sclerosis Front Neurol 201910338
Appendix Authors
Author Location Contribution
MaijaSaraste PhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Acquisition of dataanalyzed andinterpreted the data anddrafted and revised themanuscript forintellectual content
Appendix (continued)
Author Location Contribution
SvetlanaBezukladovaMSc
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Acquisition of dataanalyzed and interpretedthe data and revised themanuscript
MarkusMatilainenPhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Statistical analysisdrafted the figures andrevised the manuscript
Jouni TuiskuMSc
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Major role in theacquisition of data andrevised the manuscript
EeroRissanenMDPhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Designed andconceptualized the studyand acquisition of data
MarcusSucksdorffMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Major role in theacquisition of data
SiniLaaksonenMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Acquisition of data
AnnaVuorimaaMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Acquisition of data
Jens KuhleMD PhD
Neurologic Clinic andPoliclinic Departments ofMedicine Biomedicineand Clinical ResearchUniversity Hospital BaselSwitzerland
Major role in theacquisition of data
DavidLeppert MDPhD
Neurologic Clinic andPoliclinic Departments ofMedicine Biomedicineand Clinical ResearchUniversity Hospital BaselSwitzerland
Revised the manuscriptfor intellectual content
Laura AirasMD PhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Designed andconceptualized the studyand revised themanuscript for intellectualcontent
10 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
3 Kuhle J Kropshofer H Haering DA et al Blood neurofilament light chain as abiomarker of MS disease activity and treatment response Neurology 201992e1007ndashe1015
4 Varhaug KN Barro C Bjoslashrnevik K et al Neurofilament light chain predicts diseaseactivity in relapsing-remitting MS Neurol Neuroimmunol Neuroinflamm 20185e422
5 Novakova L Zetterberg H Sundstrom P et al Monitoring disease activity in multiplesclerosis using serum neurofilament light protein Neurology 2017892230ndash2237
6 Disanto G Barro C Benkert P et al Serum neurofilament light a biomarker ofneuronal damage in multiple sclerosis Ann Neurol 201781857ndash870
7 Siller N Kuhle J Muthuraman M et al Serum neurofilament light chain is a bio-marker of acute and chronic neuronal damage in early multiple sclerosis Mult Scler201925678ndash686
8 Barro C Benkert P Disanto G et al Serum neurofilament as a predictor of diseaseworsening and brain and spinal cord atrophy in multiple sclerosis Brain 20181412382ndash2391
9 Ferraro D Guicciardi C De Biasi S et al Plasma neurofilaments correlate withdisability in progressive multiple sclerosis patients Acta Neurol Scand 202014116ndash21
10 Jakimovski D Zivadinov R RamanthanM et al Serum neurofilament light chain levelassociations with clinical and cognitive performance in multiple sclerosis a longitu-dinal retrospective 5-year study Mult Scler 2020261670ndash1681
11 Sbardella E Tona F Petsas N Pantano P DTI measurements in multiple sclerosisevaluation of brain damage and clinical implications Mult Scler Int 20132013671730
12 Tae WS Ham BJ Pyun SB Kang SH Kim BJ Current clinical applications ofdiffusion-tensor imaging in neurological disorders J Clin Neurol 201814129ndash140
13 Kurtzke JF Rating neurologic impairment in multiple sclerosis an expanded disabilitystatus scale (EDSS) Neurology 1983331444ndash1452
14 Kuhle J Barro C Andreasson U et al Comparison of three analytical platforms forquantification of the neurofilament light chain in blood samples ELISA electro-chemiluminescence immunoassay and SimoaClinChemLabMed 2016541655ndash1661
15 Bezukladova S Tuisku J Matilainen M et al Insights into disseminated MS brainpathology with multimodal diffusion tensor and PET imaging Neurol NeuroimmunolNeuroinflamm 20207e691
16 Schmidt P Gaser C Arsic M et al An automated tool for detection of FLAIR-hyperintense white-matter lesions in Multiple Sclerosis Neuroimage 2012593774ndash3783
17 Salat DH Greve DN Pacheco JL et al Regional white matter volume differences innondemented aging and Alzheimerrsquos disease Neuroimage 2009441247ndash1258
18 Rissanen E Tuisku J Vahlberg T et al Microglial activation white matter tractdamage and disability in MS Neurol Neuroimmunol Neuroinflamm 20185e443
19 Leemans A Jeurissen B Sijbers J et al ExploreDTI a graphical toolbox for pro-cessing analysing and visualizing diffusion MR data Presented at the 17th AnnualMeeting of International Society for Magnetic Resonance in Medicine Honolulu HI2009 p 3537
20 Benjamini Y Hochberg Y Controlling the false discovery rate a practical and pow-erful approach to multiple testing J R Stat Soc Ser B 199557289ndash300
21 Gelman A Scaling regression inputs by dividing by two standard deviations Stat Med2008272865ndash2873
22 Khalil M Teunissen CE Otto M et al Neurofilaments as biomarkers in neurologicaldisorders Nat Rev Neurol 201814577ndash589
23 Canto E Barro C Zhao C et al Association between serum neurofilament light chainlevels and long-term disease course among patients with multiple sclerosis followedup for 12 years JAMA Neurol 2019761359ndash1366
24 Hogel H Rissanen E Barro C et al Serum glial fibrillary acidic protein correlates withmultiple sclerosis disease severity Mult Scler 202026210ndash219
25 Menke RA Gray E Lu CH et al CSF neurofilament light chain reflects corticospinaltract degeneration in ALS Ann Clin Transl Neurol 20152748ndash755
26 Moore EE Hohman TJ Badami FS et al Neurofilament relates to white mattermicrostructure in older adults Neurobiol Aging 201870233ndash241
27 Ljungqvist J Zetterberg H Mitsis M Blennow K Skoglund T Serum neurofilamentlight protein as a marker for diffuse axonal injury results from a case series studyJ Neurotrauma 2017341124ndash1127
28 Schultz SA Strain JF Adedokun A et al Serum neurofilament light chain levels areassociated with white matter integrity in autosomal dominant Alzheimerrsquos diseaseNeurobiol Dis 2020142104960
29 Narayanan S Fu L Pioro E et al Imaging of axonal damage in multiple sclerosisspatial distribution of magnetic resonance imaging lesions Ann Neurol 199741385ndash391
30 Ciccarelli O Werring DJ Barker GJ et al A study of the mechanisms of normal-appearing white matter damage in multiple sclerosis using diffusion tensor imagingevidence of Wallerian degeneration J Neurol 2003250287ndash292
31 Pokryszko-Dragan A Banaszek A Nowakowska-Kotas M et al Diffusion tensorimaging findings in the multiple sclerosis patients and their relationships to variousaspects of disability J Neurol Sci 2018391127ndash133
32 Kolasa M Hakulinen U Brander A et al Diffusion tensor imaging and disabilityprogression in multiple sclerosis a 4-year follow-up study Brain Behav 20199e01194
33 Andersen O Hildeman A Longfils M et al Diffusion tensor imaging in multiplesclerosis at different final outcomes Acta Neurol Scand 2018137165ndash173
34 Koenig KA Sakaie KE Lowe MJ et al The relationship between cognitive functionand high-resolution diffusion tensor MRI of the cingulum bundle in multiple sclerosisMult Scler 2015211794ndash1801
35 Pardini M Bonzano L Bergamino M et al Cingulum bundle alterations underliesubjective fatigue in multiple sclerosis Mult Scler 201521442ndash447
36 Sejbaek T Nielsen HH Penner N et al Dimethyl fumarate decreases neurofilamentlight chain in CSF and blood of treatment naıve relapsing MS patients J NeurolNeurosurg Psychiatry 2019901324ndash1330
37 Piehl F Kockum I Khademi M et al Plasma neurofilament light chain levels inpatients with MS switching from injectable therapies to fingolimod Mult Scler 2018241046ndash1054
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 11
DOI 101212NXI000000000000092620218 Neurol Neuroimmunol Neuroinflamm
Maija Saraste Svetlana Bezukladova Markus Matilainen et al High serum neurofilament associates with diffuse white matter damage in MS
This information is current as of December 8 2020
ServicesUpdated Information amp
httpnnneurologyorgcontent81e926fullhtmlincluding high resolution figures can be found at
References httpnnneurologyorgcontent81e926fullhtmlref-list-1
This article cites 36 articles 3 of which you can access for free at
Citations httpnnneurologyorgcontent81e926fullhtmlotherarticles
This article has been cited by 1 HighWire-hosted articles
Subspecialty Collections
httpnnneurologyorgcgicollectionmultiple_sclerosisMultiple sclerosis
httpnnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpnnneurologyorgmiscaboutxhtmlpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnnneurologyorgmiscaddirxhtmlreprintsusInformation about ordering reprints can be found online
Academy of Neurology All rights reserved Online ISSN 2332-7812Copyright copy 2020 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the AmericanPublished since April 2014 it is an open-access online-only continuous publication journal Copyright
is an official journal of the American Academy of NeurologyNeurol Neuroimmunol Neuroinflamm
are shown in table 1 The median (interquartile range [IQR])age of this cohort was 483 (43ndash53) years which is similar tothe age of the healthy individuals in the control group (476[44ndash53] p = 1) Overall the patient cohort was representedwith a quite stable disease as only 13 of the patients (9 RRMSand 1 SPMS) had had a relapse within 1 year before samplingSixty-eight percent of the patients (45 RRMS and 9 SPMS)were on disease-modifying therapy (dimethyl fumarate n = 3fingolimod n = 14 glatiramer acetate n = 5 interferon beta-1a n= 6 natalizumab n = 8 rituximab n = 4 and teriflunomide n =14) In MRI white matter and cortical gray matter volumeswere decreased in patients compared with healthy controlsDTI measures were obtained from both the entire cerebralNAWM and from 6 different NAWM subregions In the cin-gulate area there were significant differences in all 4 DTI pa-rameters in patients with MS compared with healthy controlsfractional anisotropy was decreased whereas mean diffusivityradial diffusivity and axial diffusivity were increased (table 1)
Serum NfL levels in patients with MS andhealthy controlsThe median (IQR) NfL level was higher in patients withSPMS compared with patients with RRMS (314 [22ndash44] vs188 [14ndash26] pgmL p = 0001 table 1) NfL levels in thewhole MS group were not different from healthy controls(217 [15ndash31] vs 231 [21ndash28] pgmL p = 03) regardlesswhether the analysis was performed with age correctionAcute inflammation did not appear to be a factor to affect theNfL level in this cohort with relatively modest acute in-flammatory activity as the median NfL value among thepatients with or without relapse within the previous year wascomparable (223 [16ndash32] vs 186 [15ndash25] pgmL p = 03Wilcoxon rank-sum data not shown)
Characterization of patients with highNfL levelsTo explore associations between increased NfL levels and DTI-measurable diffuse neuroaxonal damage patients with MS weredivided into NfL(high) and NfL(low) subgroups The divisionwas based on the median NfL value measured among healthycontrols (231 [21ndash28] pgmL) In the NfL(high) subgroup theNfL concentration was significantly elevated compared withhealthy controls (p = 0018 table 1) and in the further NfL vsDTI correlation analyses we focused on the NfL(high) subgroupBased on the demographic and clinical data the patients in theNfL(high) subgroupwere at amore advanced stage of the diseaseTheNfL(high) subgroup includedmore patients with SPMS (p =0003) and the patients in the subgroup were also older (p =0003) and had a higher EDSS score (p = 0005 table 1) Inevaluation using conventional MRI the NAWM white matter orcortical gray matter volumes and T1 and T2 lesion loads were notdifferent between NfL(low) and NfL(high) subgroups (table 1)
Associations between NfL and DTIOur results show that DTI metrics of diffuse neuroaxonaldamage within theNAWMassociatewith high serumNfL levelsin MS As a demonstration of this in the NfL(high) subgroup
several (n = 24) significant correlations were found betweenserum NfL levels and NAWM DTI indices both in the entireNAWM and in various brain subregions higher NfL levels wereassociated with lower fractional anisotropy and higher diffusivity(mean axial and radial) in the whole NAWM (figure 1) and inall its subregions except for fractional anisotropy in temporalNAWM axial diffusivity in parietal and temporal NAWM andmean diffusivity in remaining NAWM (table 2) Significance ofthese correlations was sustained after adjustment using the falsediscovery rate method for the number of DTI parameters (n =28) except for parietal fractional anisotropy Results remainedsimilar when the data were analyzed without patients who hadhad a relapse within the previous year before sampling (n = 3data not shown) No correlations between serumNfL levels andDTI metrics were observed in the NfL(low) subgroup or in theoverall MS cohort (table 2)
Multiple regression modeling was performed to further evaluatethe effect of DTI on NfL In the NfL(high) subgroup the DTIindices weremore significantly associated withNfL thanwith theclinical parameters included in the model (sex age disease typeEDSS score treatment and the presence of relapses within 1year before sampling) (table 3) On average 48 of the variancein NfL (SD 44) could be explained by DTI indices and theclinical parameters mentioned above (figure 2) There weremultiple (n = 26) significant associations between NfL and DTIin the whole NAWM and in all its 6 subregions All except 2 ofthe associations remained significant after adjustment with thefalse discovery rate method for the number of investigated var-iables (n = 32) The strongest associations were observed in thecingulate area between higher NfL and higher mean and radialdiffusivity and in deep and frontal NAWM between higher NfLand higher mean diffusivity (figure 2) Analyzing the datawithout the patients with a relapse (n = 3) within the previousyear did not alter the associations (data not shown)
