ARTICLE OPEN ACCESS Optic Nerve Topography in Multiple Sclerosis Diagnosis The Utility of Visual Evoked Potentials Angela Vidal-Jordana, MD, PhD, Alex Rovira, MD, Georgina Arrambide, MD, PhD, Correspondence Susana Otero-Romero, MD, PhD, Jordi Rı́o, MD, PhD, Manuel Comabella, MD, PhD, Carlos Nos, MD, Dr. Vidal-Jordana Joaquin Castilló, MD, PhD, Ingrid Galan, MD, Sergio Cabello, Dulce Moncho, MD, Kimia Rahnama, MD, avidal@cem-cat.org Vanessa Thonon, MD, Breogan Rodrı́guez-Acevedo, MD, Ana Zabalza, MD, Luciana Midaglia, MD, Cristina Auger, MD, Jaume Sastre-Garriga, MD, PhD, Xavier Montalban, MD, PhD,* and Mar Tintoré, MD, PhD* ® Neurology 2021;96:e482-e490. doi:10.1212/WNL.0000000000011339 Abstract RELATED ARTICLE Objective Editorial To assess the added value of the optic nerve region (by using visual evoked potentials [VEPs]) Optic Nerve in Multiple to the current diagnostic criteria. Sclerosis Diagnostic Criteria: An Aye to the Eyes? Methods From the Barcelona clinically isolated syndrome (CIS) cohort, patients with complete in- Page 139 formation to assess dissemination in space (DIS), the optic nerve region, and dissemination in time at baseline (n = 388) were selected. Modified DIS (modDIS) criteria were constructed by adding the optic nerve to the current DIS regions. The DIS and modDIS criteria were evaluated with univariable Cox proportional hazard regression analyses with the time to the second attack as the outcome. A subset of these patients who had at least 10 years of follow-up or a second attack occurring within 10 years (n = 151) were selected to assess the diagnostic performance. The analyses were also performed according to CIS topography (optic neuritis vs non–optic neuritis). Results The addition of the optic nerve as a fifth region improved the diagnostic performance by slightly increasing the accuracy (2017 DIS 75.5%, modDIS 78.1%) and the sensitivity (2017 DIS 79.2%, modDIS 82.3%) without lowering the specificity (2017 DIS 52.4%, modDIS 52.4%). When the analysis was conducted according to CIS topography, the modDIS criteria performed similarly in both optic neuritis and non–optic neuritis CIS. Conclusion The addition of the optic nerve, assessed by VEP, as a fifth region in the current DIS criteria slightly improves the diagnostic performance because it increases sensitivity without losing specificity. *These authors contributed equally to this work. From the Servicio de Neurologı́a-Neuroinmunologı́a (A V.-J., G.A., S.O.-R., J.R., M.C., C.N., J.C., I.G., S.C., B.R.-A., A.Z., L.M., J.S.-G., X.M., M.T.), Centro de Esclerosis Mú ltiple de Catalunya (Cemcat), Secció n de Neuroradiologia (A.R., C.A.), Servei de Radiologia, Servicio de Medicina Preventiva y Epidemiologia (S.O.-R.), and Servicio de Neurofisiologı́a Clı́nica (D.M., K.R., V.T.), Hospital Universitari Vall d’Hebron, Universitat Autònoma de Barcelona, Spain; and Division of Neurology (X.M.), St. Michael’s Hospital. University of Toronto, Ontario, Canada. 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. The Article Processing Charge was funded by Fondo de Investigación en Salud (FIS: PI17/02162) from Instituto de Salud Carlos III, Ministry of Science, Innovation and Universities, Spain. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal. e482 Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. Glossary CI = confidence interval; CIS = clinically isolated syndrome; DIS = dissemination in space; DIT = dissemination in time; HR = hazard ratio; MAGNIMS = Magnetic Resonance Imaging in Multiple Sclerosis; modDIS = modified DIS; MS = multiple sclerosis; OB = oligoclonal bands; OCT = optical coherence tomography; VEP = visual evoked potential. The improvement in diagnosis is one of the main advances Methods that have occurred in the multiple sclerosis (MS) field in the Study Cohort past decades. Since 2001,1 dissemination in space (DIS) and in time (DIT) can be established from MRI findings, and This observational, retrospective study was based on the since 2010, the diagnosis can be stablished shortly after pre- Barcelona CIS cohort. Clinical, demographic, biological, and radiologic data are collected prospectively following a pre- senting a typical clinically isolated syndrome (CIS) on the results of a single baseline MRI.2,3 More recently, oligoclonal specified protocol described elsewhere.6,15 Briefly, patients bands (OB) have been reincorporated into the 2017 di- <50 years of age first seen within 3 months of symptom onset were included. Demographic data, CIS topography, and agnostic criteria as an alternative criterion for diagnosing MS in the absence of DIT.3,4 presence of OB at baseline were recorded, together with the occurrence of a second attack during follow-up. Brain MRI was obtained at baseline (within 5 months after CIS onset), at Optic neuritis represents the first manifestation of MS in 25% to 35% of patients with CIS5,6 and will occur during 1 year, and every 5 years thereafter. Beginning in 2007, spinal disease course in ≈70% of patients.5 The involvement of cord MRIs were done at baseline as part of the diagnostic process in all patients with CIS. MRI scans were obtained at the optic nerve can be established either clinically or by the use of paraclinical tests such as visual evoked potentials 3.0T since 2010 and at 1.5T previously. Each MRI