ARTICLE

Asymptomatic optic nerve lesionsAn underestimated cause of silent retinal atrophy in MS

Jean-Baptiste Davion, MD, Renaud Lopes, PhD, Elodie Drumez, BST, Julien Labreuche, BST,

Nawal Hadhoum, MD, Julien Lannoy, MD, Patrick Vermersch, MD, Jean-Pierre Pruvo, MD, PhD,

Xavier Leclerc, MD, PhD, Helene Zephir, MD, PhD, and Olivier Outteryck, MD, PhD

Neurology® 2020;94:e2468-e2478. doi:10.1212/WNL.0000000000009504

Correspondence

Dr. Outteryck

olivier.outteryck@chru-lille.fr

AbstractObjectiveTo evaluate the frequency of asymptomatic optic nerve lesions and their role in the asymp-tomatic retinal neuroaxonal loss observed in multiple sclerosis (MS).

MethodsWe included patients with remitting-relapsing MS in the VWIMS study (Analysis of Neu-rodegenerative Process Within Visual Ways In Multiple Sclerosis) (ClinicalTrials.govIdentifier: 03656055). Included patients underwent optical coherence tomography (OCT),optic nerve and brain MRI, and low-contrast visual acuity measurement. In eyes of patientswith MS without optic neuritis (MS-NON), an optic nerve lesion on MRI (3D doubleinversion recovery [DIR] sequence) was considered as an asymptomatic lesion. We con-sidered the following OCT/MRI measures: peripapillary retinal nerve fiber layer thickness,macular ganglion cell + inner plexiform layer (mGCIPL) volumes, optic nerve lesion length,T2 lesion burden, and fractional anisotropy within optic radiations.

ResultsAn optic nerve lesion was detected in half of MS-NON eyes. Compared to optic nerves withoutany lesion and independently of the optic radiation lesions, the asymptomatic lesions wereassociated with thinner inner retinal layers (p < 0.0001) and a lower contrast visual acuity (p ≤0.003). Within eyes with asymptomatic optic nerve lesions, optic nerve lesion length was theonly MRI measure significantly associated with retinal neuroaxonal loss (p < 0.03). IntereyemGCIPL thickness difference (IETD) was lower in patients with bilateral optic nerve DIRhypersignal compared to patients with unilateral hypersignal (p = 0.0317). For the diagnosis ofhistory of optic neuritis, sensitivity of 3D DIR and of mGCIPL IETD were 84.9% and 63.5%,respectively.

ConclusionsAsymptomatic optic nerve lesions are an underestimated and preponderant cause of retinalneuroaxonal loss inMS. 3DDIR sequencemay bemore sensitive than IETDmeasured byOCTfor the detection of optic nerve lesions.

MORE ONLINE

CME CourseNPub.org/cmelist

From the Department of Neuroradiology, INSERM, U1171–Degenerative and Vascular Cognitive Disorders (J.-B.D., R.L., J.-P.P., X.L., O.O.), Department of Biostatistics, EA 2694–SantePublique: Epidemiologie et Qualite des Soins (E.D., J.L.), and Department of Neurology, INSERM, U995–Lille Inflammation Research International Center (N.H., J.L., P.V., H.Z.), CHU Lille,Universite de Lille, France.

Go to Neurology.org/N for full disclosures. Funding information and disclosures deemed relevant by the authors, if any, are provided at the end of the article.

e2468 Copyright © 2020 American Academy of Neurology

Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

Long-term disability in multiple sclerosis (MS) is mainlyrelated to axonal loss.1 Easily accessible to examination andcomposed solely of axons, the retinal nerve fiber layer(RNFL) is a promising imaging biomarker to study axonaldegeneration in MS. RNFL atrophy in MS is mainly due toclinical episodes of optic neuritis (ON).2 ON causes axonallesions of ganglion cells within the optic nerve, leading toretrograde axonal degeneration and RNFL atrophy. Ante-rograde axonal degeneration of ganglion cells can also drivea transsynaptic degeneration of postgeniculate neurons,leading to visual cortex atrophy. RNFL atrophy can also befound in eyes of patients withMS without history of ON (MS-NON).2 Three hypotheses have been mentioned: lesionsof the optic radiations (ORs), primary retinal pathology, andasymptomatic lesions of the optic nerve.3 OR injury in-cluding MS lesions can induce RNFL atrophy.4 Retrogradeaxonal degeneration of the postgeniculate neuron can lead toa transsynaptic axonal degeneration of ganglion cells.

Considering this latter mechanism, RNFL atrophy in MS-NON eyes may represent a window to the brain in MS butnone of the studies focusing on this mechanism has simulta-neously evaluated the possibility of asymptomatic optic nervelesions with highly sensitive tools. A primary retinal de-generative process has also been discussed.5,6 Demyelinatinglesions of the optic nerve without ON exist in MS. Pathologyshowed constant7 to frequent8 optic nerve inflammatorylesions in MS, whereas half of patients with MS experienceON during their lifetime.9 MRI displays inflammatory lesionsof the optic nerve in 20%10 to 38.5%11 of MS-NON eyes, andin 22.1% of clinically isolated syndrome (CIS) NON eyes.12

Recently, we suggested that asymptomatic optic nerve lesionwas the main cause of retinal neuroaxonal loss in CIS.13

It is unknown if these asymptomatic optic nerve lesions areassociated with RNFL atrophy in relapsing-remitting MS(RRMS). Ourmain objective was to study retinal neuroaxonal

GlossaryCIS = clinically isolated syndrome; CIS-NON = clinically isolated syndrome without history of optic neuritis; DIR = doubleinversion recovery; DTI = diffusion tensor imaging; EPI = echoplanar imaging; FA = fractional anisotropy; FLAIR = fluid-attenuated inversion recovery; FOV = field of view; IETD = intereye retinal thickness differences; INL = inner nuclear layer;IQR = interquartile range; mGCIPL = macular ganglion cell + inner plexiform layer; mINL = macular inner nuclear layer;MME = microcystic macular edema; MS = multiple sclerosis; MS-NON = multiple sclerosis without history of optic neuritis;MS-ON = multiple sclerosis with history of optic neuritis; OCT = optical coherence tomography; ON = optic neuritis; ORs =optic radiations; pRNFL = peripapillary retinal nerve fiber layer; RNFL = retinal nerve fiber layer; RRMS = relapsing-remittingmultiple sclerosis;TE = echo time;TI = inversion time;TR = repetition time;VA = visual acuity;VEP = visual evoked potential.