Higher NfL was also associated with lower NAWM andcortical gray matter and higher T2 lesion volumes (figure 2)The volumes and clinical adjustments explained 434494 and 433 of the variance in NfL respectively
DiscussionPresent results demonstrate that the DTI-MRI measures ofNAWM correlate with serumNfL in MS We found that amongpatients with more advanced disease increased serumNfL levelsassociate with DTI measures reflecting diffuse microstructuraldamage ie decreased fractional anisotropy and increasedmeanaxial and radial diffusivities in the NAWM In a multiple re-gression model which was adjusted with sex age disease typeEDSS score treatment and the presence of relapses the DTIindices were more significantly associated with NfL than theabove-mentioned demographic and clinical parameters
The pathophysiology of MS involves acute and chronic mech-anisms that lead to gradual axonal and myelin damage that
6 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
Table 3 Multiple regression analyses evaluating separately the effect of DTI-MRI indices or volumetric brain MRI data onNfL in the NfL(high) subgroup of patients with MS
Variable in model DTI Sex Age Type EDSS score DMT Relapses
Entire NAWM
FA 0018 07 08 10 03 04 02
MD 0003 07 09 10 02 04 04
AD 0010 09 10 08 01 07 05
RD 0005 07 10 10 02 04 03
Frontal NAWM
FA 0045 08 10 10 02 07 02
MD 0001 05 07 10 02 06 04
AD 0004 05 09 08 03 07 04
RD 0003 06 09 09 02 06 03
Parietal NAWM
FA 0023 09 09 09 02 05 01
MD 0004 07 07 08 01 05 02
AD 0017 08 10 10 01 06 03
RD 0006 08 08 09 02 05 02
Temporal NAWM
FA 0116 07 07 10 03 05 03
MD 0007 06 10 10 01 05 05
AD 0009 07 08 08 01 05 05
RD 0020 06 09 10 02 05 04
Occipital NAWM
FA 0011 09 06 09 03 03 01
MD 0009 08 08 09 01 06 02
AD 0086 08 07 08 01 08 03
RD 0011 08 08 08 02 05 02
Cingulate NAWM
FA 0004 05 04 08 06 04 03
MD 0001 06 09 08 06 02 07
AD 0003 08 09 08 03 04 07
RD 0001 06 07 08 07 02 06
Deep NAWM
FA 0047 05 06 09 03 05 03
MD 0001 07 09 09 05 02 07
AD 0005 10 08 10 03 04 08
RD 0002 06 08 10 05 02 08
Volume (pf)
NAWM 0019 08 08 08 03 05 05
Continued
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 7
eventually manifest by brain atrophy and clinically by steadyworsening of disability in almost all patients over time This diffusepathology which in the majority of cases leads to progressiveworsening of the MS-related symptoms is now the focus of in-terest for patients and physicians as modern disease-modifyingtherapies have led to an almost complete suppression of relapseactivity while their impact on progression is at best modest Ourincomplete understanding of the molecular mechanisms that leadto progression is one of the main reasons for the failure in de-velopment ofmore efficacious therapies for secondary progressiveform of MS Another impediment is the relative insensitivity ofcurrent measurement tools to capture features of progression atsubclinical stages and hence to quantitate the effects of anti-neurodegenerative treatments at a time point where the overallbrain structure is still largely intact Specifically conventional MRIcan only quantitate brain volume loss as an end result of thedisease but does not provide insights into the underlying micro-structural changes DTI is a nonconventional MRI technique thatprovides a measure of such diffuse changes beyond volumetry
NfL is a highly sensitive biomarker to detect neuronal damageand is the first of its kind applicable in blood-derived probes22
However the cause and underlying pathology of the increasedNfL concentration may vary greatly both inter- and intra-individually In clinically isolated syndrome and RRMS acuteNfL elevation is particularly associated with focal in-flammation ie relapses and focal lesion formation4ndash823
However the indication for the contribution of diffuse pa-thology to higher NfL levels has so far been only incidentaland consisted of associations of increased NfL with moreadvanced clinical disability and brain atrophy824 Novelfindings presented here provide more direct evidence for theeffect of diffuse pathologic process in the NAWM on in-creased serum NfL We observed that among a subgroup ofolder and more disabled patients NfL associates both withradial and axial diffusivities In the context of MS these DTIscalars are thought to reflect myelin loss and axonal damagerespectively11 NfL also associated with mean diffusivity thatcan also indicate on axonal andmyelin loss11 The associationswere most pronounced in the cingulate deep and frontalNAWM
NfL differs in 2 important aspects from MRI first as asignal of neurodegeneration it reflects ongoing pathologicprocesses in real time while imaging captures morphologicfeatures that are inherently retrospective Second NfLlevels reflect neuronal damage in the CNS comprehen-sively ie in addition to brain pathology it also capturesspinal cord pathology that is not routinely evaluated byimaging in the workup of MS for individual patients As adownside NfL levels cannot differentiate betweenacute inflammatory and chronic neurodegenerative diseaseactivity Our results suggest that DTI-MRI allows us tocategorize the pathogenic source of NfL at least semi-quantitatively We demonstrate that both diffuse axonaldamage within the NAWM measured using DTI and focalinflammatory white matter lesion load measured usingconventional MRI contributed to the elevated NfL serumlevels This highlights that both focal and diffuse pathologicchanges must be taken into account when levels of NfL inindividual patients are interpreted A somewhat largerproportion of NfL variance was explained by DTI indices inaverage than by T2 lesion load This implies that in non-relapsing patients the ongoing diffuse microstructuraldamage in the NAWM outside lesions might predominateover focal axonal damage that has occurred within lesionsas a source of NfL in serum Accordingly in other neuro-logic disease with brain diffuse pathology such as amyo-trophic lateral sclerosis Alzheimer disease traumatic braininjury and age-related white matter pathology elevatedserum NfL levels have been associated with DTI-measurable diffuse damage within the CNS25ndash28 Thesestudies are further evidence for the concept that braindiffuse chronic neuropathologic mechanisms contribute toNfL release and that nonconventional MRI techniques areable to identify the CNS areas of its morphologic source
Current results demonstrate that the cingulate area is thebrain area with the most prominent white matter tractdamage outside focal lesions Periventricular regions arecharacteristically the brain area with a particularly heavylesion load in MS29 We addressed the question whetherthe cingulum-NAWM DTI abnormalities would arise from
Table 3 Multiple regression analyses evaluating separately the effect of DTI-MRI indices or volumetric brain MRI data onNfL in the NfL(high) subgroup of patients with MS (continued)
Variable in model DTI Sex Age Type EDSS score DMT Relapses
GMctx 0004 09 02 04 003 03 06
T2 0019 06 03 07 04 07 03
T1 015 07 04 10 05 07 03
Abbreviations AD = axial diffusivity DMT = disease-modifying treatment DTI = diffusion tensor imaging EDSS = Expanded Disability Status Scale FA =fractional anisotropy GMctx = cortical gray matter MD = mean diffusivity NAWM = normal-appearing white matter NfL = neurofilament light chain pf =parenchymal fraction RD = radial diffusivityPatients withMSwere divided into NfL(low) and NfL(high) subgroups based on themedian NfL value of healthy controls (231 pgmL) Model was adjusted bysex age disease type (relapsing-remittingsecondary progressive) EDSS score treatment (no first- and second-line) and presence of relapses within 1 yearbefore sampling (yesno) Bolded p values indicate that the parameter affected the logarithmofNfL value statistically significantly whenDTI or volume and allthe adjusting variables were in the models
8 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
the dirty-appearing white matter due to close proximity to aheavy lesion load but the relative lesion volume in the cingulatearea was not found to be significantly larger compared withlesion load in other regions or in the whole NAWM Thereforeit seems unlikely that differences in the DTI indices would besolely due to dirty-appearing white matter adjacent to lesionswhereas it is more likely that remote lesions lead to DTI changesin the cingulateNAWM throughWallerian degeneration30 Alsoother studies have observed diffusion changes in the cingulatearea of patients with MS31ndash33 Because the cingulum bundle offibers projects from the cingulate gyrus to the entorhinal cortexin the brain allowing for communication between components
of the limbic system the DTI abnormalities may be the mor-phologic substrate for impaired cognitive function and fatiguesymptoms that are typical in progressive MS3435
Present results are in line with previous observations on therelation between NfL concentration and MS disease sub-type with progressive patients having greater NfL levelscompared with relapsing-remitting patients568ndash102324
However the median NfL value of the entire patient cohortwas relatively low compared with healthy controls which isthe main limitation of our study Because of this we werenot able to use the previously defined cutoff values3 for
Figure 2 The effect of DTI-MRI indices or volumetric MRI data on NfL in the NfL(high) subgroup
The NfL(high) subgroup is comprised of patients with serum NfL above the median value of healthy controls (231 pgmL) Logarithmic serum NfL wasmodeled separately by DTI indices of the entire and parcellated NAWM and volumetric brain MRI data using multiple regression analyses Models wereadjusted by sex age disease type (RRMSSPMS) the EDSS score treatment (no first and second line) and presence of relapseswithin 1 year before sampling(yesno) The results are illustrated using dot and whisker plots in which red dots represent standardized regression coefficients and red lines represent theCIs of the estimates Significant p values of the DTI parameters and the percentage of variance in the response that can be explained by the independentvariables (R2) are also shown The results were ordered according to the R2 value All except 2 of the p values (fractional anisotropy of the frontal and deepNAWM) shown in the figure remained significant after adjustment using the false discovery ratemethod for the number of investigated variables (n = 32) AD =axial diffusivity DTI = diffusion tensor imaging EDSS = Expanded Disability Status Scale FA = fractional anisotropy GMctx = cortical gray matter MD =meandiffusivity NAWM = normal-appearing white matter NfL = neurofilament light chain PF = parenchymal fraction RD = radial diffusivity RRMS = relapsing-remitting MS SPMS = secondary progressive MS WM = white matter
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 9
NfL(low) and NfL(high) subgroups Instead we used themedian value of healthy controls Although the divider usedin our study was based on a small group of individualssimilar values have been observed in previous largerstudies68 One reason for the low median NfL level couldbe that the patients with RRMS are represented with arelatively benign disease course In addition our patientswith RRMS were sampled during a relapse-free time andhad no signs of ongoing focal inflammatory activity in theMRI Moreover most of the patients (68) were using adisease-modifying treatment at the time of evaluationHence similar as observed in other studies35233637 thisresulted in low serum NfL concentration Following thesimilarity in the NfL levels between healthy controls andpatients with MS the main results of our study are based ona relatively low number of patients In addition the patientnumbers under different immunomodulatory treatmentsvaried substantially However the heterogeneity of thecohort was taken account in the multiple regression modelwhere both EDSS score and disease-modifying treatmentwere used as adjustments
Our findings suggest that elevated serum NfL in MS resultsfrom neuroaxonal damage both within the NAWM and infocal T2 lesions The association between DTI-measurablediffuse microstructural white matter damage and serumNfL isfurther conceptual evidence for the latter being a usefulmonitoring tool in assessing the degree of ongoing neuro-degenerative processes in MS
AcknowledgmentThe authors thank all patients with MS and healthycontrols who participated in this study They also thank thepersonnel of Turku PET Centre for their excellenttechnical assistance
Study fundingThis work was supported by the Finnish Academy the SigridJuselius Foundation
DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgNN for full disclosures
Publication historyReceived by Neurology Neuroimmunology amp NeuroinflammationMay 8 2020 Accepted in final form October 21 2020
References1 Reich DS Lucchinetti CF Calabresi PA Multiple sclerosis N Engl J Med 2018378
169ndash1802 Varhaug KN Torkildsen Oslash Myhr KM Vedeler CA Neurofilament light chain as a
biomarker in multiple sclerosis Front Neurol 201910338
Appendix Authors
Author Location Contribution
MaijaSaraste PhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Acquisition of dataanalyzed andinterpreted the data anddrafted and revised themanuscript forintellectual content
Appendix (continued)
Author Location Contribution
SvetlanaBezukladovaMSc
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Acquisition of dataanalyzed and interpretedthe data and revised themanuscript
MarkusMatilainenPhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Statistical analysisdrafted the figures andrevised the manuscript
Jouni TuiskuMSc
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Major role in theacquisition of data andrevised the manuscript
EeroRissanenMDPhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Designed andconceptualized the studyand acquisition of data
MarcusSucksdorffMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Major role in theacquisition of data
SiniLaaksonenMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Acquisition of data
AnnaVuorimaaMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Acquisition of data
Jens KuhleMD PhD
Neurologic Clinic andPoliclinic Departments ofMedicine Biomedicineand Clinical ResearchUniversity Hospital BaselSwitzerland
Major role in theacquisition of data
DavidLeppert MDPhD
Neurologic Clinic andPoliclinic Departments ofMedicine Biomedicineand Clinical ResearchUniversity Hospital BaselSwitzerland
Revised the manuscriptfor intellectual content
Laura AirasMD PhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Designed andconceptualized the studyand revised themanuscript for intellectualcontent
10 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
3 Kuhle J Kropshofer H Haering DA et al Blood neurofilament light chain as abiomarker of MS disease activity and treatment response Neurology 201992e1007ndashe1015
4 Varhaug KN Barro C Bjoslashrnevik K et al Neurofilament light chain predicts diseaseactivity in relapsing-remitting MS Neurol Neuroimmunol Neuroinflamm 20185e422
5 Novakova L Zetterberg H Sundstrom P et al Monitoring disease activity in multiplesclerosis using serum neurofilament light protein Neurology 2017892230ndash2237
6 Disanto G Barro C Benkert P et al Serum neurofilament light a biomarker ofneuronal damage in multiple sclerosis Ann Neurol 201781857ndash870
7 Siller N Kuhle J Muthuraman M et al Serum neurofilament light chain is a bio-marker of acute and chronic neuronal damage in early multiple sclerosis Mult Scler201925678ndash686
8 Barro C Benkert P Disanto G et al Serum neurofilament as a predictor of diseaseworsening and brain and spinal cord atrophy in multiple sclerosis Brain 20181412382ndash2391
9 Ferraro D Guicciardi C De Biasi S et al Plasma neurofilaments correlate withdisability in progressive multiple sclerosis patients Acta Neurol Scand 202014116ndash21
10 Jakimovski D Zivadinov R RamanthanM et al Serum neurofilament light chain levelassociations with clinical and cognitive performance in multiple sclerosis a longitu-dinal retrospective 5-year study Mult Scler 2020261670ndash1681
11 Sbardella E Tona F Petsas N Pantano P DTI measurements in multiple sclerosisevaluation of brain damage and clinical implications Mult Scler Int 20132013671730
12 Tae WS Ham BJ Pyun SB Kang SH Kim BJ Current clinical applications ofdiffusion-tensor imaging in neurological disorders J Clin Neurol 201814129ndash140
13 Kurtzke JF Rating neurologic impairment in multiple sclerosis an expanded disabilitystatus scale (EDSS) Neurology 1983331444ndash1452
14 Kuhle J Barro C Andreasson U et al Comparison of three analytical platforms forquantification of the neurofilament light chain in blood samples ELISA electro-chemiluminescence immunoassay and SimoaClinChemLabMed 2016541655ndash1661
15 Bezukladova S Tuisku J Matilainen M et al Insights into disseminated MS brainpathology with multimodal diffusion tensor and PET imaging Neurol NeuroimmunolNeuroinflamm 20207e691
16 Schmidt P Gaser C Arsic M et al An automated tool for detection of FLAIR-hyperintense white-matter lesions in Multiple Sclerosis Neuroimage 2012593774ndash3783
17 Salat DH Greve DN Pacheco JL et al Regional white matter volume differences innondemented aging and Alzheimerrsquos disease Neuroimage 2009441247ndash1258
18 Rissanen E Tuisku J Vahlberg T et al Microglial activation white matter tractdamage and disability in MS Neurol Neuroimmunol Neuroinflamm 20185e443
19 Leemans A Jeurissen B Sijbers J et al ExploreDTI a graphical toolbox for pro-cessing analysing and visualizing diffusion MR data Presented at the 17th AnnualMeeting of International Society for Magnetic Resonance in Medicine Honolulu HI2009 p 3537
20 Benjamini Y Hochberg Y Controlling the false discovery rate a practical and pow-erful approach to multiple testing J R Stat Soc Ser B 199557289ndash300
21 Gelman A Scaling regression inputs by dividing by two standard deviations Stat Med2008272865ndash2873
22 Khalil M Teunissen CE Otto M et al Neurofilaments as biomarkers in neurologicaldisorders Nat Rev Neurol 201814577ndash589
23 Canto E Barro C Zhao C et al Association between serum neurofilament light chainlevels and long-term disease course among patients with multiple sclerosis followedup for 12 years JAMA Neurol 2019761359ndash1366
24 Hogel H Rissanen E Barro C et al Serum glial fibrillary acidic protein correlates withmultiple sclerosis disease severity Mult Scler 202026210ndash219
25 Menke RA Gray E Lu CH et al CSF neurofilament light chain reflects corticospinaltract degeneration in ALS Ann Clin Transl Neurol 20152748ndash755
26 Moore EE Hohman TJ Badami FS et al Neurofilament relates to white mattermicrostructure in older adults Neurobiol Aging 201870233ndash241
27 Ljungqvist J Zetterberg H Mitsis M Blennow K Skoglund T Serum neurofilamentlight protein as a marker for diffuse axonal injury results from a case series studyJ Neurotrauma 2017341124ndash1127
28 Schultz SA Strain JF Adedokun A et al Serum neurofilament light chain levels areassociated with white matter integrity in autosomal dominant Alzheimerrsquos diseaseNeurobiol Dis 2020142104960
29 Narayanan S Fu L Pioro E et al Imaging of axonal damage in multiple sclerosisspatial distribution of magnetic resonance imaging lesions Ann Neurol 199741385ndash391
30 Ciccarelli O Werring DJ Barker GJ et al A study of the mechanisms of normal-appearing white matter damage in multiple sclerosis using diffusion tensor imagingevidence of Wallerian degeneration J Neurol 2003250287ndash292
31 Pokryszko-Dragan A Banaszek A Nowakowska-Kotas M et al Diffusion tensorimaging findings in the multiple sclerosis patients and their relationships to variousaspects of disability J Neurol Sci 2018391127ndash133
32 Kolasa M Hakulinen U Brander A et al Diffusion tensor imaging and disabilityprogression in multiple sclerosis a 4-year follow-up study Brain Behav 20199e01194
33 Andersen O Hildeman A Longfils M et al Diffusion tensor imaging in multiplesclerosis at different final outcomes Acta Neurol Scand 2018137165ndash173
34 Koenig KA Sakaie KE Lowe MJ et al The relationship between cognitive functionand high-resolution diffusion tensor MRI of the cingulum bundle in multiple sclerosisMult Scler 2015211794ndash1801
35 Pardini M Bonzano L Bergamino M et al Cingulum bundle alterations underliesubjective fatigue in multiple sclerosis Mult Scler 201521442ndash447
36 Sejbaek T Nielsen HH Penner N et al Dimethyl fumarate decreases neurofilamentlight chain in CSF and blood of treatment naıve relapsing MS patients J NeurolNeurosurg Psychiatry 2019901324ndash1330
37 Piehl F Kockum I Khademi M et al Plasma neurofilament light chain levels inpatients with MS switching from injectable therapies to fingolimod Mult Scler 2018241046ndash1054
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 11
DOI 101212NXI000000000000092620218 Neurol Neuroimmunol Neuroinflamm
Maija Saraste Svetlana Bezukladova Markus Matilainen et al High serum neurofilament associates with diffuse white matter damage in MS
This information is current as of December 8 2020
ServicesUpdated Information amp
httpnnneurologyorgcontent81e926fullhtmlincluding high resolution figures can be found at
References httpnnneurologyorgcontent81e926fullhtmlref-list-1
This article cites 36 articles 3 of which you can access for free at
Citations httpnnneurologyorgcontent81e926fullhtmlotherarticles
This article has been cited by 1 HighWire-hosted articles
Subspecialty Collections
httpnnneurologyorgcgicollectionmultiple_sclerosisMultiple sclerosis
httpnnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpnnneurologyorgmiscaboutxhtmlpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnnneurologyorgmiscaddirxhtmlreprintsusInformation about ordering reprints can be found online
Academy of Neurology All rights reserved Online ISSN 2332-7812Copyright copy 2020 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the AmericanPublished since April 2014 it is an open-access online-only continuous publication journal Copyright
is an official journal of the American Academy of NeurologyNeurol Neuroimmunol Neuroinflamm
Table 3 Multiple regression analyses evaluating separately the effect of DTI-MRI indices or volumetric brain MRI data onNfL in the NfL(high) subgroup of patients with MS
Variable in model DTI Sex Age Type EDSS score DMT Relapses
Entire NAWM
FA 0018 07 08 10 03 04 02
MD 0003 07 09 10 02 04 04
AD 0010 09 10 08 01 07 05
RD 0005 07 10 10 02 04 03
Frontal NAWM
FA 0045 08 10 10 02 07 02
MD 0001 05 07 10 02 06 04
AD 0004 05 09 08 03 07 04
RD 0003 06 09 09 02 06 03
Parietal NAWM
FA 0023 09 09 09 02 05 01
MD 0004 07 07 08 01 05 02
AD 0017 08 10 10 01 06 03
RD 0006 08 08 09 02 05 02
Temporal NAWM
FA 0116 07 07 10 03 05 03
MD 0007 06 10 10 01 05 05
AD 0009 07 08 08 01 05 05
RD 0020 06 09 10 02 05 04
Occipital NAWM
FA 0011 09 06 09 03 03 01
MD 0009 08 08 09 01 06 02
AD 0086 08 07 08 01 08 03
RD 0011 08 08 08 02 05 02
Cingulate NAWM
FA 0004 05 04 08 06 04 03
MD 0001 06 09 08 06 02 07
AD 0003 08 09 08 03 04 07
RD 0001 06 07 08 07 02 06
Deep NAWM
FA 0047 05 06 09 03 05 03
MD 0001 07 09 09 05 02 07
AD 0005 10 08 10 03 04 08
RD 0002 06 08 10 05 02 08
Volume (pf)
NAWM 0019 08 08 08 03 05 05
Continued
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 7
eventually manifest by brain atrophy and clinically by steadyworsening of disability in almost all patients over time This diffusepathology which in the majority of cases leads to progressiveworsening of the MS-related symptoms is now the focus of in-terest for patients and physicians as modern disease-modifyingtherapies have led to an almost complete suppression of relapseactivity while their impact on progression is at best modest Ourincomplete understanding of the molecular mechanisms that leadto progression is one of the main reasons for the failure in de-velopment ofmore efficacious therapies for secondary progressiveform of MS Another impediment is the relative insensitivity ofcurrent measurement tools to capture features of progression atsubclinical stages and hence to quantitate the effects of anti-neurodegenerative treatments at a time point where the overallbrain structure is still largely intact Specifically conventional MRIcan only quantitate brain volume loss as an end result of thedisease but does not provide insights into the underlying micro-structural changes DTI is a nonconventional MRI technique thatprovides a measure of such diffuse changes beyond volumetry
NfL is a highly sensitive biomarker to detect neuronal damageand is the first of its kind applicable in blood-derived probes22
However the cause and underlying pathology of the increasedNfL concentration may vary greatly both inter- and intra-individually In clinically isolated syndrome and RRMS acuteNfL elevation is particularly associated with focal in-flammation ie relapses and focal lesion formation4ndash823
However the indication for the contribution of diffuse pa-thology to higher NfL levels has so far been only incidentaland consisted of associations of increased NfL with moreadvanced clinical disability and brain atrophy824 Novelfindings presented here provide more direct evidence for theeffect of diffuse pathologic process in the NAWM on in-creased serum NfL We observed that among a subgroup ofolder and more disabled patients NfL associates both withradial and axial diffusivities In the context of MS these DTIscalars are thought to reflect myelin loss and axonal damagerespectively11 NfL also associated with mean diffusivity thatcan also indicate on axonal andmyelin loss11 The associationswere most pronounced in the cingulate deep and frontalNAWM
NfL differs in 2 important aspects from MRI first as asignal of neurodegeneration it reflects ongoing pathologicprocesses in real time while imaging captures morphologicfeatures that are inherently retrospective Second NfLlevels reflect neuronal damage in the CNS comprehen-sively ie in addition to brain pathology it also capturesspinal cord pathology that is not routinely evaluated byimaging in the workup of MS for individual patients As adownside NfL levels cannot differentiate betweenacute inflammatory and chronic neurodegenerative diseaseactivity Our results suggest that DTI-MRI allows us tocategorize the pathogenic source of NfL at least semi-quantitatively We demonstrate that both diffuse axonaldamage within the NAWM measured using DTI and focalinflammatory white matter lesion load measured usingconventional MRI contributed to the elevated NfL serumlevels This highlights that both focal and diffuse pathologicchanges must be taken into account when levels of NfL inindividual patients are interpreted A somewhat largerproportion of NfL variance was explained by DTI indices inaverage than by T2 lesion load This implies that in non-relapsing patients the ongoing diffuse microstructuraldamage in the NAWM outside lesions might predominateover focal axonal damage that has occurred within lesionsas a source of NfL in serum Accordingly in other neuro-logic disease with brain diffuse pathology such as amyo-trophic lateral sclerosis Alzheimer disease traumatic braininjury and age-related white matter pathology elevatedserum NfL levels have been associated with DTI-measurable diffuse damage within the CNS25ndash28 Thesestudies are further evidence for the concept that braindiffuse chronic neuropathologic mechanisms contribute toNfL release and that nonconventional MRI techniques areable to identify the CNS areas of its morphologic source
Current results demonstrate that the cingulate area is thebrain area with the most prominent white matter tractdamage outside focal lesions Periventricular regions arecharacteristically the brain area with a particularly heavylesion load in MS29 We addressed the question whetherthe cingulum-NAWM DTI abnormalities would arise from