scan was (VEPs), MRI, and optical coherence tomography (OCT). assessed by experienced neuroradiologists under normal reporting conditions. In 2016, the Magnetic Resonance Imaging in Multiple Sclerosis (MAGNIMS) group proposed the inclusion of the optic nerve as an additional region to demonstrate Pattern-reversal VEPs were performed during the diagnostic process in the same location by an expert neurophysiologist DIS.7 However, due to the lack of evidence regarding the diagnostic properties of the different tests to assess the from the Neurophysiology Unit following international optic nerve involvement in patients with CIS, the optic guidelines.16 Although this is a prospective CIS cohort, the nerve was finally not incorporated into the most recent date of VEP was not part of the initially collected information. revision of the McDonald diagnostic criteria.3 However, we have been able to retrieve this information from electronic clinical records for 288 patients (74.2%). VEPs A few studies have investigated the anterior visual pathway were performed with a mean time from CIS to VEP of 3.95 involvement in CIS cohorts using clinical features,8 months (SD 3.82 months). Preexisting eye conditions were VEP,8,9 OCT,10–13 and MRI9,12,14 alone or in combination, taken into account when VEP results were interpreted. Pa- but only 2 of them have evaluated the diagnostic perfor- tients presenting with any ocular pathology that may affect mance of adding the optic nerve as a new region in the VEP were excluded. According to our normative data, VEPs diagnostic criteria,8,9 yielding somewhat contradictory were considered abnormal if the latency of the P100 wave was results. While the first study8 found that the addition of >112 milliseconds, there was an intereye latency asymmetry optic nerve (evaluated mainly by clinical assessment) >8, milliseconds, or there was an absence of the P100 wave. would improve the diagnostic performance of 2017 Study Design McDonald criteria only in patients with optic neuritis CIS, To evaluate the effect of adding new regions to fulfill DIS, we the study conducted by the MAGNIMS group9 concluded performed a risk assessment analysis. We selected patients that there was no clear improvement of the 2010 McDo- with complete information to assess the 5 DIS regions (brain nald criteria when the optic nerve region (assessed mainly and spinal cord MRI plus VEPs) and with gadolinium ad- by VEP) was added to the 2010 DIS criteria. Moreover, the ministration at baseline to assess DIT (cohort 1). From this modified DIS criterion that included the optic nerve was cohort and to ensure that late converters were taken into the one with the lowest specificity.9 account, patients who had at least 10 years of follow-up or who experienced a second attack within 10 years of the CIS In this work, we have investigated the effect of adding the optic were selected to test the diagnostic criteria performance nerve region (evaluated by VEP) as part of the DIS criteria to (cohort 2) (figure). For both cohorts, DIS and DIT were improve the 2017 diagnostic criteria in a well-characterized assessed at baseline. The modified DIS (modDIS) criteria cohort of patients with CIS.6 We have also compared the added were constructed by adding the optic nerve region to the value of the fifth region according to optic nerve involvement current DIS topographies and using a cutoff value of 2 of 5. (symptomatic vs asymptomatic CIS). The second attack was determined when new symptoms Neurology.org/N Neurology | Volume 96, Number 4 | January 26, 2021 e483 Figure Patient Disposition CIS = clinically isolated syndrome; MS = multiple sclerosis; ON = optic neuritis. suggestive of a relapse occurred during the follow-up after an Cohort 2, Diagnostic Criteria Performance interval of at least 1 month since the CIS and in the absence of We analyzed the sensitivity, specificity, accuracy, positive fever or concurrent diseases.17 predictive value, and negative predictive value of the current and modDIS criteria alone and in combination with either OB Statistical Analysis or MRI DIT, with second attack and 2017 McDonald MS at Descriptive statistics were performed on the baseline vari- 10 years as the primary and secondary outcomes, respectively. ables. Normally distributed continuous variables were sum- A sensitivity analysis excluding patients with treatment onset marized with means and SDs. Otherwise, continuous variables before presenting a second attack was conducted. were summarized with medians and ranges; categorical vari- ables were summarized as percentages. To test whether the fulfillment of the modDIS criteria would have a different impact in symptomatic (optic neuritis) and Cohort 1, Risk Assessment Analysis asymptomatic (non–optic neuritis) patients with CIS, both For the risk assessment analysis, the number of DIS regions risk assessment and diagnostic performance analyses were fulfilled and the current and modDIS criteria were evaluated conducted according to CIS topography. with univariable Cox proportional hazard regression analyses considering the time to the second attack during the follow-up All statistical tests were performed with SPSS software version as the primary outcome and the time to 2017 McDonald MS 22 (SPSS Inc, Chicago, IL) and R version 3.6.0 (R