Neurology.org/N Neurology | Volume 94, Number 23 | June 9, 2020 e2469

Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

loss and visual disability associated with asymptomatic opticnerve lesions in patients with RRMS.

MethodsDesign, settings, and participantsWe conducted a transversal pilot study (NCT 03656055) andincluded patients with RRMS fulfilling McDonald 2010 cri-teria, aged 18 to 65 years, treated by natalizumab for at least 6months in our center and seronegative or with a low serumantibody index (<1.5) for JC virus. Natalizumab is associatedwith an early and sustained drop of new or enlarging T2lesions.14 Therefore, our population comprised patients withMS without recent inflammation, but with disease that hasbeen active enough to result in a moderate to marked lesionburden. Patients were not included if it was impossible todetermine the existence or absence of history of ON or if theyhad signs of clinical activity in the last 6 months or any factorthat could modify retinal thickness.

All included patients underwent clinical examination, retinaloptical coherence tomography (OCT), contrast vision ex-amination, and optic nerve/brain MRI. Each evaluation(OCT, MRI, contrast vision) was performed blind to anyclinical or paraclinical data.

Clinical measures were recorded by interrogation and com-plete reading of the medical records: age, sex, date of the firstrelapse, and history of clinical episode of ON. Diagnosis ofON should have been documented by a neurologist or neuro-ophthalmologist. Absence of history of ON was reaffirmedonly if there was no evocative history9 at interrogation and inmedical records.

Data acquisition and analysis

Optical coherence tomographyRetinal OCT was performed with spectral-domain OCT(Spectralis; Heidelberg Engineering, Germany) and respectedOSCAR-IB criteria.15 Our protocol has been detailed pre-viously.16 Classification of peripapillary RNFL (pRNFL)thickness values according to Heidelberg Spectralis healthycontrol database (<1st percentile, <5th percentile) were re-ported. For the calculation of intereye retinal thickness dif-ference, we considered the macular thickness within the6-mm Early Treatment Diabetic Retinopathy Study disc.

Visual acuity (VA) measuresMonocular VAwasmeasured at high (100%), low (2.5%), andvery low (1.25%) contrasts with printed scales PRECISION-VISION-2180, using logarithm of the minimum angle ofresolution (logMAR) unit.

MRI measuresMagnetic resonance images were acquired on a 3T Achievascanner (Philips, the Netherlands) using a 32-channel arrayhead coil. The imaging protocol included 3D T1 turbo field

echo (repetition time [TR]/echo time [TE] = 9.9/4.6 ms,sagittal acquisition, voxel size 1.0 × 1.0 × 1.0 mm, field of view[FOV] 256 × 256 × 160, number of slices 160, sense 2), 3Ddouble inversion recovery (DIR) (TR/TE = 5,500/252 ms,inversion time [TI]–dual 625/2,600, voxel size 1.2 × 1.2 ×1.3 mm, number of excitations 2, fat suppression spectralpresaturation with inversion recovery, FOV 250 × 250 × 195,number of slices 150, sense 2), 3D fluid-attenuated inversionrecovery (FLAIR) (TR/TE = 8,000/334 ms, TI = 2,500 ms,sagittal acquisition, voxel size 1.12 × 1.12 × 1.12 mm, FOV250 × 250 × 180, number of slices 160, sense 3), diffusiontensor imaging (DTI) (32 directions, single shot, 2 b-factors,b-max = 1,000 s/mm2, TR/TE = 12,000/56 ms, axial acqui-sition, voxel size 2.0 × 2.0 × 2.0 mm, FOV 250 × 250 × 132,number of slices 66, sense 2), and 1 B0 with a reversed phase-encoding polarity.

Detection of demyelinating lesions on optic nerve/chiasma/optic tracts was performed by a reading of 3D DIR17 and 3DFLAIR sequences by a trained investigator (O.O.) who wasblind of OCT and clinical data. Length of optic nerve DIRhypersignal was measured directly on MRI workstation.11 Ifseveral hypersignals were present on 1 nerve, length was de-fined as the sum of the length of each hypersignal. For thedetection of demyelinating optic nerve lesions with 3D DIRsequence, sensitivity and specificity were 95% and 94%, re-spectively.17 Intraobserver and interobserver agreement wasexcellent and very good for optic nerve DIR hypersignal de-tection and length measurement, respectively.11 We providethe optic nerve imaging and the corresponding OCT scans ofsome patients in figure 1.

T1-weighted images were processed using FreeSurfer soft-ware (v5.3, surfer.nmr.mgh.harvard.edu/). This includedthe preprocessing steps of nonuniform signal correction,signal and spatial normalizations, skull stripping, and braintissues segmentation. The primary visual cortex (V1) wasidentified from Brodmann area atlas of the Martinos Centerfor Biomedical Imaging. The visual cortex volume wasmeasured as the sum of left and right primary visual cortex,normalized on the intracranial volume estimated by Free-Surfer. To segment the ORs, we warped to the T1 space theJuelich histologic atlas, made from postmortem histologicexamination of 10 human brains from patients withoutneurologic affection. T2 lesion volumes in the ORs weremeasured as follows: brain MS lesions were semi-automatically segmented on 3D FLAIR with ITK-SNAP(v3.6.0, itksnap.org) to create a lesion mask, which waswarped into T1 space. Then, an OR lesion mask was createdby keeping voxels of the warped lesion mask that were in theOR mask. Diffusion tensor images were corrected for eddycurrent and motion artifacts using FSL software (fsl.fmrib.ox.ac.uk/fsl/fslwiki/). Then the susceptibility-induced off-resonance field, inherent to echoplanar imaging (EPI) ac-quisition schemes and responsible for geometric and signalartifacts, was estimated using images with reversed phase-encode blips. This field was applied to correct all diffusion

e2470 Neurology | Volume 94, Number 23 | June 9, 2020 Neurology.org/N

Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

tensor images. OR segmentations defined in the T1 spacewere transformed into the DTI space to calculate a meanfractional anisotropy (FA) in the ORs. We provide imagesof the OR postprocessing analysis for one patient in figure 2.