Table 3 Multiple regression analyses evaluating separately the effect of DTI-MRI indices or volumetric brain MRI data onNfL in the NfL(high) subgroup of patients with MS (continued)
Variable in model DTI Sex Age Type EDSS score DMT Relapses
GMctx 0004 09 02 04 003 03 06
T2 0019 06 03 07 04 07 03
T1 015 07 04 10 05 07 03
Abbreviations AD = axial diffusivity DMT = disease-modifying treatment DTI = diffusion tensor imaging EDSS = Expanded Disability Status Scale FA =fractional anisotropy GMctx = cortical gray matter MD = mean diffusivity NAWM = normal-appearing white matter NfL = neurofilament light chain pf =parenchymal fraction RD = radial diffusivityPatients withMSwere divided into NfL(low) and NfL(high) subgroups based on themedian NfL value of healthy controls (231 pgmL) Model was adjusted bysex age disease type (relapsing-remittingsecondary progressive) EDSS score treatment (no first- and second-line) and presence of relapses within 1 yearbefore sampling (yesno) Bolded p values indicate that the parameter affected the logarithmofNfL value statistically significantly whenDTI or volume and allthe adjusting variables were in the models
8 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
the dirty-appearing white matter due to close proximity to aheavy lesion load but the relative lesion volume in the cingulatearea was not found to be significantly larger compared withlesion load in other regions or in the whole NAWM Thereforeit seems unlikely that differences in the DTI indices would besolely due to dirty-appearing white matter adjacent to lesionswhereas it is more likely that remote lesions lead to DTI changesin the cingulateNAWM throughWallerian degeneration30 Alsoother studies have observed diffusion changes in the cingulatearea of patients with MS31ndash33 Because the cingulum bundle offibers projects from the cingulate gyrus to the entorhinal cortexin the brain allowing for communication between components
of the limbic system the DTI abnormalities may be the mor-phologic substrate for impaired cognitive function and fatiguesymptoms that are typical in progressive MS3435
Present results are in line with previous observations on therelation between NfL concentration and MS disease sub-type with progressive patients having greater NfL levelscompared with relapsing-remitting patients568ndash102324
However the median NfL value of the entire patient cohortwas relatively low compared with healthy controls which isthe main limitation of our study Because of this we werenot able to use the previously defined cutoff values3 for
Figure 2 The effect of DTI-MRI indices or volumetric MRI data on NfL in the NfL(high) subgroup
The NfL(high) subgroup is comprised of patients with serum NfL above the median value of healthy controls (231 pgmL) Logarithmic serum NfL wasmodeled separately by DTI indices of the entire and parcellated NAWM and volumetric brain MRI data using multiple regression analyses Models wereadjusted by sex age disease type (RRMSSPMS) the EDSS score treatment (no first and second line) and presence of relapseswithin 1 year before sampling(yesno) The results are illustrated using dot and whisker plots in which red dots represent standardized regression coefficients and red lines represent theCIs of the estimates Significant p values of the DTI parameters and the percentage of variance in the response that can be explained by the independentvariables (R2) are also shown The results were ordered according to the R2 value All except 2 of the p values (fractional anisotropy of the frontal and deepNAWM) shown in the figure remained significant after adjustment using the false discovery ratemethod for the number of investigated variables (n = 32) AD =axial diffusivity DTI = diffusion tensor imaging EDSS = Expanded Disability Status Scale FA = fractional anisotropy GMctx = cortical gray matter MD =meandiffusivity NAWM = normal-appearing white matter NfL = neurofilament light chain PF = parenchymal fraction RD = radial diffusivity RRMS = relapsing-remitting MS SPMS = secondary progressive MS WM = white matter
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 9
NfL(low) and NfL(high) subgroups Instead we used themedian value of healthy controls Although the divider usedin our study was based on a small group of individualssimilar values have been observed in previous largerstudies68 One reason for the low median NfL level couldbe that the patients with RRMS are represented with arelatively benign disease course In addition our patientswith RRMS were sampled during a relapse-free time andhad no signs of ongoing focal inflammatory activity in theMRI Moreover most of the patients (68) were using adisease-modifying treatment at the time of evaluationHence similar as observed in other studies35233637 thisresulted in low serum NfL concentration Following thesimilarity in the NfL levels between healthy controls andpatients with MS the main results of our study are based ona relatively low number of patients In addition the patientnumbers under different immunomodulatory treatmentsvaried substantially However the heterogeneity of thecohort was taken account in the multiple regression modelwhere both EDSS score and disease-modifying treatmentwere used as adjustments
Our findings suggest that elevated serum NfL in MS resultsfrom neuroaxonal damage both within the NAWM and infocal T2 lesions The association between DTI-measurablediffuse microstructural white matter damage and serumNfL isfurther conceptual evidence for the latter being a usefulmonitoring tool in assessing the degree of ongoing neuro-degenerative processes in MS
AcknowledgmentThe authors thank all patients with MS and healthycontrols who participated in this study They also thank thepersonnel of Turku PET Centre for their excellenttechnical assistance
Study fundingThis work was supported by the Finnish Academy the SigridJuselius Foundation
DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgNN for full disclosures
Publication historyReceived by Neurology Neuroimmunology amp NeuroinflammationMay 8 2020 Accepted in final form October 21 2020
References1 Reich DS Lucchinetti CF Calabresi PA Multiple sclerosis N Engl J Med 2018378
169ndash1802 Varhaug KN Torkildsen Oslash Myhr KM Vedeler CA Neurofilament light chain as a
biomarker in multiple sclerosis Front Neurol 201910338
Appendix Authors
Author Location Contribution
MaijaSaraste PhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Acquisition of dataanalyzed andinterpreted the data anddrafted and revised themanuscript forintellectual content
Appendix (continued)
Author Location Contribution
SvetlanaBezukladovaMSc
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Acquisition of dataanalyzed and interpretedthe data and revised themanuscript
MarkusMatilainenPhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Statistical analysisdrafted the figures andrevised the manuscript
Jouni TuiskuMSc
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Major role in theacquisition of data andrevised the manuscript
EeroRissanenMDPhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Designed andconceptualized the studyand acquisition of data
MarcusSucksdorffMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Major role in theacquisition of data
SiniLaaksonenMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Acquisition of data
AnnaVuorimaaMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Acquisition of data
Jens KuhleMD PhD
Neurologic Clinic andPoliclinic Departments ofMedicine Biomedicineand Clinical ResearchUniversity Hospital BaselSwitzerland
Major role in theacquisition of data
DavidLeppert MDPhD
Neurologic Clinic andPoliclinic Departments ofMedicine Biomedicineand Clinical ResearchUniversity Hospital BaselSwitzerland
Revised the manuscriptfor intellectual content
Laura AirasMD PhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Designed andconceptualized the studyand revised themanuscript for intellectualcontent
10 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
3 Kuhle J Kropshofer H Haering DA et al Blood neurofilament light chain as abiomarker of MS disease activity and treatment response Neurology 201992e1007ndashe1015
4 Varhaug KN Barro C Bjoslashrnevik K et al Neurofilament light chain predicts diseaseactivity in relapsing-remitting MS Neurol Neuroimmunol Neuroinflamm 20185e422
5 Novakova L Zetterberg H Sundstrom P et al Monitoring disease activity in multiplesclerosis using serum neurofilament light protein Neurology 2017892230ndash2237
6 Disanto G Barro C Benkert P et al Serum neurofilament light a biomarker ofneuronal damage in multiple sclerosis Ann Neurol 201781857ndash870
7 Siller N Kuhle J Muthuraman M et al Serum neurofilament light chain is a bio-marker of acute and chronic neuronal damage in early multiple sclerosis Mult Scler201925678ndash686
8 Barro C Benkert P Disanto G et al Serum neurofilament as a predictor of diseaseworsening and brain and spinal cord atrophy in multiple sclerosis Brain 20181412382ndash2391
9 Ferraro D Guicciardi C De Biasi S et al Plasma neurofilaments correlate withdisability in progressive multiple sclerosis patients Acta Neurol Scand 202014116ndash21
10 Jakimovski D Zivadinov R RamanthanM et al Serum neurofilament light chain levelassociations with clinical and cognitive performance in multiple sclerosis a longitu-dinal retrospective 5-year study Mult Scler 2020261670ndash1681
11 Sbardella E Tona F Petsas N Pantano P DTI measurements in multiple sclerosisevaluation of brain damage and clinical implications Mult Scler Int 20132013671730
12 Tae WS Ham BJ Pyun SB Kang SH Kim BJ Current clinical applications ofdiffusion-tensor imaging in neurological disorders J Clin Neurol 201814129ndash140
13 Kurtzke JF Rating neurologic impairment in multiple sclerosis an expanded disabilitystatus scale (EDSS) Neurology 1983331444ndash1452
14 Kuhle J Barro C Andreasson U et al Comparison of three analytical platforms forquantification of the neurofilament light chain in blood samples ELISA electro-chemiluminescence immunoassay and SimoaClinChemLabMed 2016541655ndash1661
15 Bezukladova S Tuisku J Matilainen M et al Insights into disseminated MS brainpathology with multimodal diffusion tensor and PET imaging Neurol NeuroimmunolNeuroinflamm 20207e691
16 Schmidt P Gaser C Arsic M et al An automated tool for detection of FLAIR-hyperintense white-matter lesions in Multiple Sclerosis Neuroimage 2012593774ndash3783
17 Salat DH Greve DN Pacheco JL et al Regional white matter volume differences innondemented aging and Alzheimerrsquos disease Neuroimage 2009441247ndash1258
18 Rissanen E Tuisku J Vahlberg T et al Microglial activation white matter tractdamage and disability in MS Neurol Neuroimmunol Neuroinflamm 20185e443
19 Leemans A Jeurissen B Sijbers J et al ExploreDTI a graphical toolbox for pro-cessing analysing and visualizing diffusion MR data Presented at the 17th AnnualMeeting of International Society for Magnetic Resonance in Medicine Honolulu HI2009 p 3537
20 Benjamini Y Hochberg Y Controlling the false discovery rate a practical and pow-erful approach to multiple testing J R Stat Soc Ser B 199557289ndash300
21 Gelman A Scaling regression inputs by dividing by two standard deviations Stat Med2008272865ndash2873
22 Khalil M Teunissen CE Otto M et al Neurofilaments as biomarkers in neurologicaldisorders Nat Rev Neurol 201814577ndash589
23 Canto E Barro C Zhao C et al Association between serum neurofilament light chainlevels and long-term disease course among patients with multiple sclerosis followedup for 12 years JAMA Neurol 2019761359ndash1366
24 Hogel H Rissanen E Barro C et al Serum glial fibrillary acidic protein correlates withmultiple sclerosis disease severity Mult Scler 202026210ndash219
25 Menke RA Gray E Lu CH et al CSF neurofilament light chain reflects corticospinaltract degeneration in ALS Ann Clin Transl Neurol 20152748ndash755
26 Moore EE Hohman TJ Badami FS et al Neurofilament relates to white mattermicrostructure in older adults Neurobiol Aging 201870233ndash241
27 Ljungqvist J Zetterberg H Mitsis M Blennow K Skoglund T Serum neurofilamentlight protein as a marker for diffuse axonal injury results from a case series studyJ Neurotrauma 2017341124ndash1127
28 Schultz SA Strain JF Adedokun A et al Serum neurofilament light chain levels areassociated with white matter integrity in autosomal dominant Alzheimerrsquos diseaseNeurobiol Dis 2020142104960
29 Narayanan S Fu L Pioro E et al Imaging of axonal damage in multiple sclerosisspatial distribution of magnetic resonance imaging lesions Ann Neurol 199741385ndash391
30 Ciccarelli O Werring DJ Barker GJ et al A study of the mechanisms of normal-appearing white matter damage in multiple sclerosis using diffusion tensor imagingevidence of Wallerian degeneration J Neurol 2003250287ndash292
31 Pokryszko-Dragan A Banaszek A Nowakowska-Kotas M et al Diffusion tensorimaging findings in the multiple sclerosis patients and their relationships to variousaspects of disability J Neurol Sci 2018391127ndash133
32 Kolasa M Hakulinen U Brander A et al Diffusion tensor imaging and disabilityprogression in multiple sclerosis a 4-year follow-up study Brain Behav 20199e01194
33 Andersen O Hildeman A Longfils M et al Diffusion tensor imaging in multiplesclerosis at different final outcomes Acta Neurol Scand 2018137165ndash173
34 Koenig KA Sakaie KE Lowe MJ et al The relationship between cognitive functionand high-resolution diffusion tensor MRI of the cingulum bundle in multiple sclerosisMult Scler 2015211794ndash1801
35 Pardini M Bonzano L Bergamino M et al Cingulum bundle alterations underliesubjective fatigue in multiple sclerosis Mult Scler 201521442ndash447
36 Sejbaek T Nielsen HH Penner N et al Dimethyl fumarate decreases neurofilamentlight chain in CSF and blood of treatment naıve relapsing MS patients J NeurolNeurosurg Psychiatry 2019901324ndash1330