Foundation criteria fulfillment as the secondary outcome. The results are for Statistical Computing, Vienna, Austria). expressed as the hazard ratios (HRs) with 95% confidence intervals (CIs). The Schoenfeld residuals have been evaluated to test the proportional risk assumption. Finally, because Standard Protocol Approvals and initiating a disease-modifying treatment before a second at- Patient Consents tack could have affected the primary outcome, a sensitivity This study received approval from the Clinical Research analysis with treatment exposure as a time-dependent variable Ethics Committee at Vall d’Hebron University Hospital. All was also conducted. patients signed written informed consents. e484 Neurology | Volume 96, Number 4 | January 26, 2021 Neurology.org/N Data Availability for 1 criterion to HR 22.7, 95% CI 7.9–65.7 for 5 criteria Anonymized data will be shared by request from any qualified affected) (table 2). investigator. Both the 2017 DIS and modDIS criteria at baseline conferred a similar risk for developing a second attack during follow-up Results compared to patients not fulfilling these criteria (2017 DIS: HR 4.3, 95% CI 2.8–6.5; modDIS: HR 4.8, 95% CI 3.0–7.5) (table From January 1995 to December 2017, 1,337 patients were 2). Similar findings were obtained with the 2017 McDonald MS included in the Barcelona CIS inception cohort; 99 patients as the outcome although with wider CIs (2017 DIS: HR 20.4, were excluded for different reasons (figure). Of the remaining 95% CI 13.6–30.6; modDIS: HR 21.8, 95% CI 14.1–33.9). 1,238 patients, 388 fulfilled the criteria for the risk assessment analysis (cohort 1), and 151 were included in the diagnostic In the sensitivity analysis including treatment exposure as a time- criteria performance analysis (cohort 2). Baseline character- dependent variable, similar results were found (HR 4.6, 95% CI istics of the whole cohort, cohort 1, and cohort 2 are detailed 2.9–7.2 for 2017 DIS; HR 5.1, 95% CI 3.1–8.3 for modDIS). in table 1. Of note, in the whole cohort (n = 1,238), patients were younger, and there was a higher proportion of patients When the risk assessment analysis was conducted according with positive OB and with abnormal VEPs. to CIS topography, fulfilling either the 2017 DIS or modDIS criteria (including the optic nerve region) conferred a similar VEPs were abnormal in 126 of 388 patients (32.5%), being higher risk for developing a second attack during the follow- asymptomatic in 13.6%. Mean time of follow-up was 62.7 up in patients with both optic neuritis and non–optic neuritis months (SD 40.4 months) for cohort 1 and 88.9 months (SD CIS (table 3). 37.9 months) for cohort 2. Patients with the longest follow-up (cohort 2) had a greater proportion of abnormal baseline MRI Diagnostic Criteria Performance Analysis and positive OB results. One hundred thirty patients (33.5%) (Cohort 2) presented a second attack, and all occurred within the first 10 The addition of the optic nerve region to the DIS criteria years after CIS onset (mean time 19.6 months, SD 22.3 improved the current DIS diagnostic performance by in- months). creasing the accuracy and sensitivity without decreasing the specificity (table 4). When the current and modDIS criteria Risk Assessment Analysis (Cohort 1) were considered along with the presence of OB or DIT, the After presentation of a CIS, the risk of developing a second modified criteria remained more accurate and more sensitive attack during the follow-up increased as the number of regions and had a specificity similar to that of the current 2017 fulfilled at baseline increased (from HR 5.6, 95% CI 1.9–16.5 McDonald MS criteria, although the improvement was less Table 1 Baseline Characteristics Whole Cohort (n = 1,238) Cohort 1 (n = 388) Cohort 2 (n = 151) Age at CIS, mean (SD), y 31.6 (8.2) 33.5 (7.8) 32.4 (7.4) Female, n (%) 842 (68.0) 263 (67.8) 108 (71.5) CIS topography, n (%) Optic neuritis 446 (36.0) 138 (35.6) 40 (26.5) Non–optic neuritis 787 (63.6) 250 (64.4) 111 (73.5) Positive OB, n (%) 582/990 (58.8) 182/335 (54.3) 100/136 (73.5) Baseline EDSS score, median (range) 1.0 (0–5.5) 1.0 (0–4.5) 1.5 (0–4.5) Abnormal MRI, n (%) 810 (71.2) 275 (71.1) 136 (90.1) Gd-enhancing lesions, n (%) 235 (29.3) 114 (29.4) 71 (47.0) Patients with abnormal VEPs, n (%) 415/1,098 (37.8) 126 (32.5) 53 (35.1) Optic neuritis 297/416 (71.4) 92/138 (66.7) 28/40 (70.0) Non–optic neuritis 118/682 (17.3) 34/250 (13.6) 25/111 (22.5) Time of follow-up, mean (SD), mo 107.9 (98.0) 62.7 (40.4) 88.9 (37.9) Abbreviations: CIS = clinically isolated syndrome; EDSS = Expanded Disability Status Scale; OB = oligoclonal bands; VEP = visual evoked potential. Neurology.org/N Neurology | Volume 96, Number 4 | January 26, 2021 e485 Table 2 Risk Assessment Analysis for the Whole Cohort: HRs for Second Attack as the Outcome No. (%) Second Attack, n (%) HR (95% CI) No. of regions affected 1 criterion 78 (20.1) 19/78 (24.4) 5.6 (1.9–16.5) 2 criteria 49 (12.6) 21/49 (42.9) 11.7 (4.0–34.2) 3 criteria 71 (18.3) 38/71 (53.5) 15.8 (5.6–44.4) 4 criteria 57 (14.7) 24/57 (42.1) 12.5 (4.3–36.1) 5 criteria 39 (10.1) 24/39 (61.5) 22.7 (7.9–65.7) 2017 DIS vs modDIS 2017 DIS 205 (52.8) 103/205 (50.2) 4.3 (2.8–6.5) modDIS 216 (55.7) 107/216 (49.5) 4.8 (3.0–7.5) Abbreviations: CI = confidence interval; DIS = dissemination