Statistical analysisThickness and volume of different retinal layers and VA atdifferent contrasts were first compared between eyes withasymptomatic lesions (i.e., eyes without history of ON with

Figure 1 Optic nerve MRI and optical coherence tomography (OCT) of patients with multiple sclerosis (MS)

Patients 1–3: Patients with MS without optic neuritis (MS-NON) and without asymptomatic optic nerve double inversion recovery (DIR) hypersignal. Peri-papillary retinal nerve fiber layer (pRNFL) thicknesses are in normal range without any significant intereye thickness difference (IETD). Patients 4–6: PatientswithMSwith optic neuritis (MS-ON)with unilateral symptomatic optic nerveDIR hypersignal (red arrows) andwithout asymptomatic involvement in the felloweye (patient 4) or presenting asymptomatic optic nerve DIR hypersignal (yellow arrows) of the fellow eye (patients 5 and 6). pRNFL thicknesses of the eyeassociated with optic neuritis (ON) are lower than in the fellow eye. IETD are quite high. Patients 7–12: Patients with MS without ON but with unilateral(patients 7 and 8) or bilateral (patients 9–12) asymptomatic optic nerve DIR hypersignal (yellow arrows). pRNFL thicknesses are lower than normal and IETDare low or very low. Green disc: eyes without ON andwithout asymptomatic optic nerve DIR hypersignal. Yellow disc: eyes without ON but with asymptomaticoptic nerve DIR hypersignal. Red disc: eyes with ON and with symptomatic optic nerve DIR hypersignal. Yellow arrows point to the asymptomatic optic nerveDIR hypersignals and the corresponding hyperintensities on 3D fluid-attenuated inversion recovery (FLAIR) sequence. Red arrows point to the symptomaticoptic nerve DIR hypersignals and the corresponding hyperintensities on 3D FLAIR sequence.

Neurology.org/N Neurology | Volume 94, Number 23 | June 9, 2020 e2471

Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

a homolateral optic nerve hypersignal on 3D DIR) and eyeswithout lesions (i.e., eyes without history of ON withouthomolateral optic nerve hypersignal on 3D DIR), and secondly,compared between eyes with asymptomatic lesions and eyes withsymptomatic lesions (i.e., ON history with a homolateral opticnerve hypersignal on 3D DIR) using linear mixed models byincluding eye subgroups as fixed effect and patients as randomeffects, to account for the correlation between eyes in the samepatient. Comparisons were adjusted further for prespecifiedconfounding factors: age, sex,MS duration, lesion volume inORsdivided by intracranial volume, andmean FA inORs (included asfixed effects into linear mixed models). We used multivariablelinear mixed models to study the associations of the differentretinal layer thicknesses and volumes with the length of the opticnerve lesions onMRI in eyes with asymptomatic or symptomaticlesion subgroups, with the normalized lesion volume in ORs, andwith the mean FA in ORs, including prespecified confoundingfactors (age, sex, and MS duration) as fixed effects and patientsas random effect. We also used a linear mixed model includinga random patient effect to compare the length of the optic nerveDIR hypersignal between symptomatic and asymptomatic cases.Finally, we used an analysis covariance adjusted for age, sex,MS duration, normalized lesion volume, and mean FA in ORsto compare, in patients without history of ON, the primarynormalized visual cortex volume between those with asymp-tomatic hypersignal on one or both optic nerves and thosewithout hypersignal on either optic nerve. Normality of modelresiduals were checked using normal quantile–quantile plots.

Intereye retinal thickness differences (IETD) were comparedbetweenMS subgroups using variance analysis. Post hoc pairwisecomparisons were performed using the Bonferroni correction.

Considering a history of unilateral or bilateral ON as the goldstandard, optimal IETD thresholds for the detection of opticnerve involvement were calculated from the receiver operatingcharacteristic curve by maximizing the Youden index. Diagnosticvalues of the observed optimal IETD (expressed as absolutedifference) thresholds were evaluated by calculating sensibilityand specificity. Successively, we evaluated the ability of IETD andof 3D DIR MRI sequence to detect symptomatic optic nerveinvolvement within the whole cohort (occurrence of unilateral orbilateral clinical episode of ON as gold standard).

All statistical tests were done at the 2-tailed α level of 0.05.Data were analyzed using SAS software (version 9.4; SASInstitute Inc., Cary, NC).

Standard protocol approvals, registrations,and patient consentsThe study is registered at Clinicaltrials.gov (NCT02766205).The study was approved by the independent ethics committeeof Dijon, France, and was performed in accordance with theDeclaration of Helsinki. All patients provided written in-formed consent.

Data availabilityDe-identified participant data are available upon reasonablerequest.

ResultsDescription of the populationBetween March and December 2017, we included 98 patients(72 women) with a mean age at inclusion of 41.5 ± 11.7 years

Figure 2 Postprocessing MRI analysis

The primary visual cortex volume was measuredusing the atlas included in FreeSurfer (A). Thelesions were semiautomatically segmented withITK-SNAP on the fluid-attenuated inversion re-covery (FLAIR) sequence (B), and warped in theT1 space using statistical parametric mapping(SPM) (C). The optic radiationsmask was creatingby warping the Juelich atlas to the T1 space withSPM (D). The optic radiations lesion volume wasmeasured as the lesion volume inside the opticradiation mask (E). The mean fractional anisot-ropy was calculated inside the optic radiationmask warped in the double inversion recoveryspace (F).

e2472 Neurology | Volume 94, Number 23 | June 9, 2020 Neurology.org/N

Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

(range, 19.5–65.0), a median delay since first relapse of 11.6years (interquartile range [IQR], 7.1–16.7; range, 0.8–28.0),and a median natalizumab treatment duration without in-terruption of 5.4 years (IQR, 2.0–8.4; range, 6 months–10.4years). Delay from last ON episode was more than 6 months.No patient developed progressive multifocal leukoencephal-opathy at 23 months follow-up. Lesion volume in ORs variedfrom a minor to a severe burden (extremes: 0.01–6.95 cm3).Over 196 eyes (figure 3), 73 presented at least 1 episode ofON (MS-ON eyes: 37.2% of all eyes, 54.1% of patients). Overthese 73 eyes with ON, we found 60 homolateral optic nerveDIR hypersignals on MRI (MS-ON-DIRpositive eyes, 82.2%).Over 123 MS-NON eyes), 60 had at least 1 asymptomaticoptic nerve DIR hypersignal on MRI (MS-NON-DIRpositive

eyes, 48.8%) and 63 did not (MS-NON-DIRnegative eyes,51.2%). These asymptomatic optic nerve hypersignals in-volved 42 patients (42.9% of our population), being bilateralin 18 patients. Among the 196 eyes, 120 had an optic nerveDIR hypersignal on MRI (MS-DIRpositive eyes, 61.2%). Wefound no patient with optic tract or chiasm involvement.We found microcystic macular edema (MME) on OCT of 2MS-ON-DIRpositive eyes (2 different patients). MME was notfound in any other eyes subgroup. In 196 eyes, 77 (39.3%)presented a global pRNFL <5th percentile and 44 (22.5%)a global pRNFL <1st percentile. In 98 patients, 49 (50%)

presented at least on 1 side a global pRNFL <5th percentile,and 33 (33.7%) a global pRNFL <1st percentile. In patientswith only asymptomatic optic nerve lesions (n = 25), 13(52.0%) presented at least on 1 side a global pRNFL <5thpercentile, and 8 (32.0%) a global pRNFL <1st percentile. Inpatients with only symptomatic optic nerve lesions (n = 36),22 (61.1%) presented at least on 1 side a global pRNFL <5thpercentile, and 15 (41.7%) a global pRNFL <1st percentile.