37 Piehl F Kockum I Khademi M et al Plasma neurofilament light chain levels inpatients with MS switching from injectable therapies to fingolimod Mult Scler 2018241046ndash1054
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 11
DOI 101212NXI000000000000092620218 Neurol Neuroimmunol Neuroinflamm
Maija Saraste Svetlana Bezukladova Markus Matilainen et al High serum neurofilament associates with diffuse white matter damage in MS
This information is current as of December 8 2020
ServicesUpdated Information amp
httpnnneurologyorgcontent81e926fullhtmlincluding high resolution figures can be found at
References httpnnneurologyorgcontent81e926fullhtmlref-list-1
This article cites 36 articles 3 of which you can access for free at
Citations httpnnneurologyorgcontent81e926fullhtmlotherarticles
This article has been cited by 1 HighWire-hosted articles
Subspecialty Collections
httpnnneurologyorgcgicollectionmultiple_sclerosisMultiple sclerosis
httpnnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpnnneurologyorgmiscaboutxhtmlpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnnneurologyorgmiscaddirxhtmlreprintsusInformation about ordering reprints can be found online
Academy of Neurology All rights reserved Online ISSN 2332-7812Copyright copy 2020 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the AmericanPublished since April 2014 it is an open-access online-only continuous publication journal Copyright
is an official journal of the American Academy of NeurologyNeurol Neuroimmunol Neuroinflamm
eventually manifest by brain atrophy and clinically by steadyworsening of disability in almost all patients over time This diffusepathology which in the majority of cases leads to progressiveworsening of the MS-related symptoms is now the focus of in-terest for patients and physicians as modern disease-modifyingtherapies have led to an almost complete suppression of relapseactivity while their impact on progression is at best modest Ourincomplete understanding of the molecular mechanisms that leadto progression is one of the main reasons for the failure in de-velopment ofmore efficacious therapies for secondary progressiveform of MS Another impediment is the relative insensitivity ofcurrent measurement tools to capture features of progression atsubclinical stages and hence to quantitate the effects of anti-neurodegenerative treatments at a time point where the overallbrain structure is still largely intact Specifically conventional MRIcan only quantitate brain volume loss as an end result of thedisease but does not provide insights into the underlying micro-structural changes DTI is a nonconventional MRI technique thatprovides a measure of such diffuse changes beyond volumetry
NfL is a highly sensitive biomarker to detect neuronal damageand is the first of its kind applicable in blood-derived probes22
However the cause and underlying pathology of the increasedNfL concentration may vary greatly both inter- and intra-individually In clinically isolated syndrome and RRMS acuteNfL elevation is particularly associated with focal in-flammation ie relapses and focal lesion formation4ndash823
However the indication for the contribution of diffuse pa-thology to higher NfL levels has so far been only incidentaland consisted of associations of increased NfL with moreadvanced clinical disability and brain atrophy824 Novelfindings presented here provide more direct evidence for theeffect of diffuse pathologic process in the NAWM on in-creased serum NfL We observed that among a subgroup ofolder and more disabled patients NfL associates both withradial and axial diffusivities In the context of MS these DTIscalars are thought to reflect myelin loss and axonal damagerespectively11 NfL also associated with mean diffusivity thatcan also indicate on axonal andmyelin loss11 The associationswere most pronounced in the cingulate deep and frontalNAWM
NfL differs in 2 important aspects from MRI first as asignal of neurodegeneration it reflects ongoing pathologicprocesses in real time while imaging captures morphologicfeatures that are inherently retrospective Second NfLlevels reflect neuronal damage in the CNS comprehen-sively ie in addition to brain pathology it also capturesspinal cord pathology that is not routinely evaluated byimaging in the workup of MS for individual patients As adownside NfL levels cannot differentiate betweenacute inflammatory and chronic neurodegenerative diseaseactivity Our results suggest that DTI-MRI allows us tocategorize the pathogenic source of NfL at least semi-quantitatively We demonstrate that both diffuse axonaldamage within the NAWM measured using DTI and focalinflammatory white matter lesion load measured usingconventional MRI contributed to the elevated NfL serumlevels This highlights that both focal and diffuse pathologicchanges must be taken into account when levels of NfL inindividual patients are interpreted A somewhat largerproportion of NfL variance was explained by DTI indices inaverage than by T2 lesion load This implies that in non-relapsing patients the ongoing diffuse microstructuraldamage in the NAWM outside lesions might predominateover focal axonal damage that has occurred within lesionsas a source of NfL in serum Accordingly in other neuro-logic disease with brain diffuse pathology such as amyo-trophic lateral sclerosis Alzheimer disease traumatic braininjury and age-related white matter pathology elevatedserum NfL levels have been associated with DTI-measurable diffuse damage within the CNS25ndash28 Thesestudies are further evidence for the concept that braindiffuse chronic neuropathologic mechanisms contribute toNfL release and that nonconventional MRI techniques areable to identify the CNS areas of its morphologic source
Current results demonstrate that the cingulate area is thebrain area with the most prominent white matter tractdamage outside focal lesions Periventricular regions arecharacteristically the brain area with a particularly heavylesion load in MS29 We addressed the question whetherthe cingulum-NAWM DTI abnormalities would arise from
Table 3 Multiple regression analyses evaluating separately the effect of DTI-MRI indices or volumetric brain MRI data onNfL in the NfL(high) subgroup of patients with MS (continued)
Variable in model DTI Sex Age Type EDSS score DMT Relapses
GMctx 0004 09 02 04 003 03 06
T2 0019 06 03 07 04 07 03
T1 015 07 04 10 05 07 03
Abbreviations AD = axial diffusivity DMT = disease-modifying treatment DTI = diffusion tensor imaging EDSS = Expanded Disability Status Scale FA =fractional anisotropy GMctx = cortical gray matter MD = mean diffusivity NAWM = normal-appearing white matter NfL = neurofilament light chain pf =parenchymal fraction RD = radial diffusivityPatients withMSwere divided into NfL(low) and NfL(high) subgroups based on themedian NfL value of healthy controls (231 pgmL) Model was adjusted bysex age disease type (relapsing-remittingsecondary progressive) EDSS score treatment (no first- and second-line) and presence of relapses within 1 yearbefore sampling (yesno) Bolded p values indicate that the parameter affected the logarithmofNfL value statistically significantly whenDTI or volume and allthe adjusting variables were in the models
8 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
the dirty-appearing white matter due to close proximity to aheavy lesion load but the relative lesion volume in the cingulatearea was not found to be significantly larger compared withlesion load in other regions or in the whole NAWM Thereforeit seems unlikely that differences in the DTI indices would besolely due to dirty-appearing white matter adjacent to lesionswhereas it is more likely that remote lesions lead to DTI changesin the cingulateNAWM throughWallerian degeneration30 Alsoother studies have observed diffusion changes in the cingulatearea of patients with MS31ndash33 Because the cingulum bundle offibers projects from the cingulate gyrus to the entorhinal cortexin the brain allowing for communication between components
of the limbic system the DTI abnormalities may be the mor-phologic substrate for impaired cognitive function and fatiguesymptoms that are typical in progressive MS3435
Present results are in line with previous observations on therelation between NfL concentration and MS disease sub-type with progressive patients having greater NfL levelscompared with relapsing-remitting patients568ndash102324
However the median NfL value of the entire patient cohortwas relatively low compared with healthy controls which isthe main limitation of our study Because of this we werenot able to use the previously defined cutoff values3 for
Figure 2 The effect of DTI-MRI indices or volumetric MRI data on NfL in the NfL(high) subgroup
The NfL(high) subgroup is comprised of patients with serum NfL above the median value of healthy controls (231 pgmL) Logarithmic serum NfL wasmodeled separately by DTI indices of the entire and parcellated NAWM and volumetric brain MRI data using multiple regression analyses Models wereadjusted by sex age disease type (RRMSSPMS) the EDSS score treatment (no first and second line) and presence of relapseswithin 1 year before sampling(yesno) The results are illustrated using dot and whisker plots in which red dots represent standardized regression coefficients and red lines represent theCIs of the estimates Significant p values of the DTI parameters and the percentage of variance in the response that can be explained by the independentvariables (R2) are also shown The results were ordered according to the R2 value All except 2 of the p values (fractional anisotropy of the frontal and deepNAWM) shown in the figure remained significant after adjustment using the false discovery ratemethod for the number of investigated variables (n = 32) AD =axial diffusivity DTI = diffusion tensor imaging EDSS = Expanded Disability Status Scale FA = fractional anisotropy GMctx = cortical gray matter MD =meandiffusivity NAWM = normal-appearing white matter NfL = neurofilament light chain PF = parenchymal fraction RD = radial diffusivity RRMS = relapsing-remitting MS SPMS = secondary progressive MS WM = white matter
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 9
NfL(low) and NfL(high) subgroups Instead we used themedian value of healthy controls Although the divider usedin our study was based on a small group of individualssimilar values have been observed in previous largerstudies68 One reason for the low median NfL level couldbe that the patients with RRMS are represented with arelatively benign disease course In addition our patientswith RRMS were sampled during a relapse-free time andhad no signs of ongoing focal inflammatory activity in theMRI Moreover most of the patients (68) were using adisease-modifying treatment at the time of evaluationHence similar as observed in other studies35233637 thisresulted in low serum NfL concentration Following thesimilarity in the NfL levels between healthy controls andpatients with MS the main results of our study are based ona relatively low number of patients In addition the patientnumbers under different immunomodulatory treatmentsvaried substantially However the heterogeneity of thecohort was taken account in the multiple regression modelwhere both EDSS score and disease-modifying treatmentwere used as adjustments
Our findings suggest that elevated serum NfL in MS resultsfrom neuroaxonal damage both within the NAWM and infocal T2 lesions The association between DTI-measurablediffuse microstructural white matter damage and serumNfL isfurther conceptual evidence for the latter being a usefulmonitoring tool in assessing the degree of ongoing neuro-degenerative processes in MS
AcknowledgmentThe authors thank all patients with MS and healthycontrols who participated in this study They also thank thepersonnel of Turku PET Centre for their excellenttechnical assistance
Study fundingThis work was supported by the Finnish Academy the SigridJuselius Foundation
DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgNN for full disclosures
Publication historyReceived by Neurology Neuroimmunology amp NeuroinflammationMay 8 2020 Accepted in final form October 21 2020
References1 Reich DS Lucchinetti CF Calabresi PA Multiple sclerosis N Engl J Med 2018378
169ndash1802 Varhaug KN Torkildsen Oslash Myhr KM Vedeler CA Neurofilament light chain as a
biomarker in multiple sclerosis Front Neurol 201910338
Appendix Authors
Author Location Contribution
MaijaSaraste PhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Acquisition of dataanalyzed andinterpreted the data anddrafted and revised themanuscript forintellectual content
Appendix (continued)
Author Location Contribution
SvetlanaBezukladovaMSc
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Acquisition of dataanalyzed and interpretedthe data and revised themanuscript
MarkusMatilainenPhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Statistical analysisdrafted the figures andrevised the manuscript
Jouni TuiskuMSc
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Major role in theacquisition of data andrevised the manuscript
EeroRissanenMDPhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Designed andconceptualized the studyand acquisition of data
MarcusSucksdorffMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Major role in theacquisition of data
SiniLaaksonenMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Acquisition of data
AnnaVuorimaaMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Acquisition of data
Jens KuhleMD PhD
Neurologic Clinic andPoliclinic Departments ofMedicine Biomedicineand Clinical ResearchUniversity Hospital BaselSwitzerland
Major role in theacquisition of data
DavidLeppert MDPhD
Neurologic Clinic andPoliclinic Departments ofMedicine Biomedicineand Clinical ResearchUniversity Hospital BaselSwitzerland
Revised the manuscriptfor intellectual content