in space; HR = hazard ratio; modDIS = modified DIS. The 2017 DIS criteria as defined in Thompson et al.3: at least 1 lesion in at least 2 of 4 areas of the CNS (periventricular, cortico-juxtacortical, infratentorial, spinal cord). modDIS added optic nerve assessed by visual evoked potential: at least 1 lesion in at least 2 of 5 areas of the CNS (periventricular, cortico- juxtacortical, infratentorial, spinal cord, optic nerve). noticeable than when DIS was analyzed alone (table 4). McDonald MS criteria, although this improvement was less Similar results were obtained with 2017 McDonald MS as the noticeable (table 5). outcome (2017 DIS vs modDIS: sensitivity 80.7% vs 83.6%, specificity 100% vs 100%, accuracy 82.1% vs 84.8%) and in the sensitivity analysis including only patients who either were Discussion never treated or started treatment after presenting a second attack (data not shown). In patients presenting a CIS, the addition of the optic nerve (both in symptomatic and asymptomatic patients) as a fifth When the diagnostic performance analysis was conducted region to the current DIS criteria confers a higher risk for according to CIS topography, the addition of the optic nerve developing a second attack during follow-up and slightly region improved the current DIS diagnostic performance in improves the diagnostic criteria performance by increasing patients with both optic neuritis vs non–optic neuritis CIS, sensitivity without losing specificity. although this improvement was more noticeable in symp- tomatic (optic neuritis) patients (table 5). Again, when the In our work, we have demonstrated that patients fulfilling the different DIS criteria were considered in combination with modDIS criteria at baseline were at a high risk of developing a presence of OB or DIT, adding the optic nerve region into the second attack during follow-up, which was similar to the risk DIS criteria also resulted in more accurate, more sensitive observed in patients fulfilling the 2017 DIS criteria. These diagnostic criteria, with a specificity similar to that of the 2017 findings are in line with the work by Filippi et al.9 that Table 3 Risk Assessment Analysis According to Optic Nerve Involvement: HRs for Second Attack as the Outcome No. (%) Second Attack n (%) HR (95% CI) Optic neuritis CIS (n = 138) 2017 DIS 61 (44.2) 23/61 (37.7) 3.8 (1.8–8.0) modDIS 70 (50.7) 25/70 (35.7) 4.3 (1.9–9.6) Non–optic neuritis CIS (n = 250) 2017 DIS 144 (57.6) 80/144 (55.6) 4.1 (2.4–6.9) modDIS 146 (58.4) 82/146 (56.2) 4.6 (2.7–8.0) Abbreviations: CI = confidence interval; CIS = clinically isolated syndrome; DIS = dissemination in space; HR = hazard ratio; modDIS = modified DIS. The 2017 DIS criteria as defined in Thompson et al.3: at least 1 lesion in at least 2 of 4 areas of the CNS (periventricular, cortico-juxtacortical, infratentorial, spinal cord). modDIS added optic nerve assessed by visual evoked potential: at least 1 lesion in at least 2 of 5 areas of the CNS (periventricular, cortico- juxtacortical, infratentorial, spinal cord, optic nerve). e486 Neurology | Volume 96, Number 4 | January 26, 2021 Neurology.org/N Table 4 Diagnostic Performance Analysis: Diagnostic Properties at 10 years With Second Attack as the Outcome % (95% CI) Sensitivity Specificity Accuracy PPV NPV 2017 DIS 79.2 (71.2–85.8) 52.4 (29.8–74.3) 75.5 (67.8–82.1) 91.1 (86.7–94.2) 28.9 (19.4–40.9) modDIS 82.3 (74.6–88.4) 52.4 (29.8–74.3) 78.1 (70.7–84.5) 91.4 (87.1–94.4) 32.3 (21.6–45.4) 2017 DIS + DIT and/or OB 73.9 (65.4–81.2) 66.7 (43.0–85.4) 72.9 (62.0–79.8) 93.2 (88.1–96.2) 29.2 (21.3–38.5) modDIS + DIT and/or OB 74.2 (65.9–81.5) 66.7 (43.0–85.4) 73.2 (65.5–80.0) 93.2 (88.4–96.3) 29.2 (21.3–38.5) Abbreviations: CI = confidence interval; DIS = dissemination in space; DIT = dissemination in time; modDIS = modified DIS; NPV = negative predictive value; OB = oligoclonal bands; PPV = positive predictive value. The 2017 DIS criteria as defined in Thompson et al.3: at least 1 lesion in at least 2 of 4 areas of the CNS (periventricular, cortico-juxtacortical, infratentorial, spinal cord). modDIS added optic nerve assessed by visual evoked potential: at least 1 lesion in at least 2 of 5 areas of the CNS (periventricular, cortico- juxtacortical, infratentorial, spinal cord, optic nerve). evaluated the optic nerve involvement (mainly using VEPs) in As for the diagnostic performance analysis, we have proved 241 patients with CIS, reporting a similar higher risk for de- that the addition of the optic nerve to the current DIS criteria veloping a second attack when patients fulfilled either the slightly increases the accuracy and sensitivity without low- 2010 DIS criteria or the 2010 modDIS criteria including the ering the specificity, both in the whole cohort and in optic nerve (2010 DIS: 3.48, 95% CI 2.2–5.6; 2010 DIS plus symptomatic and asymptomatic patients with CIS. These optic nerve: 3.34, 95% CI 1.9–5.6). In addition, we have results are partially discordant with the results published in 2 conducted the risk assessment analysis based on CIS topog- previous works showing that the addition of the optic nerve raphy to test whether the fulfillment of the modDIS criteria to the DIS criteria increased the sensitivity at the expense of would have a different impact in symptomatic and asymp- decreasing