Asymptomatic optic nerve DIR hypersignal vsno optic nerve lesionsIn MS-NON eyes, eyes with asymptomatic optic nerve DIRhypersignal had a significantly lower temporal (p < 0.0001)and global pRNFL (p < 0.0001) thicknesses and a lowermacular ganglion cell + inner plexiform layer (mGCIPL)volume (p < 0.0001) compared to those with absence ofhypersignal (table 1).

Asymptomatic optic nerve DIR hypersignals were associatedwith a significantly worst contrast VA at 2.5% (p = 0.002) and1.25% (p = 0.003).

Symptomatic vs asymptomatic optic nerveDIR hypersignalAmongMS-DIRpositive eyes (n = 120), eyes with symptomaticoptic nerve DIR hypersignal had significantly lower temporal

Figure 3 Flowchart of our population

DIR = double inversion recovery; MS =multiple sclerosis; MS-NON =multiple sclerosis without history of optic neuritis; MS-ON =multiple sclerosis with historyof optic neuritis.

Neurology.org/N Neurology | Volume 94, Number 23 | June 9, 2020 e2473

Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

and global pRNFL, lower mGCIPL, and higher macular innernuclear layer (mINL) volumes compared to asymptomatichypersignal (table 1). Moreover, eyes with symptomaticoptic nerve DIR hypersignal had a significantly worse VA at1.25%, 2.5%, and 100% contrast. Themean length of the opticnerve hypersignal was significantly higher in symptomaticthan in asymptomatic cases (20.55 mm ± 9.8 vs 13.32 mm ±9.3, p < 0.0001).

Measures independently associated withretinal thickness/volumeAmong MS-NON-DIRpositive eyes (n = 60, table 2), a higherlength of optic nerve DIR hypersignal was significantly asso-ciated with a thinner temporal pRNFL (β = −0.43 μm/mm ±0.18, p = 0.027) and mGCIPL (β = −0.002 mm3/mm ± 0.001,p = 0.022). The quantitative measures of injury within ORs(normalized lesion volume, mean FA) were associated withno retinal measures.

Among MS-ON-DIRpositive eyes (n = 60, table 2), a higherlength of optic nerve DIR hypersignal was significantly asso-ciated with a thinner temporal (β = −0.93 μm/mm ± 0.17, p <0.001) and global pRNFL (β = −0.86 μm/mm ± 0.14, p< 0.001), and with a lowermGCIPL volume (β = −0.007mm3/mm± 0.001, p < 0.001). The normalized lesion volume in ORswas significantly associated with a lower mGCIPL and a highermINL volumes (β = −0.37 mm3/% ± 0.14, p = 0.018 andβ = 0.19 mm3/% ± 0.06, p = 0.009, respectively). The mean FAin ORs was associated with none of the retinal measures.

Among MS-NON-DIRnegative eyes (n = 63, table 2), thequantitative measures of injury within ORs (normalized lesion

volume, mean FA) were significantly associated with a highermGCIPL volume (β = 0.34 mm3/% ± 0.12, p = 0.011 andβ = 0.012 ± 0.0039, p = 0.003, respectively), but not with theother retinal measures.

Impact for optic nerve DIR hypersignal onprimary visual cortex volumeWithin MS-NON, the 25 patients with unilateral or bilateralasymptomatic optic nerve DIR hypersignal had a significantlylower normalized primary visual cortex volume, compared tothe 20 patients without optic nerve DIR hypersignal (0.58% ±0.09 vs 0.67% ± 0.14, p = 0.018). After adjustment for age, sex,MS duration, normalized lesion volume, andmean FA inORs,this difference was not significant (p = 0.089).

Within the whole cohort, the 36 patients with unilateral orbilateral symptomatic optic nerve DIR hypersignal andwithout asymptomatic optic nerve involvement had a sig-nificantly lower normalized primary visual cortex volumecompared to the 20 patients without optic nerve DIRhypersignal (0.60% ± 0.11 vs 0.67% ± 0.14, p = 0.039). Afteradjustment for age, sex, MS duration, normalized lesionvolume, and mean FA in ORs, this difference remainedsignificant (p = 0.029).

Intereye retinal thickness differenceIntereye retinal thickness differences according to different MSsubgroups are described in tables 3 and 4. Patients with uni-lateral optic nerve DIR hypersignal presented a higher pRNFLand mGCIPL-IETD than patients without optic nerve DIRhypersignal (p = 0.0018 and p = 0.0008, respectively) andpatients with bilateral optic nerve hypersignal (p = 0.0493 and

Table 1 Measurement of different retinal layers and visual acuity at different contrasts in eyes without history of opticneuritis according to the presence or absence of an optic nerve lesion and in eyes with an optic nerve lesion onMRI according to their symptomatic or asymptomatic nature

Eyes without optic neuritis Eyes with an optic nerve lesion on MRI

Asymptomaticlesion (n = 60)

No lesion(n = 63)

Adjustedp valuea

Asymptomaticlesion (n = 60)

Symptomaticlesions (n = 60)

Adjustedp valuea

Retinal thickness/volume

Temporal pRNFL, μm 52.5 ± 11.91 66.14 ± 9.3 <0.001 52.5 ± 11.91 48.45 ± 16.25 0.039

Global pRNFL, μm 82.2 ± 9.36 96.70 ± 10.0 <0.001 82.2 ± 9.36 76.55 ± 13.92 <0.001

mGCIPL, mm3 0.49 ± 0.08 0.57 ± 0.06 <0.001 0.49 ± 0.08 0.43 ± 0.09 <0.001

mINL, mm3 0.247 ± 0.023 0.241 ±0.025

0.41 0.247 ± 0.023 0.252 ± 0.028 0.004

Contrast vision acuity

100% contrast 0.03 ± 0.14 −0.03 ± 0.11 0.70 0.03 ± 0.14 0.11 ± 0.27 <0.001

2.5% contrast 0.72 ± 0.16 0.57 ± 0.15 0.002 0.72 ± 0.16 0.81 ± 0.21 <0.001

1.25% contrast 1.01 ± 0.10 0.91 ± 0.14 0.003 1.01 ± 0.10 1.05 ± 0.08 0.001

Abbreviations: mGCIPL = macular ganglion cell + inner plexiform layer; mINL = macular inner nuclear layer; pRNFL = peripapillary nerve fiber layer.Values are expressed as mean ± SD.a Adjusted for age, sex, multiple sclerosis duration, normalized lesion volume in optic radiations, andmean fractional anisotropy in optic radiations. Contrastvision acuity expressed in logarithm of the minimum angle of resolution unit.

e2474 Neurology | Volume 94, Number 23 | June 9, 2020 Neurology.org/N

Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

p = 0.0317, respectively). There was no significant differencebetween IETD of patients with bilateral optic nerve DIRhypersignal and IETD of patients without optic nerve DIRhypersignal (p = 0.1863 for pRNFL and p = 0.1399 formGCIPL).