Laura AirasMD PhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Designed andconceptualized the studyand revised themanuscript for intellectualcontent
10 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
3 Kuhle J Kropshofer H Haering DA et al Blood neurofilament light chain as abiomarker of MS disease activity and treatment response Neurology 201992e1007ndashe1015
4 Varhaug KN Barro C Bjoslashrnevik K et al Neurofilament light chain predicts diseaseactivity in relapsing-remitting MS Neurol Neuroimmunol Neuroinflamm 20185e422
5 Novakova L Zetterberg H Sundstrom P et al Monitoring disease activity in multiplesclerosis using serum neurofilament light protein Neurology 2017892230ndash2237
6 Disanto G Barro C Benkert P et al Serum neurofilament light a biomarker ofneuronal damage in multiple sclerosis Ann Neurol 201781857ndash870
7 Siller N Kuhle J Muthuraman M et al Serum neurofilament light chain is a bio-marker of acute and chronic neuronal damage in early multiple sclerosis Mult Scler201925678ndash686
8 Barro C Benkert P Disanto G et al Serum neurofilament as a predictor of diseaseworsening and brain and spinal cord atrophy in multiple sclerosis Brain 20181412382ndash2391
9 Ferraro D Guicciardi C De Biasi S et al Plasma neurofilaments correlate withdisability in progressive multiple sclerosis patients Acta Neurol Scand 202014116ndash21
10 Jakimovski D Zivadinov R RamanthanM et al Serum neurofilament light chain levelassociations with clinical and cognitive performance in multiple sclerosis a longitu-dinal retrospective 5-year study Mult Scler 2020261670ndash1681
11 Sbardella E Tona F Petsas N Pantano P DTI measurements in multiple sclerosisevaluation of brain damage and clinical implications Mult Scler Int 20132013671730
12 Tae WS Ham BJ Pyun SB Kang SH Kim BJ Current clinical applications ofdiffusion-tensor imaging in neurological disorders J Clin Neurol 201814129ndash140
13 Kurtzke JF Rating neurologic impairment in multiple sclerosis an expanded disabilitystatus scale (EDSS) Neurology 1983331444ndash1452
14 Kuhle J Barro C Andreasson U et al Comparison of three analytical platforms forquantification of the neurofilament light chain in blood samples ELISA electro-chemiluminescence immunoassay and SimoaClinChemLabMed 2016541655ndash1661
15 Bezukladova S Tuisku J Matilainen M et al Insights into disseminated MS brainpathology with multimodal diffusion tensor and PET imaging Neurol NeuroimmunolNeuroinflamm 20207e691
16 Schmidt P Gaser C Arsic M et al An automated tool for detection of FLAIR-hyperintense white-matter lesions in Multiple Sclerosis Neuroimage 2012593774ndash3783
17 Salat DH Greve DN Pacheco JL et al Regional white matter volume differences innondemented aging and Alzheimerrsquos disease Neuroimage 2009441247ndash1258
18 Rissanen E Tuisku J Vahlberg T et al Microglial activation white matter tractdamage and disability in MS Neurol Neuroimmunol Neuroinflamm 20185e443
19 Leemans A Jeurissen B Sijbers J et al ExploreDTI a graphical toolbox for pro-cessing analysing and visualizing diffusion MR data Presented at the 17th AnnualMeeting of International Society for Magnetic Resonance in Medicine Honolulu HI2009 p 3537
20 Benjamini Y Hochberg Y Controlling the false discovery rate a practical and pow-erful approach to multiple testing J R Stat Soc Ser B 199557289ndash300
21 Gelman A Scaling regression inputs by dividing by two standard deviations Stat Med2008272865ndash2873
22 Khalil M Teunissen CE Otto M et al Neurofilaments as biomarkers in neurologicaldisorders Nat Rev Neurol 201814577ndash589
23 Canto E Barro C Zhao C et al Association between serum neurofilament light chainlevels and long-term disease course among patients with multiple sclerosis followedup for 12 years JAMA Neurol 2019761359ndash1366
24 Hogel H Rissanen E Barro C et al Serum glial fibrillary acidic protein correlates withmultiple sclerosis disease severity Mult Scler 202026210ndash219
25 Menke RA Gray E Lu CH et al CSF neurofilament light chain reflects corticospinaltract degeneration in ALS Ann Clin Transl Neurol 20152748ndash755
26 Moore EE Hohman TJ Badami FS et al Neurofilament relates to white mattermicrostructure in older adults Neurobiol Aging 201870233ndash241
27 Ljungqvist J Zetterberg H Mitsis M Blennow K Skoglund T Serum neurofilamentlight protein as a marker for diffuse axonal injury results from a case series studyJ Neurotrauma 2017341124ndash1127
28 Schultz SA Strain JF Adedokun A et al Serum neurofilament light chain levels areassociated with white matter integrity in autosomal dominant Alzheimerrsquos diseaseNeurobiol Dis 2020142104960
29 Narayanan S Fu L Pioro E et al Imaging of axonal damage in multiple sclerosisspatial distribution of magnetic resonance imaging lesions Ann Neurol 199741385ndash391
30 Ciccarelli O Werring DJ Barker GJ et al A study of the mechanisms of normal-appearing white matter damage in multiple sclerosis using diffusion tensor imagingevidence of Wallerian degeneration J Neurol 2003250287ndash292
31 Pokryszko-Dragan A Banaszek A Nowakowska-Kotas M et al Diffusion tensorimaging findings in the multiple sclerosis patients and their relationships to variousaspects of disability J Neurol Sci 2018391127ndash133
32 Kolasa M Hakulinen U Brander A et al Diffusion tensor imaging and disabilityprogression in multiple sclerosis a 4-year follow-up study Brain Behav 20199e01194
33 Andersen O Hildeman A Longfils M et al Diffusion tensor imaging in multiplesclerosis at different final outcomes Acta Neurol Scand 2018137165ndash173
34 Koenig KA Sakaie KE Lowe MJ et al The relationship between cognitive functionand high-resolution diffusion tensor MRI of the cingulum bundle in multiple sclerosisMult Scler 2015211794ndash1801
35 Pardini M Bonzano L Bergamino M et al Cingulum bundle alterations underliesubjective fatigue in multiple sclerosis Mult Scler 201521442ndash447
36 Sejbaek T Nielsen HH Penner N et al Dimethyl fumarate decreases neurofilamentlight chain in CSF and blood of treatment naıve relapsing MS patients J NeurolNeurosurg Psychiatry 2019901324ndash1330
37 Piehl F Kockum I Khademi M et al Plasma neurofilament light chain levels inpatients with MS switching from injectable therapies to fingolimod Mult Scler 2018241046ndash1054
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 11
DOI 101212NXI000000000000092620218 Neurol Neuroimmunol Neuroinflamm
Maija Saraste Svetlana Bezukladova Markus Matilainen et al High serum neurofilament associates with diffuse white matter damage in MS
This information is current as of December 8 2020
ServicesUpdated Information amp
httpnnneurologyorgcontent81e926fullhtmlincluding high resolution figures can be found at
References httpnnneurologyorgcontent81e926fullhtmlref-list-1
This article cites 36 articles 3 of which you can access for free at
Citations httpnnneurologyorgcontent81e926fullhtmlotherarticles
This article has been cited by 1 HighWire-hosted articles
Subspecialty Collections
httpnnneurologyorgcgicollectionmultiple_sclerosisMultiple sclerosis
httpnnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpnnneurologyorgmiscaboutxhtmlpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnnneurologyorgmiscaddirxhtmlreprintsusInformation about ordering reprints can be found online
Academy of Neurology All rights reserved Online ISSN 2332-7812Copyright copy 2020 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the AmericanPublished since April 2014 it is an open-access online-only continuous publication journal Copyright
is an official journal of the American Academy of NeurologyNeurol Neuroimmunol Neuroinflamm
the dirty-appearing white matter due to close proximity to aheavy lesion load but the relative lesion volume in the cingulatearea was not found to be significantly larger compared withlesion load in other regions or in the whole NAWM Thereforeit seems unlikely that differences in the DTI indices would besolely due to dirty-appearing white matter adjacent to lesionswhereas it is more likely that remote lesions lead to DTI changesin the cingulateNAWM throughWallerian degeneration30 Alsoother studies have observed diffusion changes in the cingulatearea of patients with MS31ndash33 Because the cingulum bundle offibers projects from the cingulate gyrus to the entorhinal cortexin the brain allowing for communication between components
of the limbic system the DTI abnormalities may be the mor-phologic substrate for impaired cognitive function and fatiguesymptoms that are typical in progressive MS3435
Present results are in line with previous observations on therelation between NfL concentration and MS disease sub-type with progressive patients having greater NfL levelscompared with relapsing-remitting patients568ndash102324
However the median NfL value of the entire patient cohortwas relatively low compared with healthy controls which isthe main limitation of our study Because of this we werenot able to use the previously defined cutoff values3 for
Figure 2 The effect of DTI-MRI indices or volumetric MRI data on NfL in the NfL(high) subgroup
The NfL(high) subgroup is comprised of patients with serum NfL above the median value of healthy controls (231 pgmL) Logarithmic serum NfL wasmodeled separately by DTI indices of the entire and parcellated NAWM and volumetric brain MRI data using multiple regression analyses Models wereadjusted by sex age disease type (RRMSSPMS) the EDSS score treatment (no first and second line) and presence of relapseswithin 1 year before sampling(yesno) The results are illustrated using dot and whisker plots in which red dots represent standardized regression coefficients and red lines represent theCIs of the estimates Significant p values of the DTI parameters and the percentage of variance in the response that can be explained by the independentvariables (R2) are also shown The results were ordered according to the R2 value All except 2 of the p values (fractional anisotropy of the frontal and deepNAWM) shown in the figure remained significant after adjustment using the false discovery ratemethod for the number of investigated variables (n = 32) AD =axial diffusivity DTI = diffusion tensor imaging EDSS = Expanded Disability Status Scale FA = fractional anisotropy GMctx = cortical gray matter MD =meandiffusivity NAWM = normal-appearing white matter NfL = neurofilament light chain PF = parenchymal fraction RD = radial diffusivity RRMS = relapsing-remitting MS SPMS = secondary progressive MS WM = white matter
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 9
NfL(low) and NfL(high) subgroups Instead we used themedian value of healthy controls Although the divider usedin our study was based on a small group of individualssimilar values have been observed in previous largerstudies68 One reason for the low median NfL level couldbe that the patients with RRMS are represented with arelatively benign disease course In addition our patientswith RRMS were sampled during a relapse-free time andhad no signs of ongoing focal inflammatory activity in theMRI Moreover most of the patients (68) were using adisease-modifying treatment at the time of evaluationHence similar as observed in other studies35233637 thisresulted in low serum NfL concentration Following thesimilarity in the NfL levels between healthy controls andpatients with MS the main results of our study are based ona relatively low number of patients In addition the patientnumbers under different immunomodulatory treatmentsvaried substantially However the heterogeneity of thecohort was taken account in the multiple regression modelwhere both EDSS score and disease-modifying treatmentwere used as adjustments
Our findings suggest that elevated serum NfL in MS resultsfrom neuroaxonal damage both within the NAWM and infocal T2 lesions The association between DTI-measurablediffuse microstructural white matter damage and serumNfL isfurther conceptual evidence for the latter being a usefulmonitoring tool in assessing the degree of ongoing neuro-degenerative processes in MS
AcknowledgmentThe authors thank all patients with MS and healthycontrols who participated in this study They also thank thepersonnel of Turku PET Centre for their excellenttechnical assistance
Study fundingThis work was supported by the Finnish Academy the SigridJuselius Foundation
DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgNN for full disclosures
Publication historyReceived by Neurology Neuroimmunology amp NeuroinflammationMay 8 2020 Accepted in final form October 21 2020
References1 Reich DS Lucchinetti CF Calabresi PA Multiple sclerosis N Engl J Med 2018378
169ndash1802 Varhaug KN Torkildsen Oslash Myhr KM Vedeler CA Neurofilament light chain as a
biomarker in multiple sclerosis Front Neurol 201910338
Appendix Authors
Author Location Contribution
MaijaSaraste PhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Acquisition of dataanalyzed andinterpreted the data anddrafted and revised themanuscript forintellectual content
Appendix (continued)
Author Location Contribution
SvetlanaBezukladovaMSc
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Acquisition of dataanalyzed and interpretedthe data and revised themanuscript
MarkusMatilainenPhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Statistical analysisdrafted the figures andrevised the manuscript
Jouni TuiskuMSc
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Major role in theacquisition of data andrevised the manuscript
EeroRissanenMDPhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Designed andconceptualized the studyand acquisition of data
MarcusSucksdorffMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Major role in theacquisition of data
SiniLaaksonenMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Acquisition of data
AnnaVuorimaaMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Acquisition of data
Jens KuhleMD PhD
Neurologic Clinic andPoliclinic Departments ofMedicine Biomedicineand Clinical ResearchUniversity Hospital BaselSwitzerland
Major role in theacquisition of data