the specificity.8,9 In the first work, Filippi et al.9 tomatic patients with CIS (optic neuritis vs non–optic neuritis found that the addition of the optic nerve region to the 2010 CIS) and found similar results. This probably means that DIS criteria increased the sensitivity by lowering the speci- VEPs were able to capture optic nerve involvement in both ficity of the 2010 criteria with second attack 5 years after symptomatic and asymptomatic eyes18 and reinforces the fact presenting the CIS used as the outcome (sensitivity: 2010 that symptomatic lesions should be taken into account when DIS 0.87 vs 2010 DIS + optic nerve 0.90; specificity: 2010 considering the diagnosis of MS in patients with CIS.19,20 DIS 0.33 vs 2010 DIS + optic nerve 0.26).9 More recently, Table 5 Diagnostic Performance Analysis According to Optic Nerve Involvement: Diagnostic Properties at 10 years With Second Attack as the Outcome % (95% CI) Sensitivity Specificity Accuracy PPV NPV Optic neuritis CIS (n = 40) 2017 DIS 69.7 (51.3–84.4) 57.1 (18.4–90.1) 67.5 (51.0–81.4) 88.5 (76.0–94.9) 28.6 (14.9–47.7) modDIS 75.8 (57.7–89.0) 57.1 (18.4–90.1) 72.5 (56.1–85.4) 89.3 (77.6–95.2) 33.3 (17.2–54.7) 2017 DIS + DIT and/or OB 63.6 (45.1–79.6) 85.7 (42.1–99.6) 67.5 (50.9–81.4) 95.5 (77.1–99.2) 33.3 (22.5–46.3) modDIS + DIT and/or OB 66.7 (48.2–82.0) 85.7 (42.1–99.6) 70.0 (53.5–83.4) 95.7 (77.9–99.3) 35.3 (23.6–49.1) Non–optic neuritis CIS (n = 111) 2017 DIS 82.5 (73.4–89.5) 50.0 (23.0–77.0) 78.4 (69.6–85.6) 91.9 (87.0–95.1) 29.2 (17.3–44.8) modDIS 84.5 (75.8–91.1) 50.0 (23.0–77.0) 80.2 (71.5–87.1) 92.1 (87.3–95.2) 31.8 (18.8–48.5) 2017 DIS + DIT and/or OB 77.3 (67.7–85.2) 57.1 (28.9–82.3) 74.8 (65.7–82.5) 92.6 (87.1–95.9) 26.7 (16.9-39-5) modDIS + DIT and/or OB 78.4 (68.8–86.1) 57.1 (28.9–82.3) 75.7 (66.6–83.3) 92.7 (87.3–95.9) 27.6 (17.4–40.8) Abbreviations: CI = confidence interval; CIS = clinically isolated syndrome; DIS = dissemination in space; DIT = dissemination in time; modDIS = modified DIS; NPV = negative predictive value; OB = oligoclonal bands; PPV = positive predictive value. The 2017 DIS criteria as defined in Thompson et al.3: at least 1 lesion in at least 2 of 4 areas of the CNS (periventricular, cortico-juxtacortical, infratentorial, spinal cord). modDIS added optic nerve assessed by visual evoked potential: at least 1 lesion in at least 2 of 5 areas of the CNS (periventricular, cortico- juxtacortical, infratentorial, spinal cord, optic nerve). Neurology.org/N Neurology | Volume 96, Number 4 | January 26, 2021 e487 Brownlee et al.8 analyzed the diagnostic performance of the diagnosis of MS has evolved tremendously since the use of new modified criteria according to CIS topography (optic MRI, our work demonstrates the utility of VEPs (in com- neuritis and non–optic neuritis CIS) using the presence of a bination with brain and spinal cord MRI) in MS diagnosis, second attack after a mean follow-up period of almost 15 supporting the reintroduction of this test in the diagnostic years as the outcome. They reported that the inclusion of the workup of patients with CIS. It is important to bear in mind symptomatic optic nerve involvement (assessed mainly that VEP results may be influenced by several factors clinically) increased the sensitivity and decreased the spec- concerning test acquisition such as patient cooperation, ificity of the 2017 DIS criteria (sensitivity: 2017 DIS 83% vs luminescence and contrast patterns, and stimuli pre- 2017 DIS + optic nerve 95%; specificity: 2017 DIS 68%, sentation rate.30,31 To minimize these factors, VEPs should 2017 DIS + optic nerve 57%).8 Both works reported that the be acquired following international guidelines,16 and a decrease in specificity was no longer seen when the different normative dataset obtained from normal controls using DIS criteria were evaluated along with DIT criteria. The local equipment is commonly used to define the range of differences in the test specificity results between our work normal variation. and the previous published works may be explained in part by the longer follow-up of our cohort compared with Filippi Some limitations should be taken into consideration in the et al., which may have allowed us to detect late converters, or interpretation of our results. VEPs were obtained as part of by the different methods used to evaluate optic nerve in- the routine diagnostic process of patients with CIS, but in- volvement in the Brownlee et al. work. Moreover, we have formation on the time elapsed since CIS to VEP was avail- demonstrated that the addition of the optic nerve to the able for only three-quarters of the patients. Because patients 2017 McDonald criteria increased the diagnostic perfor- included in this analysis were slightly different from the full mance in patients with both optic neuritis (symptomatic) Barcelona CIS cohort (n = 1,238), we cannot exclude some and non–optic neuritis (asymptomatic) CIS, although to a minor selection bias. However, differences in VEP abnor- greater extent in the former group. These results are partially malities were small (37.8 vs 32.5), and our results were still discordant with the results published by Brownlee et al. and significant even with a lower proportion of abnormal VEPs may be partly explained by a slight overrepresentation of in both cohorts 1 and 