Among the whole cohort and by considering a history of uni-lateral or bilateral ON as the gold standard, optimal IETDthresholds for the detection of optic nerve involvement were≥6 mm for global pRNFL and ≥2.83 mm for mGCIPL. Sen-sitivity of these optimal pRNFL and mGCIPL-IETD thresh-olds and sensitivity of optic nerve MRI (3D DIR) were 56.6%,67.3%, and 84.9%, respectively. Specificity of these optimalpRNFL and mGCIPL-IETD thresholds and specificity of opticnerve MRI were 86.7%, 67.4%, and 44.4%, respectively.

DiscussionWe found that asymptomatic optic nerve involvement inRRMS is significantly associated with asymptomatic retinalneuroaxonal loss and higher visual disability. As we adjusted tonormalized T2 lesion volume and to microstructural integrityof ORs, these associations seems to be independent of

demyelinating and degenerative processes in the brain, and ofa possible transsynaptic degeneration. Evidence supports thatasymptomatic optic nerve lesions induce retinal atrophy: thestrength of the associations, their independence from anotherexplanation, and the analogy with symptomatic lesions re-sponsible for ON. The significant relation between the opticnerve lesion length and the temporal pRNFL and mGCIPLthicknesses argues for causality. Therefore asymptomaticoptic nerve lesions seem to be an additional important ex-planation of retinal neuroaxonal loss in MS-NON eyes.

A thinning of retinal layers associated with lesions of the ORsis demonstrated experimentally in primates18 but also in MSwith a methodology close to ours,19 explained by a retrogradetranssynaptic degeneration. We observed a positive associa-tion of the mean FA within the ORs and the mGCIPLthickness in the eyes without optic nerve lesions (MS-NON-DIRnegative): a lower mean FA may reflect a poorer conser-vation of the microstructural architecture within the ORs,which might lead to a retrograde transsynaptic degenerationand thus to a greater retinal degeneration.20 Conversely, wealso observed a positive association of the normalized ORlesion volume and the mGCIPL thickness in the same eyeswithout optic nerve lesion (MS-NON-DIRnegative), which is

Table 2 Association between the thickness or volume of different retinal layers and MRI measures in eyes with anasymptomatic optic nerve lesion (n = 60), in eyes with a symptomatic optic nerve lesion (n = 60), and in eyeswithout optic nerve lesion (n = 63) in multivariate analysis including age, sex, and multiple sclerosis duration asprespecified confounding factors

MRI measures

Temporal pRNFL, μm Global pRNFL, μm mGCIPL, mm3 mINL, mm3

β ± SEpValue β ± SE

pValue β ± SE

pValue β ± SE

pValue

Eyes with an asymptomaticoptic nerve lesion (n = 60)

Length of optic nervelesions, mm

−0.43 ± 0.18 0.027 −0.19 ± 0.11 0.10 −0.002 ± 0.001 0.022 0.0002 ± 0.0002 0.47

Normalized opticradiations lesion volume, %

−10.61 ± 18.78 0.58 2.09 ± 19.08 0.91 −0.19 ± 0.155 0.23 0.02 ± 0.04 0.59

Mean FA in optic radiations, % 0.07 ± 0.65 0.91 0.66 ± 0.64 0.32 −0.001 ± 0.005 0.80 0.00004 ± 0.0012 0.97

Eyes with a symptomatic opticnerve lesion (n = 60)

Length of optic nervelesions, mm

−0.93 ± 0.17 <0.001 −0.86 ± 0.14 <0.001 −0.007 ± 0.001 <0.001 0.0007 ± 0.0003 0.06

Normalized optic radiationslesion volume, %

−26.11 ± 32.83 0.44 −37.63 ± 26.85 0.18 −0.37 ± 0.14 0.018 0.19 ± 0.06 0.009

Mean FA in optic radiations, % 0.34 ± 1.12 0.77 −0.07 ± 0.92 0.94 −0.005 ± 0.005 0.28 0.002 ± 0.002 0.28

Eyes without optic nerve lesion(n = 63)

Normalized optic radiationslesion volume, %

12.79 ± 20.75 0.54 27.14 ± 23.47 0.26 0.34 ± 0.12 0.011 −0.09 ± 0.06 0.13

Mean FA in optic radiations, % 0.98 ± 0.68 0.16 1.46 ± 0.76 0.070 0.012 ± 0.0039 0.003 −0.003 ± 0.002 0.083

Abbreviations: FA = fractional anisotropy; mGCIPL = macular ganglion cell + inner plexiform layer; mINL = macular inner nuclear layer; pRNFL = peripapillaryretinal nerve fiber layer.β indicates regression coefficient derived from the multivariable linear mixed model.

Neurology.org/N Neurology | Volume 94, Number 23 | June 9, 2020 e2475

Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

inconsistent with a retrograde transsynaptic degenerationinduced by OR lesions. No significant association was ob-served between the normalized OR lesion volume and thepRNFL. These results may argue against an important role ofOR T2 lesions in retinal neurodegeneration occurring in MS.

The significant association of asymptomatic lesions with low-contrast VA is consistent with previous data,11 showing low-contrast scales to be more prone to detecting post-ON visual

impairment.21 We hypothesize that the asymptomatic nature ofoptic nerve lesions in MS results from a difference in size, in-sufficient to alter the 100% contrast VA. Our results suggest thatasymptomatic optic nerve lesions induce the same morphologicand functional changes on optic ways, but to a lesser extent thansymptomatic lesions. Asymptomatic optic nerve lesions wereshorter and contrary to symptomatic lesions, we failed to dem-onstrate an anterograde transsynaptic degeneration associatedwith asymptomatic lesions.