DavidLeppert MDPhD
Neurologic Clinic andPoliclinic Departments ofMedicine Biomedicineand Clinical ResearchUniversity Hospital BaselSwitzerland
Revised the manuscriptfor intellectual content
Laura AirasMD PhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Designed andconceptualized the studyand revised themanuscript for intellectualcontent
10 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
3 Kuhle J Kropshofer H Haering DA et al Blood neurofilament light chain as abiomarker of MS disease activity and treatment response Neurology 201992e1007ndashe1015
4 Varhaug KN Barro C Bjoslashrnevik K et al Neurofilament light chain predicts diseaseactivity in relapsing-remitting MS Neurol Neuroimmunol Neuroinflamm 20185e422
5 Novakova L Zetterberg H Sundstrom P et al Monitoring disease activity in multiplesclerosis using serum neurofilament light protein Neurology 2017892230ndash2237
6 Disanto G Barro C Benkert P et al Serum neurofilament light a biomarker ofneuronal damage in multiple sclerosis Ann Neurol 201781857ndash870
7 Siller N Kuhle J Muthuraman M et al Serum neurofilament light chain is a bio-marker of acute and chronic neuronal damage in early multiple sclerosis Mult Scler201925678ndash686
8 Barro C Benkert P Disanto G et al Serum neurofilament as a predictor of diseaseworsening and brain and spinal cord atrophy in multiple sclerosis Brain 20181412382ndash2391
9 Ferraro D Guicciardi C De Biasi S et al Plasma neurofilaments correlate withdisability in progressive multiple sclerosis patients Acta Neurol Scand 202014116ndash21
10 Jakimovski D Zivadinov R RamanthanM et al Serum neurofilament light chain levelassociations with clinical and cognitive performance in multiple sclerosis a longitu-dinal retrospective 5-year study Mult Scler 2020261670ndash1681
11 Sbardella E Tona F Petsas N Pantano P DTI measurements in multiple sclerosisevaluation of brain damage and clinical implications Mult Scler Int 20132013671730
12 Tae WS Ham BJ Pyun SB Kang SH Kim BJ Current clinical applications ofdiffusion-tensor imaging in neurological disorders J Clin Neurol 201814129ndash140
13 Kurtzke JF Rating neurologic impairment in multiple sclerosis an expanded disabilitystatus scale (EDSS) Neurology 1983331444ndash1452
14 Kuhle J Barro C Andreasson U et al Comparison of three analytical platforms forquantification of the neurofilament light chain in blood samples ELISA electro-chemiluminescence immunoassay and SimoaClinChemLabMed 2016541655ndash1661
15 Bezukladova S Tuisku J Matilainen M et al Insights into disseminated MS brainpathology with multimodal diffusion tensor and PET imaging Neurol NeuroimmunolNeuroinflamm 20207e691
16 Schmidt P Gaser C Arsic M et al An automated tool for detection of FLAIR-hyperintense white-matter lesions in Multiple Sclerosis Neuroimage 2012593774ndash3783
17 Salat DH Greve DN Pacheco JL et al Regional white matter volume differences innondemented aging and Alzheimerrsquos disease Neuroimage 2009441247ndash1258
18 Rissanen E Tuisku J Vahlberg T et al Microglial activation white matter tractdamage and disability in MS Neurol Neuroimmunol Neuroinflamm 20185e443
19 Leemans A Jeurissen B Sijbers J et al ExploreDTI a graphical toolbox for pro-cessing analysing and visualizing diffusion MR data Presented at the 17th AnnualMeeting of International Society for Magnetic Resonance in Medicine Honolulu HI2009 p 3537
20 Benjamini Y Hochberg Y Controlling the false discovery rate a practical and pow-erful approach to multiple testing J R Stat Soc Ser B 199557289ndash300
21 Gelman A Scaling regression inputs by dividing by two standard deviations Stat Med2008272865ndash2873
22 Khalil M Teunissen CE Otto M et al Neurofilaments as biomarkers in neurologicaldisorders Nat Rev Neurol 201814577ndash589
23 Canto E Barro C Zhao C et al Association between serum neurofilament light chainlevels and long-term disease course among patients with multiple sclerosis followedup for 12 years JAMA Neurol 2019761359ndash1366
24 Hogel H Rissanen E Barro C et al Serum glial fibrillary acidic protein correlates withmultiple sclerosis disease severity Mult Scler 202026210ndash219
25 Menke RA Gray E Lu CH et al CSF neurofilament light chain reflects corticospinaltract degeneration in ALS Ann Clin Transl Neurol 20152748ndash755
26 Moore EE Hohman TJ Badami FS et al Neurofilament relates to white mattermicrostructure in older adults Neurobiol Aging 201870233ndash241
27 Ljungqvist J Zetterberg H Mitsis M Blennow K Skoglund T Serum neurofilamentlight protein as a marker for diffuse axonal injury results from a case series studyJ Neurotrauma 2017341124ndash1127
28 Schultz SA Strain JF Adedokun A et al Serum neurofilament light chain levels areassociated with white matter integrity in autosomal dominant Alzheimerrsquos diseaseNeurobiol Dis 2020142104960
29 Narayanan S Fu L Pioro E et al Imaging of axonal damage in multiple sclerosisspatial distribution of magnetic resonance imaging lesions Ann Neurol 199741385ndash391
30 Ciccarelli O Werring DJ Barker GJ et al A study of the mechanisms of normal-appearing white matter damage in multiple sclerosis using diffusion tensor imagingevidence of Wallerian degeneration J Neurol 2003250287ndash292
31 Pokryszko-Dragan A Banaszek A Nowakowska-Kotas M et al Diffusion tensorimaging findings in the multiple sclerosis patients and their relationships to variousaspects of disability J Neurol Sci 2018391127ndash133
32 Kolasa M Hakulinen U Brander A et al Diffusion tensor imaging and disabilityprogression in multiple sclerosis a 4-year follow-up study Brain Behav 20199e01194
33 Andersen O Hildeman A Longfils M et al Diffusion tensor imaging in multiplesclerosis at different final outcomes Acta Neurol Scand 2018137165ndash173
34 Koenig KA Sakaie KE Lowe MJ et al The relationship between cognitive functionand high-resolution diffusion tensor MRI of the cingulum bundle in multiple sclerosisMult Scler 2015211794ndash1801
35 Pardini M Bonzano L Bergamino M et al Cingulum bundle alterations underliesubjective fatigue in multiple sclerosis Mult Scler 201521442ndash447
36 Sejbaek T Nielsen HH Penner N et al Dimethyl fumarate decreases neurofilamentlight chain in CSF and blood of treatment naıve relapsing MS patients J NeurolNeurosurg Psychiatry 2019901324ndash1330
37 Piehl F Kockum I Khademi M et al Plasma neurofilament light chain levels inpatients with MS switching from injectable therapies to fingolimod Mult Scler 2018241046ndash1054
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 11
DOI 101212NXI000000000000092620218 Neurol Neuroimmunol Neuroinflamm
Maija Saraste Svetlana Bezukladova Markus Matilainen et al High serum neurofilament associates with diffuse white matter damage in MS
This information is current as of December 8 2020
ServicesUpdated Information amp
httpnnneurologyorgcontent81e926fullhtmlincluding high resolution figures can be found at
References httpnnneurologyorgcontent81e926fullhtmlref-list-1
This article cites 36 articles 3 of which you can access for free at
Citations httpnnneurologyorgcontent81e926fullhtmlotherarticles
This article has been cited by 1 HighWire-hosted articles
Subspecialty Collections
httpnnneurologyorgcgicollectionmultiple_sclerosisMultiple sclerosis
httpnnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpnnneurologyorgmiscaboutxhtmlpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnnneurologyorgmiscaddirxhtmlreprintsusInformation about ordering reprints can be found online
Academy of Neurology All rights reserved Online ISSN 2332-7812Copyright copy 2020 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the AmericanPublished since April 2014 it is an open-access online-only continuous publication journal Copyright
is an official journal of the American Academy of NeurologyNeurol Neuroimmunol Neuroinflamm
NfL(low) and NfL(high) subgroups Instead we used themedian value of healthy controls Although the divider usedin our study was based on a small group of individualssimilar values have been observed in previous largerstudies68 One reason for the low median NfL level couldbe that the patients with RRMS are represented with arelatively benign disease course In addition our patientswith RRMS were sampled during a relapse-free time andhad no signs of ongoing focal inflammatory activity in theMRI Moreover most of the patients (68) were using adisease-modifying treatment at the time of evaluationHence similar as observed in other studies35233637 thisresulted in low serum NfL concentration Following thesimilarity in the NfL levels between healthy controls andpatients with MS the main results of our study are based ona relatively low number of patients In addition the patientnumbers under different immunomodulatory treatmentsvaried substantially However the heterogeneity of thecohort was taken account in the multiple regression modelwhere both EDSS score and disease-modifying treatmentwere used as adjustments
Our findings suggest that elevated serum NfL in MS resultsfrom neuroaxonal damage both within the NAWM and infocal T2 lesions The association between DTI-measurablediffuse microstructural white matter damage and serumNfL isfurther conceptual evidence for the latter being a usefulmonitoring tool in assessing the degree of ongoing neuro-degenerative processes in MS
AcknowledgmentThe authors thank all patients with MS and healthycontrols who participated in this study They also thank thepersonnel of Turku PET Centre for their excellenttechnical assistance
Study fundingThis work was supported by the Finnish Academy the SigridJuselius Foundation
DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgNN for full disclosures
Publication historyReceived by Neurology Neuroimmunology amp NeuroinflammationMay 8 2020 Accepted in final form October 21 2020
References1 Reich DS Lucchinetti CF Calabresi PA Multiple sclerosis N Engl J Med 2018378
169ndash1802 Varhaug KN Torkildsen Oslash Myhr KM Vedeler CA Neurofilament light chain as a
biomarker in multiple sclerosis Front Neurol 201910338
Appendix Authors
Author Location Contribution
MaijaSaraste PhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Acquisition of dataanalyzed andinterpreted the data anddrafted and revised themanuscript forintellectual content
Appendix (continued)
Author Location Contribution
SvetlanaBezukladovaMSc
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Acquisition of dataanalyzed and interpretedthe data and revised themanuscript
MarkusMatilainenPhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Statistical analysisdrafted the figures andrevised the manuscript
Jouni TuiskuMSc
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuFinland
Major role in theacquisition of data andrevised the manuscript
EeroRissanenMDPhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Designed andconceptualized the studyand acquisition of data
MarcusSucksdorffMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Major role in theacquisition of data
SiniLaaksonenMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Acquisition of data
AnnaVuorimaaMD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Acquisition of data
Jens KuhleMD PhD
Neurologic Clinic andPoliclinic Departments ofMedicine Biomedicineand Clinical ResearchUniversity Hospital BaselSwitzerland
Major role in theacquisition of data
DavidLeppert MDPhD
Neurologic Clinic andPoliclinic Departments ofMedicine Biomedicineand Clinical ResearchUniversity Hospital BaselSwitzerland
Revised the manuscriptfor intellectual content
Laura AirasMD PhD
Turku PET Centre TurkuUniversity Hospital andUniversity of TurkuDivision of ClinicalNeurosciences TurkuUniversity Hospital TurkuFinland
Designed andconceptualized the studyand revised themanuscript for intellectualcontent
10 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 NeurologyorgNN
3 Kuhle J Kropshofer H Haering DA et al Blood neurofilament light chain as abiomarker of MS disease activity and treatment response Neurology 201992e1007ndashe1015
4 Varhaug KN Barro C Bjoslashrnevik K et al Neurofilament light chain predicts diseaseactivity in relapsing-remitting MS Neurol Neuroimmunol Neuroinflamm 20185e422
5 Novakova L Zetterberg H Sundstrom P et al Monitoring disease activity in multiplesclerosis using serum neurofilament light protein Neurology 2017892230ndash2237
6 Disanto G Barro C Benkert P et al Serum neurofilament light a biomarker ofneuronal damage in multiple sclerosis Ann Neurol 201781857ndash870
7 Siller N Kuhle J Muthuraman M et al Serum neurofilament light chain is a bio-marker of acute and chronic neuronal damage in early multiple sclerosis Mult Scler201925678ndash686
8 Barro C Benkert P Disanto G et al Serum neurofilament as a predictor of diseaseworsening and brain and spinal cord atrophy in multiple sclerosis Brain 20181412382ndash2391
9 Ferraro D Guicciardi C De Biasi S et al Plasma neurofilaments correlate withdisability in progressive multiple sclerosis patients Acta Neurol Scand 202014116ndash21
10 Jakimovski D Zivadinov R RamanthanM et al Serum neurofilament light chain levelassociations with clinical and cognitive performance in multiple sclerosis a longitu-dinal retrospective 5-year study Mult Scler 2020261670ndash1681
11 Sbardella E Tona F Petsas N Pantano P DTI measurements in multiple sclerosisevaluation of brain damage and clinical implications Mult Scler Int 20132013671730
12 Tae WS Ham BJ Pyun SB Kang SH Kim BJ Current clinical applications ofdiffusion-tensor imaging in neurological disorders J Clin Neurol 201814129ndash140
13 Kurtzke JF Rating neurologic impairment in multiple sclerosis an expanded disabilitystatus scale (EDSS) Neurology 1983331444ndash1452
14 Kuhle J Barro C Andreasson U et al Comparison of three analytical platforms forquantification of the neurofilament light chain in blood samples ELISA electro-chemiluminescence immunoassay and SimoaClinChemLabMed 2016541655ndash1661