2. We also acknowledge that by optic neuritis in their work8 or by the fact that VEPs are more selecting only patients with complete information to analyze capable of detecting subclinical optic nerve involvement the modDIS criteria (having brain and spinal cord MRI to- than clinical assessment alone.18,21–23 gether with VEP information) and DIT criteria (adminis- tration of gadolinium at baseline MRI), we could have The 2016 MAGNIMS proposal suggested that optic nerve enriched our cohort. Last, we are aware that disease- involvement may be ascertained clinically (by detecting modifying treatment started before second attack could have optic nerve atrophy or pallor), with neurophysiologic tests interfered with our results; however, no differences were (VEPs), or by imaging with either MRI to detect optic found in our sensitivity analysis when this information was nerve lesions or OCT to detect peripapillary retinal taken into account. nerve fiber layer thinning.7 Assessment of optic nerve involvement based only on clinical findings may be chal- Since 2010, the McDonald diagnostic criteria have allowed lenging if not assessed by a trained neuro-ophthalmolo- the diagnosis of MS immediately after presentation of a typical gist,24 and thus, the use of paraclinical tests to confirm optic CIS,2,3 reducing patients’ uncertainty and improving earlier nerve damage may be advisable. In addition, if we are trying treatment onset. However, with the exclusion of the optic to increase the sensitivity of a diagnostic test, we may want nerve as one of the relevant DIS regions, this early diagnosis to use paraclinical tests that may allow us to detect a higher may not be applied to all patients equally because symp- proportion of patients with asymptomatic optic nerve in- tomatic spinal cord and brainstem syndromes are more volvement compared to clinical tools alone. Full-field VEPs meaningful in the diagnostic process than optic neuritis.32–34 have classically been used to provide paraclinical evidence Thus, in a patient presenting a brainstem CIS, the diagnosis of of demyelination to support the diagnosis of MS.17 Very MS can be established by showing 1 periventricular lesion and few studies have assessed VEPs in purely CIS cohorts,25–29 1 enhancing brainstem lesion in brain MRI, while it will not be reporting rates of abnormal VEP results that range from the case if the same patient had presented an optic neuritis 15% to 49% and that are in line with the abnormal VEP rate with an abnormal VEP result and a brain MRI showing the of our study (32.5%). As expected, we have found VEPs to same enhancing periventricular lesion. be more frequently affected in patients with optic neuritis CIS (66.7%), which is a result similar to that reported in the Our results show that the addition of the optic nerve, assessed literature (67.8%–87%).25,29 Most of the studies evaluating by VEP, as a fifth region in the current DIS criteria improves VEPs in CIS cohorts were published before the imple- the diagnostic performance by slightly increasing sensitivity mentation of the McDonald criteria25–27 to compare the without losing specificity. Thus, our work provides additional diagnostic properties of VEP and MRI to detect patients at evidence that argues in favor of including the optic nerve as a risk of developing MS during follow-up, concluding that new region in the diagnostic criteria. Whether the optic nerve the MRI was the most sensitive tool.25–27 While the involvement in patients with CIS should be confirmed only by e488 Neurology | Volume 96, Number 4 | January 26, 2021 Neurology.org/N means of VEPs, by the use of other structural tests (MRI or OCT), or by a combination of them requires further in- Appendix Authors vestigation in prospective studies with a systematic examina- Name Location Contribution tion of the optic nerve in patients with CIS. Angela Vidal- Hospital Universitario Major role in the acquisition of Jordana, MD, Vall d’Hebron, data; designed and Study Funding PhD Barcelona, Spain conceptualized study; No targeted funding reported. analyzed the data; performed statistical analysis; drafted and revised the manuscript for Disclosures intellectual content Angela Vidal-Jordana receives support for contracts Juan Alex Rovira, Hospital Universitario Major role in the acquisition of Rodes (JR16/00024) from Fondo de Investigaciones Sani- MD Vall d’Hebron, data; interpreted the data; Barcelona, Spain revised the manuscript for tarias, Instituto de Salud Carlos III, Spain, and has engaged in intellectual content consulting and/or participated as speaker in events organized by Novartis, Roche, Teva, Mylan, Biogen, Sanofi-Genzyme, Georgina Hospital Universitario Major role in the acquisition Arrambide, Vall d’Hebron, of data; interpreted and Merck. Alex Rovira serves on scientific advisory boards MD, PhD Barcelona, Spain the data; revised the for Novartis, Sanofi-Genzyme, Icometrix, Synthetic MR, manuscript for intellectual Bayer, Merck, Biogen, and OLEA Medical and has received content speaker honoraria from Bayer, Sanofi-Genzyme, Bracco, Susana Hospital Universitario Interpreted the data; revised Merck-Serono, Teva Pharmaceutical Industries Ltd, Novartis, Otero- Vall d’Hebron, the manuscript for intellectual Roche, and Biogen. Georgina Arrambide has received