Table 3 Retinal thickness and intereye retinal thickness difference among patient subgroups classified according tounilateral/bilateral/no, symptomatic, or asymptomatic optic nerve involvement

Patients with MS, without optic neuritisPatients with MS, with optic neuritis, and withsymptomatic optic nerve DIR hypersignal

No/no (n = 20)

No/asymptomatic(n = 7)

Asymptomatic/asymptomatic(n = 18)

No/symptomatic(n = 8)

Symptomatic/asymptomatic(n = 16)

Symptomatic/symptomatic(n = 16)

Optic nerve DIRhypersignal accordingto both eyes

No/no No/asymptomatic Asymptomatic/asymptomatic

No/symptomatic Symptomatic/asymptomatic

Symptomatic/symptomatic

Median retinalthickness of both eyes

Global pRNFL, μm 99.3 (78.5–120.0) 95.5 (77.0–99.5) 81.3 (63.0–92.5) 83.8 (71.5–101.0) 73.5 (63.0–98.0) 80.3 (52.0–96.0)

mGCIPL, μm 70.7 (61.5–81.0) 61.9 (58.7–71.6) 59.1 (53.2–74.5) 61.6 (56.4–72.2) 57.6 (44.6–70.9) 56.8 (41.0–68.3)

Absolute IETD

Global pRNFL, μm 2.0 (0.0–9.0) 5.0 (4.0–14.0) 2.0 (0.0–15.0) 15.5 (4.0–50.0) 6.0 (0.0–46.0) 5.5 (0.0–23.0)

mGCIPL, μm 1.4 (0.0–7.4) 4.9 (0.7–9.9) 1.8 (0.0–7.8) 9.4 (3.5–22.6) 4.2 (0.0–20.9) 2.1 (0.0–14.9)

Abbreviations: DIR = double inversion recovery; mGCIPL = macular ganglion cell + inner plexiform layer; IETD = intereye thickness difference; MS-NON =multiple sclerosis without clinical episode of optic neuritis; MS-ON = multiple sclerosis with history of clinical episode of optic neuritis; pRNFL = peripapillaryretinal nerve fiber layer.OpticnerveDIRhypersignalmaybeabsent (no)orasymptomaticor symptomatic. Everypairingof resultshasbeenconsidered.Valuesare reportedasmedian (range).

Table 4 Retinal thickness and intereye retinal thickness difference (IETD) among patient subgroups classified accordingto unilateral or bilateral or no optic nerve double inversion recovery (DIR) hypersignal (symptomatic orasymptomatic)

No lesion,group 0 (n = 20)

One unilateral lesion,group 1 (n = 15)

Bilateral lesions,group 2 (n = 50) p Global 0 vs 1a 0 vs 2a 1 vs 2a

pRNFL IETD

Median (interquartile range) 2.0 (1.0–3.5) 8.0 (4.0–19.0) 4.0 (2.0–8.0) 0.0026 0.0018 0.1863 0.0493

99th percentile 9 50 46

95th percentile 7.5 50 23

mGCIPL IETD

Median (interquartile range) 1.4 (0.7–2.3) 5.8 (3.9–11.3) 2.5 (1.1–5.5) 0.0012 0.0008 0.1399 0.0317

99th percentile 7.4 22.6 20.9

95th percentile 5.5 22.6 15.6

Abbreviations: mGCIPL = macular ganglion cell + inner plexiform layer; MS-NON = multiple sclerosis without past clinical episode of optic neuritis; MS-ON =multiple sclerosis with past clinical episode of optic neuritis; pRNFL = peripapillary retinal nerve fiber layer.Group 0 includesMS-NON patients without asymptomatic optic nerve DIR hypersignal. Group 1 includesMS-NON patients with only unilateral asymptomaticoptic nerveDIR hypersignal andMS-ONpatientswith only unilateral symptomatic optic nerveDIR hypersignal. Group 2 includesMS-ONandMS-NONpatientswith bilateral optic nerve DIR hypersignal whether it was symptomatic or asymptomatic.a p Value after Bonferroni correction.

e2476 Neurology | Volume 94, Number 23 | June 9, 2020 Neurology.org/N

Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

Frequency of asymptomatic optic nerve lesions was high inour population. This frequency is higher than in previousstudies, possibly because of differences in methods andpopulation: Miller et al.10 found 20% in CIS and Hadhoumet al.11 found 38.5% in patients with less advanced MS. InMS, asymptomatic lesions are frequently observed in spinalcord, brainstem, and brain. Thus it seems unsurprising tofind many asymptomatic lesions on optic nerves. Indeed,pathologic studies reported that optic nerve demyelinatinglesions were near constant in MS.7,8 Some MRI studiescontrast with our results and did not report any asymp-tomatic optic nerve lesion with T2-fat-sat4 and 3D DIR.22

Other studies tried to detect these asymptomatic lesions bysearching for pRNFL asymmetry between the eyes.23

However, asymptomatic optic nerve lesions, which are fre-quently bilateral in our population, would probably makethis method not sensitive enough. We cannot firmly dem-onstrate with our data that optic nerve MRI was better thanOCT (IETD) for the detection of asymptomatic optic nervelesions since we would need another gold standard. How-ever, in case of bilateral optic nerve lesions, IETD was notdifferent from IETD of patients without optic nerve in-volvement. Furthermore, optic nerve MRI (3D DIR) clearlypresented a higher sensitivity to detect a history of clinicalepisode of ON than IETD. If not better, 3D DIR sequencecan at least be considered as a sensitive tool for the detectionof demyelinating optic nerve lesions in MS. In our study,specificity of 3D DIR sequence was clearly underestimatedby the identification of multiple asymptomatic optic nervelesions in eyes without history of ON. Recently, it has beenshown that optic nerve imaging with 3D DIR sequence maybe more sensitive than visual evoked potentials (VEPs).24

If we confirm the preponderant role of symptomatic optic nervelesions noted by others4,20 and highlight the role of asymp-tomatic optic nerve lesions in the retinal neuroaxonal loss ofpatients with RRMS, we cannot exclude a concomitant primaryretinal pathology, previously described as a macular thinning inthe absence of retrograde degeneration of RNFL.5 MacularGCIPL atrophy without global pRNFL atrophy has beenreported recently in eyes without history of ON in patients withearly MS25 and in eyes without history of ON in patients withCIS (CIS-NON).13 In this latter study, asymptomatic retinalneuronal loss in CIS-NON was associated with a temporalpRNFL thinning and the presence of asymptomatic optic nervelesion. Thus, the previously reported primary retinal neuron-opathy might actually be due to an asymptomatic optic nervelesion, itself responsible for macular and temporal pRNFLthinning without significant global pRNFL thinning. TemporalpRNFL values have not been studied in articles focusing on thepotential existence of a primary retinal pathology.5,6