15 Bezukladova S Tuisku J Matilainen M et al Insights into disseminated MS brainpathology with multimodal diffusion tensor and PET imaging Neurol NeuroimmunolNeuroinflamm 20207e691
16 Schmidt P Gaser C Arsic M et al An automated tool for detection of FLAIR-hyperintense white-matter lesions in Multiple Sclerosis Neuroimage 2012593774ndash3783
17 Salat DH Greve DN Pacheco JL et al Regional white matter volume differences innondemented aging and Alzheimerrsquos disease Neuroimage 2009441247ndash1258
18 Rissanen E Tuisku J Vahlberg T et al Microglial activation white matter tractdamage and disability in MS Neurol Neuroimmunol Neuroinflamm 20185e443
19 Leemans A Jeurissen B Sijbers J et al ExploreDTI a graphical toolbox for pro-cessing analysing and visualizing diffusion MR data Presented at the 17th AnnualMeeting of International Society for Magnetic Resonance in Medicine Honolulu HI2009 p 3537
20 Benjamini Y Hochberg Y Controlling the false discovery rate a practical and pow-erful approach to multiple testing J R Stat Soc Ser B 199557289ndash300
21 Gelman A Scaling regression inputs by dividing by two standard deviations Stat Med2008272865ndash2873
22 Khalil M Teunissen CE Otto M et al Neurofilaments as biomarkers in neurologicaldisorders Nat Rev Neurol 201814577ndash589
23 Canto E Barro C Zhao C et al Association between serum neurofilament light chainlevels and long-term disease course among patients with multiple sclerosis followedup for 12 years JAMA Neurol 2019761359ndash1366
24 Hogel H Rissanen E Barro C et al Serum glial fibrillary acidic protein correlates withmultiple sclerosis disease severity Mult Scler 202026210ndash219
25 Menke RA Gray E Lu CH et al CSF neurofilament light chain reflects corticospinaltract degeneration in ALS Ann Clin Transl Neurol 20152748ndash755
26 Moore EE Hohman TJ Badami FS et al Neurofilament relates to white mattermicrostructure in older adults Neurobiol Aging 201870233ndash241
27 Ljungqvist J Zetterberg H Mitsis M Blennow K Skoglund T Serum neurofilamentlight protein as a marker for diffuse axonal injury results from a case series studyJ Neurotrauma 2017341124ndash1127
28 Schultz SA Strain JF Adedokun A et al Serum neurofilament light chain levels areassociated with white matter integrity in autosomal dominant Alzheimerrsquos diseaseNeurobiol Dis 2020142104960
29 Narayanan S Fu L Pioro E et al Imaging of axonal damage in multiple sclerosisspatial distribution of magnetic resonance imaging lesions Ann Neurol 199741385ndash391
30 Ciccarelli O Werring DJ Barker GJ et al A study of the mechanisms of normal-appearing white matter damage in multiple sclerosis using diffusion tensor imagingevidence of Wallerian degeneration J Neurol 2003250287ndash292
31 Pokryszko-Dragan A Banaszek A Nowakowska-Kotas M et al Diffusion tensorimaging findings in the multiple sclerosis patients and their relationships to variousaspects of disability J Neurol Sci 2018391127ndash133
32 Kolasa M Hakulinen U Brander A et al Diffusion tensor imaging and disabilityprogression in multiple sclerosis a 4-year follow-up study Brain Behav 20199e01194
33 Andersen O Hildeman A Longfils M et al Diffusion tensor imaging in multiplesclerosis at different final outcomes Acta Neurol Scand 2018137165ndash173
34 Koenig KA Sakaie KE Lowe MJ et al The relationship between cognitive functionand high-resolution diffusion tensor MRI of the cingulum bundle in multiple sclerosisMult Scler 2015211794ndash1801
35 Pardini M Bonzano L Bergamino M et al Cingulum bundle alterations underliesubjective fatigue in multiple sclerosis Mult Scler 201521442ndash447
36 Sejbaek T Nielsen HH Penner N et al Dimethyl fumarate decreases neurofilamentlight chain in CSF and blood of treatment naıve relapsing MS patients J NeurolNeurosurg Psychiatry 2019901324ndash1330
37 Piehl F Kockum I Khademi M et al Plasma neurofilament light chain levels inpatients with MS switching from injectable therapies to fingolimod Mult Scler 2018241046ndash1054
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 11
DOI 101212NXI000000000000092620218 Neurol Neuroimmunol Neuroinflamm
Maija Saraste Svetlana Bezukladova Markus Matilainen et al High serum neurofilament associates with diffuse white matter damage in MS
This information is current as of December 8 2020
ServicesUpdated Information amp
httpnnneurologyorgcontent81e926fullhtmlincluding high resolution figures can be found at
References httpnnneurologyorgcontent81e926fullhtmlref-list-1
This article cites 36 articles 3 of which you can access for free at
Citations httpnnneurologyorgcontent81e926fullhtmlotherarticles
This article has been cited by 1 HighWire-hosted articles
Subspecialty Collections
httpnnneurologyorgcgicollectionmultiple_sclerosisMultiple sclerosis
httpnnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpnnneurologyorgmiscaboutxhtmlpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnnneurologyorgmiscaddirxhtmlreprintsusInformation about ordering reprints can be found online
Academy of Neurology All rights reserved Online ISSN 2332-7812Copyright copy 2020 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the AmericanPublished since April 2014 it is an open-access online-only continuous publication journal Copyright
is an official journal of the American Academy of NeurologyNeurol Neuroimmunol Neuroinflamm
3 Kuhle J Kropshofer H Haering DA et al Blood neurofilament light chain as abiomarker of MS disease activity and treatment response Neurology 201992e1007ndashe1015
4 Varhaug KN Barro C Bjoslashrnevik K et al Neurofilament light chain predicts diseaseactivity in relapsing-remitting MS Neurol Neuroimmunol Neuroinflamm 20185e422
5 Novakova L Zetterberg H Sundstrom P et al Monitoring disease activity in multiplesclerosis using serum neurofilament light protein Neurology 2017892230ndash2237
6 Disanto G Barro C Benkert P et al Serum neurofilament light a biomarker ofneuronal damage in multiple sclerosis Ann Neurol 201781857ndash870
7 Siller N Kuhle J Muthuraman M et al Serum neurofilament light chain is a bio-marker of acute and chronic neuronal damage in early multiple sclerosis Mult Scler201925678ndash686
8 Barro C Benkert P Disanto G et al Serum neurofilament as a predictor of diseaseworsening and brain and spinal cord atrophy in multiple sclerosis Brain 20181412382ndash2391
9 Ferraro D Guicciardi C De Biasi S et al Plasma neurofilaments correlate withdisability in progressive multiple sclerosis patients Acta Neurol Scand 202014116ndash21
10 Jakimovski D Zivadinov R RamanthanM et al Serum neurofilament light chain levelassociations with clinical and cognitive performance in multiple sclerosis a longitu-dinal retrospective 5-year study Mult Scler 2020261670ndash1681
11 Sbardella E Tona F Petsas N Pantano P DTI measurements in multiple sclerosisevaluation of brain damage and clinical implications Mult Scler Int 20132013671730
12 Tae WS Ham BJ Pyun SB Kang SH Kim BJ Current clinical applications ofdiffusion-tensor imaging in neurological disorders J Clin Neurol 201814129ndash140
13 Kurtzke JF Rating neurologic impairment in multiple sclerosis an expanded disabilitystatus scale (EDSS) Neurology 1983331444ndash1452
14 Kuhle J Barro C Andreasson U et al Comparison of three analytical platforms forquantification of the neurofilament light chain in blood samples ELISA electro-chemiluminescence immunoassay and SimoaClinChemLabMed 2016541655ndash1661
15 Bezukladova S Tuisku J Matilainen M et al Insights into disseminated MS brainpathology with multimodal diffusion tensor and PET imaging Neurol NeuroimmunolNeuroinflamm 20207e691
16 Schmidt P Gaser C Arsic M et al An automated tool for detection of FLAIR-hyperintense white-matter lesions in Multiple Sclerosis Neuroimage 2012593774ndash3783
17 Salat DH Greve DN Pacheco JL et al Regional white matter volume differences innondemented aging and Alzheimerrsquos disease Neuroimage 2009441247ndash1258
18 Rissanen E Tuisku J Vahlberg T et al Microglial activation white matter tractdamage and disability in MS Neurol Neuroimmunol Neuroinflamm 20185e443
19 Leemans A Jeurissen B Sijbers J et al ExploreDTI a graphical toolbox for pro-cessing analysing and visualizing diffusion MR data Presented at the 17th AnnualMeeting of International Society for Magnetic Resonance in Medicine Honolulu HI2009 p 3537
20 Benjamini Y Hochberg Y Controlling the false discovery rate a practical and pow-erful approach to multiple testing J R Stat Soc Ser B 199557289ndash300
21 Gelman A Scaling regression inputs by dividing by two standard deviations Stat Med2008272865ndash2873
22 Khalil M Teunissen CE Otto M et al Neurofilaments as biomarkers in neurologicaldisorders Nat Rev Neurol 201814577ndash589
23 Canto E Barro C Zhao C et al Association between serum neurofilament light chainlevels and long-term disease course among patients with multiple sclerosis followedup for 12 years JAMA Neurol 2019761359ndash1366
24 Hogel H Rissanen E Barro C et al Serum glial fibrillary acidic protein correlates withmultiple sclerosis disease severity Mult Scler 202026210ndash219
25 Menke RA Gray E Lu CH et al CSF neurofilament light chain reflects corticospinaltract degeneration in ALS Ann Clin Transl Neurol 20152748ndash755
26 Moore EE Hohman TJ Badami FS et al Neurofilament relates to white mattermicrostructure in older adults Neurobiol Aging 201870233ndash241
27 Ljungqvist J Zetterberg H Mitsis M Blennow K Skoglund T Serum neurofilamentlight protein as a marker for diffuse axonal injury results from a case series studyJ Neurotrauma 2017341124ndash1127
28 Schultz SA Strain JF Adedokun A et al Serum neurofilament light chain levels areassociated with white matter integrity in autosomal dominant Alzheimerrsquos diseaseNeurobiol Dis 2020142104960
29 Narayanan S Fu L Pioro E et al Imaging of axonal damage in multiple sclerosisspatial distribution of magnetic resonance imaging lesions Ann Neurol 199741385ndash391
30 Ciccarelli O Werring DJ Barker GJ et al A study of the mechanisms of normal-appearing white matter damage in multiple sclerosis using diffusion tensor imagingevidence of Wallerian degeneration J Neurol 2003250287ndash292
31 Pokryszko-Dragan A Banaszek A Nowakowska-Kotas M et al Diffusion tensorimaging findings in the multiple sclerosis patients and their relationships to variousaspects of disability J Neurol Sci 2018391127ndash133
32 Kolasa M Hakulinen U Brander A et al Diffusion tensor imaging and disabilityprogression in multiple sclerosis a 4-year follow-up study Brain Behav 20199e01194
33 Andersen O Hildeman A Longfils M et al Diffusion tensor imaging in multiplesclerosis at different final outcomes Acta Neurol Scand 2018137165ndash173
34 Koenig KA Sakaie KE Lowe MJ et al The relationship between cognitive functionand high-resolution diffusion tensor MRI of the cingulum bundle in multiple sclerosisMult Scler 2015211794ndash1801
35 Pardini M Bonzano L Bergamino M et al Cingulum bundle alterations underliesubjective fatigue in multiple sclerosis Mult Scler 201521442ndash447
36 Sejbaek T Nielsen HH Penner N et al Dimethyl fumarate decreases neurofilamentlight chain in CSF and blood of treatment naıve relapsing MS patients J NeurolNeurosurg Psychiatry 2019901324ndash1330
37 Piehl F Kockum I Khademi M et al Plasma neurofilament light chain levels inpatients with MS switching from injectable therapies to fingolimod Mult Scler 2018241046ndash1054
NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 1 | January 2021 11
DOI 101212NXI000000000000092620218 Neurol Neuroimmunol Neuroinflamm
Maija Saraste Svetlana Bezukladova Markus Matilainen et al High serum neurofilament associates with diffuse white matter damage in MS
This information is current as of December 8 2020
ServicesUpdated Information amp
httpnnneurologyorgcontent81e926fullhtmlincluding high resolution figures can be found at
References httpnnneurologyorgcontent81e926fullhtmlref-list-1
This article cites 36 articles 3 of which you can access for free at
Citations httpnnneurologyorgcontent81e926fullhtmlotherarticles
This article has been cited by 1 HighWire-hosted articles
Subspecialty Collections
httpnnneurologyorgcgicollectionmultiple_sclerosisMultiple sclerosis
httpnnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpnnneurologyorgmiscaboutxhtmlpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnnneurologyorgmiscaddirxhtmlreprintsusInformation about ordering reprints can be found online
Academy of Neurology All rights reserved Online ISSN 2332-7812Copyright copy 2020 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the AmericanPublished since April 2014 it is an open-access online-only continuous publication journal Copyright
is an official journal of the American Academy of NeurologyNeurol Neuroimmunol Neuroinflamm
DOI 101212NXI000000000000092620218 Neurol Neuroimmunol Neuroinflamm
Maija Saraste Svetlana Bezukladova Markus Matilainen et al High serum neurofilament associates with diffuse white matter damage in MS
This information is current as of December 8 2020
ServicesUpdated Information amp
httpnnneurologyorgcontent81e926fullhtmlincluding high resolution figures can be found at
References httpnnneurologyorgcontent81e926fullhtmlref-list-1
This article cites 36 articles 3 of which you can access for free at
Citations httpnnneurologyorgcontent81e926fullhtmlotherarticles
This article has been cited by 1 HighWire-hosted articles
Subspecialty Collections
httpnnneurologyorgcgicollectionmultiple_sclerosisMultiple sclerosis
httpnnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpnnneurologyorgmiscaboutxhtmlpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnnneurologyorgmiscaddirxhtmlreprintsusInformation about ordering reprints can be found online
Academy of Neurology All rights reserved Online ISSN 2332-7812Copyright copy 2020 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the AmericanPublished since April 2014 it is an open-access online-only continuous publication journal Copyright
is an official journal of the American Academy of NeurologyNeurol Neuroimmunol Neuroinflamm
Recommended