com- Romero, MD, Barcelona, Spain content PhD pensation for consulting services or participation in advisory boards from Sanofi, Merck, and Roche; research support from Jordi Rio, MD, Hospital Universitario Major role in the acquisition of PhD Vall d’Hebron, data; revised the manuscript Novartis; travel expenses for scientific meetings from Novartis Barcelona, Spain for intellectual content and Roche; and speaking honoraria from Stendhal, Sanofi, Manuel Hospital Universitario Major role in the acquisition of and Merck. Susana Otero-Romero has received speaking Comabella, Vall d’Hebron, data; revised the manuscript honoraria and consulting fees from Genzyme, Biogen-Idec, MD, PhD Barcelona, Spain for intellectual content Novartis, and MSD and research support from Novartis. Jordi Carlos Nos, Hospital Universitario Major role in the acquisition of Rı́o has engaged in consulting and/or participated as speaker/ MD Vall d’Hebron, data chair in events organized by Merck, Novartis, Sanofi, Roche, Barcelona, Spain Biogen, and Excemed. Manuel Comabella has received Joaquin Hospital Universitario Major role in the acquisition of compensation for consulting services and speaking honoraria Castilló, MD, Vall d’Hebron, data PhD Barcelona, Spain from Bayer Schering Pharma, Merk Serono, Biogen-Idec, Teva Pharmaceuticals, Sanofi-Aventis, Genzyme, and Ingrid Galan, Hospital Universitario Major role in the acquisition of MD Vall d’Hebron, data Novartis. Breogan Rodrı́guez-Acevedo has received honoraria Barcelona, Spain for consulting services from Wellspect. Jaume Sastre-Garriga Sergio Hospital Universitario Major role in the acquisition of has engaged in consulting and/or participated as speaker/ Cabello Vall d’Hebron, data chair in events organized by Merck, Novartis, Celgene, Sanofi, Barcelona, Spain Biogen, and Excemed. Xavier Montalban has received Dulce Hospital Universitario Major role in the acquisition of speaking honoraria and travel expenses for participation in Moncho, MD Vall d’Hebron, data scientific meetings and has been a steering committee mem- Barcelona, Spain ber of clinical trials or participated in advisory boards of Kimia Hospital Universitario Major role in the acquisition of clinical trials in the past years with Actelion, Bayer, Biogen, Rahnama, Vall d’Hebron, data MD Barcelona, Spain Celgene, Genzyme, Merck, Novartis, Roche, Sanofi- Genzyme, Teva Pharmaceutical, Excemed, MSIF, and Vanessa Hospital Universitario Major role in the acquisition of Thonon, MD Vall d’Hebron, data NMSS. Mar Tintoré has received compensation for consul- Barcelona, Spain ting services and speaking honoraria from Almirall, Bayer Breogan Hospital Universitario Major role in the acquisition of Schering Pharma, Biogen-Idec, Genzyme, Merck-Serono, Rodriguez- Vall d’Hebron, data Novartis, Roche, Sanofi-Aventis. and Teva Pharmaceuticals. Acevedo, MD Barcelona, Spain M. Tintoré is coeditor of Multiple Sclerosis Journal-ETC. Ana Zabalza, Hospital Universitario Major role in the acquisition of Carlos Nos, Joaquin Castilló, Ingrid Galan, Sergio Cabello, MD Vall d’Hebron, data Barcelona, Spain Dulce Moncho, Kimia Rahnama, Vanessa Thonon, Ana Zabalza, Luciana Midaglia, and Cristina Auger report no Luciana Hospital Universitario Major role in the acquisition of disclosures. Go to Neurology.org/N for full disclosures. Midaglia, MD Vall d’Hebron, data Barcelona, Spain Publication History Cristina Hospital Universitario Major role in the acquisition of Auger, MD Vall d’Hebron, data; revised the manuscript Received by Neurology January 15, 2020. Accepted in final form Barcelona, Spain for intellectual content September 1, 2020. Continued Neurology.org/N Neurology | Volume 96, Number 4 | January 26, 2021 e489 15. Arrambide G, Rovira A, Sastre-Garriga J, et al. Spinal cord lesions: a modest con- tributor to diagnosis in clinically isolated syndromes but a relevant prognostic factor. Appendix (continued) Mult Scler J 2018;24:301–312. 16. Odom JV, Bach M, Brigell M, et al. International Society for Clinical Electrophysiology Name Location Contribution of Vision. ISCEV standard for clinical visual evoked potentials: (2016 update). Doc Ophthalmol 2016;133:1–9. Jaume Hospital Universitario Major role in the acquisition of 17. Poser CM, Paty DW, Scheinberg L, et al. New diagnostic criteria for multiple sclerosis: Sastre- Vall d’Hebron, data; revised the manuscript guidelines for research protocols. Ann Neurol 1983;13:227–231. Garriga, MD, Barcelona, Spain for intellectual content 18. Leocani L, Guerrieri S, Comi G. Visual evoked potentials as a biomarker in PhD multiple sclerosis and associated optic neuritis. J Neuroophthalmol 2018;38: 350–357. Xavier Hospital Universitario Interpreted the data, revised 19. Tintore M, Otero-Romero S, Rı́o J, et al. Contribution of the symptomatic lesion in Montalban, Vall d’Hebron, the manuscript for intellectual establishing MS diagnosis and prognosis. Neurology 2016;87:1368–1374. MD, PhD Barcelona, Spain content 20. Brownlee WJ, Swanton JK, Miszkiel KA, Miller DH, Ciccarelli O. Should the symptomatic region be included in dissemination in space in MRI criteria for MS? Mar Tintoré, Hospital Universitario Designed and conceptualized Neurology 2016;87:680–683. MD, PhD Vall d’Hebron, study, interpreted the data, 21. Gundogan FC, Tas A, Altun S, Oz O, Erdem U, Sobaci G. Color vision versus pattern Barcelona, Spain revised the manuscript for visual evoked potentials in the assessment of subclinical optic pathway involvement in intellectual content multiple sclerosis. Indian J Ophthalmol 2013;61:100–103. 