We reported MME in 2% of our population. This MMEprevalence is in line with previous studies reporting MME in0%–6% of patients with MS.16,26–29 Patients with symp-tomatic optic nerve lesions presented a thicker inner nuclearlayer (INL) than patients with asymptomatic lesions and

contrary to some previous studies,27,29 we did not observeMME in eyes without ON. mINL volume has been corre-lated with the optic nerve lesion length in CIS13 but thiscorrelation is weaker than with inner retinal layers.11,13 Inlongitudinal OCT studies, baseline mINL volume has beencorrelated with annualized new T2 lesions.30 In our cross-sectional study, we have some clues in favor of correlationbetween T2 lesion load in ORs and mINL volume but thislink seems weak and is not found in every eye subgroup.

Many studies have attempted to correlate nonvisual measuresto retinal OCT in MS. In MS-NON eyes, retinal atrophy waswell correlated with cerebral volume, cognitive impairment,3

or risk of disability worsening.31 Retinal OCT might be a re-flection of the total axonal loss in the CNS, as a window to thebrain. However, none of these studies looked at asymptomaticoptic nerve lesions with a highly sensitive method. Our resultsare not contradictory with those, but it sheds the light on oneadditional important explanation for asymptomatic retinalneuroaxonal loss in RRMS. OCT seems to be a window to theoptic nerve and every study trying to associate a nonvisualmeasure to retinal OCT should not only consider symp-tomatic but also asymptomatic optic nerve lesions.

Our study has several limitations. We could not excludefalse-positive findings regarding multiple comparisons. Ourpopulation may not be representative of patients withmildly active MS, or of patients with primary or secondaryprogressive MS. Nevertheless, the characteristics of thepatients were varied. Our method of detection of ON his-tory was retrospective. Sensitivity of the 3DDIR sequence isnot perfect. In our population, 3D DIR showed 61.2% ofoptic nerve lesions (symptomatic or not), whether histologyshowed up to 100%.7 It is therefore possible that someasymptomatic lesions were not identified, and difficult tocertify that all optic nerve hypersignals correspond to in-flammatory demyelinating lesions, and not to lesions ofanother type. OR location was obtained from an anatomicalatlas, constituted from healthy subjects; MS-related cerebralatrophy can make this localization imprecise. Finally, wedid not present exhaustive visual function measurements(i.e., visual field testing) and did not perform VEP, whichwould have helped us to look for optic nerve demyelinatinglesions.

Frequency of asymptomatic optic nerve lesions has beenunderestimated in MS-NON eyes. Asymptomatic opticnerve lesions in MS are associated with structural changesand functional changes of the optic ways. Asymptomaticoptic nerve lesions may be the main explanation of retinalatrophy in eyes without ON, before transsynaptic degenerationinduced by OR lesions. Our results demonstrate the im-portance of studying the optic ways as a whole. Finally, wealso suggest that optic nerve MRI may be more sensitivethan IETD for the detection of optic nerve lesions in MS,because of the bilateral optic nerve involvement we fre-quently observed.

Neurology.org/N Neurology | Volume 94, Number 23 | June 9, 2020 e2477

Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

AcknowledgmentThe authors thank the In-vivo Imaging and Functions corefacility (ci2c.fr) for its help with data analysis: Romain Viard,Julien Dumont, Matthieu Vanhoutte, and Clement Bour-nonville; Maxime Thoor and Chloe Crinquette for MRIacquisition; and Julie Petit for help with data management.

Study fundingNo targeted funding reported.

DisclosureThe authors report no disclosures relevant to the manuscript.Go to Neurology.org/N for full disclosures.

Publication historyReceived by Neurology June 6, 2019. Accepted in final formJanuary 14, 2020.

References1. Criste G, Trapp B, Dutta R. Axonal loss in multiple sclerosis: causes and mechanisms.

Handb Clin Neurol 2014;122:101–113.2. Petzold A, Balcer LJ, Calabresi PA, et al. Retinal layer segmentation in multiple

sclerosis: a systematic review and meta-analysis. Lancet Neurol 2017;16:797–812.3. Oertel FC, Zimmermann HG, Brandt AU, Paul F. Novel uses of retinal imaging with

optical coherence tomography in multiple sclerosis. Expert Rev Neurother 2018:1–13.4. Gabilondo I, Martınez-Lapiscina EH, Martınez-Heras E, et al. Trans-synaptic axonal

degeneration in the visual pathway in multiple sclerosis. Ann Neurol 2014;75:98–107.5. Saidha S, Syc SB, Ibrahim MA, et al. Primary retinal pathology in multiple sclerosis as

detected by optical coherence tomography. Brain 2011;134:518–533.6. Brandt AU, Oberwahrenbrock T, Ringelstein M, et al. Primary retinal pathology in

multiple sclerosis as detected by optical coherence tomography. Brain 2011;134:e193.7. Mogensen PH. Histopathology of anterior parts of the optic pathway in patients with

multiple sclerose. Acta Ophthalmol 1990;68:218–220.8. Toussaint D, Perier O, Verstappen A, Bervoets S. Clinicopathological study of the

visual pathways, eyes, and cerebral hemispheres in 32 cases of disseminated sclerosis.J Clin Neuroophthalmol 1983;3:211–220.

9. Balcer LJ. Clinical practice: optic neuritis. N Engl J Med 2006;354:1273–1280.10. Miller DH, Newton MR, van der Poel JC, et al. Magnetic resonance imaging of the

optic nerve in optic neuritis. Neurology 1988;38:175–179.11. Hadhoum N, Hodel J, Defoort-Dhellemmes S, et al. Length of optic nerve double

inversion recovery hypersignal is associated with retinal axonal loss. Mult Scler 2016;22:649–658.

12. London F, Zephir H, Hadhoum N, et al. Optic nerve double inversion recoveryhypersignal in patients with clinically isolated syndrome is associated with asymp-tomatic gadolinium-enhanced lesion. Mult Scler 2019;25:1888–1895.

13. London F, Zephir H, Drumez E, et al. Optical coherence tomography: a window tothe optic nerve in clinically isolated syndrome. Brain 2019;142:903–915.

14. Polman CH, O’Connor PW, Havrdova E, et al. A randomized, placebo-controlled trialof natalizumab for relapsing multiple sclerosis. N Engl J Med 2006;354:899–910.