22. Sisto D, Trojano M, Vetrugno M, Trabucco T, Iliceto G, Sborgia C. Subclinical visual involvement in multiple sclerosis: a study by MRI, VEPs, frequency-doubling perimetry, standard perimetry, and contrast sensitivity. Investig Opthalmology Vis Sci 2005;46:1264. References 23. Yuksel B, Dogan B, Koctekin B, et al. Color vision testing versus pattern visual evoked 1. McDonald WI, Compston A, Edan G, et al. Recommended diagnostic criteria for potentials and optical coherence tomography parameters in subclinical optic nerve multiple sclerosis: guidelines from the International Panel on the Diagnosis of Mul- involvement in multiple sclerosis. J Clin Neurosci 2019;61:48–53. tiple Sclerosis. Ann Neurol 2001;50:121–127. 24. Stunkel L, Kung NH, Wilson B, McClelland CM, Van Stavern GP. Incidence and 2. Polman CH, Reingold SC, Banwell B, et al. Diagnostic criteria for multiple sclerosis: causes of overdiagnosis of optic neuritis. JAMA Ophthalmol 2018;136:76–81. 2010 revisions to the McDonald criteria. Ann Neurol 2011;69:292–302. 25. Paty DW, Oger JJ, Kastrukoff LF, et al. MRI in the diagnosis of MS: a prospective 3. Thompson AJ, Banwell BL, Barkhof F, et al. Diagnosis of multiple sclerosis: 2017 study with comparison of clinical evaluation, evoked potentials, oligoclonal banding, revisions of the McDonald criteria. Lancet Neurol 2018;17:162–173. and CT. Neurology 1988;38:180–185. 4. Arrambide G, Tintore M, Espejo C, et al. The value of oligoclonal bands in the 26. Lee KH, Hashimoto SA, Hooge JP, et al. Magnetic resonance imaging of the head in the multiple sclerosis diagnostic criteria. Brain 2018;141:1075–1084. diagnosis of multiple sclerosis: a prospective 2-year follow-up with comparison of clinical 5. Toosy AT, Mason DF, Miller DH. Optic neuritis. Lancet Neurol 2014;13:83–99. evaluation, evoked potentials, oligoclonal banding, and CT. Neurology 1991;41:657–660. 6. Tintore M, Rovira À, Rı́o J, et al. Defining high, medium and low impact prognostic 27. Filippini G, Comi GC, Cosi V, et al. Sensitivities and predictive values of paraclinical factors for developing multiple sclerosis. Brain 2015;138:1863–1874. tests for diagnosing multiple sclerosis. J Neurol 1994;241:132–137. 7. Filippi M, Rocca MA, Ciccarelli O, et al. MRI criteria for the diagnosis of multiple 28. Martinelli V, Dalla Costa G, Messina MJ, et al. Multiple biomarkers improve the sclerosis: MAGNIMS consensus guidelines. Lancet Neurol 2016;15:292–303. prediction of multiple sclerosis in clinically isolated syndromes. Acta Neurol Scand 8. Brownlee WJ, Miszkiel KA, Tur C, Barkhof F, Miller DH, Ciccarelli O. Inclusion of 2017;136:454–461. optic nerve involvement in dissemination in space criteria for multiple sclerosis. 29. Gabelić T, Radmilović M, Posavec V, et al. Differences in oligoclonal bands and visual Neurology 2018;91:e1130–e1134. evoked potentials in patients with radiologically and clinically isolated syndrome. Acta 9. Filippi M, Preziosa P, Meani A, et al. Prediction of a multiple sclerosis diagnosis in Neurol Belg 2013;113:13–17. patients with clinically isolated syndrome using the 2016 MAGNIMS and 2010 30. Walsh P. The clinical role of evoked potentials. J Neurol Neurosurg Psychiatry 2005; McDonald criteria: a retrospective study. Lancet Neurol 2018;17:133–142. 76:ii16–ii22. 10. Knier B, Berthele A, Buck D, et al. Optical coherence tomography indicates disease activity 31. Barton JL, Garber JY, Klistorner A, Barnett MH. The electrophysiological as- prior to clinical onset of central nervous system demyelination. Mult Scler J 2015;22:893–900. sessment of visual function in multiple sclerosis. Clin Neurophysiol Pract 2019; 11. Zimmermann HG, Knier B, Oberwahrenbrock T, et al. Association of retinal ganglion 4:90–96. cell layer thickness with future disease activity in patients with clinically isolated 32. Galetta SL, Balcer LJ. The optic nerve should be included as one of the typical CNS syndrome. JAMA Neurol 2018;75:1071–1079. regions for establishing dissemination in space when diagnosing MS: yes. Mult Scler J 12. London F, Zéphir H, Drumez E, et al. Optical coherence tomography: a window to 2018;24:121–122. the optic nerve in clinically isolated syndrome. Brain 2019;142:903–915. 33. Tintoré M, Montalban X. The optic nerve should be included as one of the typical 13. Oberwahrenbrock T, Ringelstein M, Jentschke S, et al. Retinal ganglion cell and inner CNS regions for establishing dissemination in space when diagnosing MS: no. Mult plexiform layer thinning in clinically isolated syndrome. Mult Scler J 2013;19:1887–1895. Scler J 2018;24:123–125. 14. London F, Zéphir H, Hadhoum N, et al. Optic nerve double inversion recovery 34. Barkhof F. The optic nerve should be included as one of the typical CNS regions hypersignal in patients with clinically isolated syndrome is associated with asymp- for establishing dissemination in space when diagnosing MS: commentary. tomatic gadolinium-enhanced lesion. Mult Scler J 2019;25:1888–1895. Mult Scler J 2018;24:125–126. e490 Neurology | Volume 96, Number 4 | January 26, 2021 Neurology.org/N