15. Schippling S, Balk LJ, Costello F, et al. Quality control for retinal OCT in multiplesclerosis: validation of the OSCAR-IB criteria. Mult Scler 2015;21:163–170.

16. Outteryck O, Majed B, Defoort-Dhellemmes S, et al. A comparative optical coherencetomography study in neuromyelitis optica spectrum disorder and multiple sclerosis.Mult Scler 2015;21:1781–1793.

17. Hodel J, Outteryck O, Bocher AL, et al. Comparison of 3D double inversion recoveryand 2D STIR FLAIR MR sequences for the imaging of optic neuritis: pilot study. EurRadiol 2014;24:3069–3075.

18. Johnson H, Cowey A. Transneuronal retrograde degeneration of retinal ganglion cellsfollowing restricted lesions of striate cortex in the monkey. Exp Brain Res 2000;132:269–275.

19. Reich DS, Smith SA, Gordon-Lipkin EM, et al. Damage to the optic radiation inmultiple sclerosis is associated with retinal injury and visual disability. Arch Neurol2009;66:998–1006.

20. Balk LJ, Steenwijk MD, Tewarie P, et al. Bidirectional trans-synaptic axonal de-generation in the visual pathway in multiple sclerosis. J Neurol Neurosurg Psychiatry2015;86:419–424.

21. Talman LS, Bisker ER, Sackel DJ, et al. Longitudinal study of vision and retinal nervefiber layer thickness in multiple sclerosis. Ann Neurol 2010;67:749–760.

22. Puthenparampil M, Federle L, Poggiali D, et al. Trans-synaptic degeneration in theoptic pathway: a study in clinically isolated syndrome and early relapsing-remittingmultiple sclerosis with or without optic neuritis. PLoS One 2017;12:e0183957.

23. Nolan RC, LiuM, Akhand O, et al. Optimal inter-eye difference thresholds by OCT inMS: an international study. Ann Neurol 2019;85:618–629.

24. Riederer I, Muhlau M, Hoshi MM, et al. Detecting optic nerve lesions in clinicallyisolated syndrome and multiple sclerosis: double-inversion recoverymagnetic resonanceimaging in comparison with visually evoked potentials. J Neurol 2018;266:148–156.

25. Pietroboni AM, Dell’Arti L, Caprioli M, et al. The loss of macular ganglion cells beginsfrom the early stages of disease and correlates with brain atrophy in multiple sclerosispatients. Mult Scler 2019;25:31–38.

26. Gelfand JM, Nolan R, Schwartz DM, et al. Microcystic macular oedema in multiplesclerosis is associated with disease severity. Brain 2012;135:1786–1793.

27. Burggraaff MC, Trieu J, de Vries-Knoppert WAEJ, et al. The clinical spectrum ofmicrocystic macular edema. Invest Ophthalmol Vis Sci 2014;55:952–961.

28. Abegg M, Dysli M, Wolf S, et al. Microcystic macular edema: retrograde maculopathycaused by optic neuropathy. Ophthalmology 2014;121:142–149.

29. Schneider E, Zimmermann H, Oberwahrenbrock T, et al. Optical coherence to-mography reveals distinct patterns of retinal damage in neuromyelitis optica andmultiple sclerosis. PLoS One 2013;8:e66151.

30. Knier B, Schmidt P, Aly L, et al. Retinal inner nuclear layer volume reflects response toimmunotherapy in multiple sclerosis. Brain 2016;139:2855–2863.

31. Martinez-Lapiscina EH, Arnow S, Wilson JA, et al. Retinal thickness measured withoptical coherence tomography and risk of disability worsening in multiple sclerosis:a cohort study. Lancet Neurol 2016;15:574–584.

Appendix Authors

Name Location Contribution

J.B. Davion,MD

University ofLille, France

Acquisition, analysis, or interpretationof data; critical revision of themanuscript

R. Lopes, PhD University ofLille, France

Acquisition, analysis, or interpretationof data; critical revision of themanuscript

E. Drumez,BST

University ofLille, France

Concept and design; acquisition,analysis, or interpretation of data;statistical analyses

J. Labreuche,BST

University ofLille, France

Concept and design; acquisition,analysis, or interpretation of data;statistical analyses

N. Hadhoum,MD

University ofLille, France

Acquisition, analysis, or interpretationof data; critical revision of themanuscript

J. Lannoy, MD University ofLille, France

Acquisition, analysis, or interpretationof data; critical revision of themanuscript

P. Vermersch,MD

University ofLille, France

Critical revision of the manuscript

J.P. Pruvo,MD, PhD

University ofLille, France

Critical revision of the manuscript

X. Leclerc,MD, PhD

University ofLille, France

Critical revision of the manuscript

H. Zephir,MD, PhD

University ofLille, France

Concept and design; acquisition,analysis, or interpretation of data;critical revision of the manuscript

O. Outteryck,MD, PhD

University ofLille, France

Concept and design; acquisition,analysis, or interpretation of data;critical revision of themanuscript; studysupervision

e2478 Neurology | Volume 94, Number 23 | June 9, 2020 Neurology.org/N

Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

DOI 10.1212/WNL.00000000000095042020;94;e2468-e2478 Published Online before print May 20, 2020Neurology 

Jean-Baptiste Davion, Renaud Lopes, Élodie Drumez, et al. MS

Asymptomatic optic nerve lesions: An underestimated cause of silent retinal atrophy in

This information is current as of May 20, 2020

ServicesUpdated Information &

http://n.neurology.org/content/94/23/e2468.fullincluding high resolution figures, can be found at:

References http://n.neurology.org/content/94/23/e2468.full#ref-list-1

This article cites 30 articles, 3 of which you can access for free at:

Subspecialty Collections

http://n.neurology.org/cgi/collection/optic_nerveOptic nerve

http://n.neurology.org/cgi/collection/multiple_sclerosisMultiple sclerosis

http://n.neurology.org/cgi/collection/mriMRIfollowing collection(s): This article, along with others on similar topics, appears in the

  Permissions & Licensing

http://www.neurology.org/about/about_the_journal#permissionsits entirety can be found online at:Information about reproducing this article in parts (figures,tables) or in

  Reprints

http://n.neurology.org/subscribers/advertiseInformation about ordering reprints can be found online:

rights reserved. Print ISSN: 0028-3878. Online ISSN: 1526-632X.1951, it is now a weekly with 48 issues per year. Copyright © 2020 American Academy of Neurology. All

® is the official journal of the American Academy of Neurology. Published continuously sinceNeurology