Journal of Neurology https://doi.org/10.1007/s00415-021-10823-z ORIGINAL COMMUNICATION Treatment response scoring systems to assess long‑term prognosis in self‑injectable DMTs relapsing–remitting multiple sclerosis patients Jordi Río1 · Àlex Rovira2 · Claudio Gasperini3 · Mar Tintoré1 · Luca Prosperini3 · Susana Otero‑Romero1 · Manuel Comabella1 · Ángela Vidal‑Jordana1 · Ingrid Galán1 · Luciana Midaglia1 · Breogán Rodriguez‑Acevedo1 · Ana Zabalza1 · Joaquim Castilló1 · Georgina Arrambide1 · Carlos Nos1 · Álvaro Cobo1 · Carmen Tur1 · Cristina Auger2 · Jaume Sastre‑Garriga1 · Xavier Montalban1 Received: 2 June 2021 / Revised: 15 September 2021 / Accepted: 24 September 2021 © Springer-Verlag GmbH Germany, part of Springer Nature 2021 Abstract Background and objectives Different treatment response scoring systems in treated MS patients exist. The objective was to assess the long-term predictive value of these systems in RRMS patients treated with self-injectable DMTs. Methods RRMS-treated patients underwent brain MRI before the onset of therapy and 12 months thereafter, and neurological assessments every 6 months. Clinical and demographic characteristics were collected at baseline. After the first year of treat- ment, several scoring systems [Rio score (RS), modified Rio score (MRS), MAGNIMS score (MS), and ROAD score (RoS)] were calculated. Cox-Regression and survival analyses were performed to identify scores predicting long-term disability. Results We included 319 RRMS patients. Survival analyses showed that patients with RS > 1 and RoS > 3 had a significant risk of reaching an EDSS of 4.0 and 6.0 The score with the best sensitivity (61%) was the RoS, while the MRS showed the best specificity (88%). The RS showed the best positive predictive value (42%) and the best accuracy (81%). Conclusions The combined measures integrated into different scores have an acceptable prognostic value for identifying patients with long-term disability. Thus, these data reinforce the concept of early treatment optimization to minimize the risk of long-term disability. Keywords Multiple sclerosis · Treatment response · Scores Introduction appropriate consensus and algorithms for the management of treatment response. The increasing number of treatments for the management Previous evidence indicates that clinical and MRI param- of multiple sclerosis (MS) makes it necessary to identify eters during the first months of treatment are good predictors early factors that predict long-term outcomes with the objec- of either treatment response or early Expanded Disability tives of optimizing therapy and of facilitating evidence- Status Scale (EDSS) worsening in the short term [2, 3]. based therapeutic decision-making [1]. However, we lack Based on these parameters, several score systems have been created. Although some of the best-known scores, such as the Rio score (RS), modified Rio score (MRS), MAGNIMS * Jordi Río score (MS), or Risk of ambulatory disability score (RoAD), jrio@cem-cat.org are used to predict short-term response, their long-term 1 Centre d’Esclerosi Múltiple de Catalunya (Cemcat), Servicio predictive value is currently unknown [4–7], Nevertheless, de Neurologia/Neuroimmunolgia, Hospital Universitari Vall some data indicate that the presence of both early on-treat- d’Hebron, Passeig Vall d’Hebron 119‑129, 08035 Barcelona, ment MRI activity and relapses are strongly correlated with Spain severe long-term disability [8]. Moreover, previous data 2 Sección de Neuroradiología, Servicio de Radiologia, from our cohort also demonstrated that scores based on the Hospital Universitari Vall d’Hebron, Universitat Autònoma combination of clinical and MRI measures predict long-term de Barcelona, Barcelona, Spain EDSS worsening [9]. 3 Department of Neurosciences, San Camillo-Forlanini Hospital, Rome, Italy 13 Vol.:(0123456789) Journal of Neurology On the other hand, current scoring systems for predicting (± 1 month) since treatment onset, acquired with the same treatment response are based solely on either clinical factors protocol. collected at MS onset [10–12] or clinical and MRI measures 5. No concomitant participation in experimental trials. collected within the first year after disease-modifying thera- 6. At least 10 years of follow-up after treatment start. pies (DMTs) start [4–6], and provide information only about treatment response in the short term. The recent RoAD score Scores of early disease activity during treatment demonstrated that the combination of baseline factors and 1 year factors was able to predict disability worsening after Patients were classified according to their clinical and MRI 10 years of follow-up [7]. activity in the first year of therapy, as previously described Therefore, based on the above premises, the objec- [4] (Fig. 1). We considered activity after 1 year based on tives were the following: 1/to compare different treatment the following: the presence of relapses, sustained and con- response scoring systems and 2/to evaluate if these scores firmed EDSS worsening and active MRI lesions [new T2 were able to predict long-term disability. or contrast-enhancing lesions (CEL)]. Different scores were assigned and analyzed according to different combinations of these measures of activity, [11–14, 18] namely, RS, MRS MS, and RoAD score. Patients and methods The RS was obtained after the first year of therapy as fol- lows: 1 point for patients with ≥ 3 new T2 and/or CE lesions, Study design and patient disposition 1 point for patients with ≥ 1 relapse, and 1 point for patients with an EDSS score increase in ≥ 1 point, sustained at the This was a single-center, longitudinal, observational study end of follow-up. A total score (0–3) was calculated for each based on a previously described prospective cohort of scoring system, and patients were classified into one of two patients with MS who started treatment with injectable categories: low (score 0–1) or high (score 2–3) risk. drugs (interferon beta or glatiramer acetate) (Fig. 1) from The MRS was obtained after the first year of therapy as 2001 to 2008 [4]. All patients were treatment naive. We col- follows: 1 point for patients with ≥ 5 new T2 lesions, 1 point lected data of patients with RRMS according to the follow- for patients experiencing 1 relapse, and 2 points for patients ing criteria: experiencing ≥ 2 relapses. A total score (0–3) was calculated for each scoring system, and patients were classified into one 1. Subject received any formulation of IFNβ or glatiramer of two categories: low (score 0–1) or high (score 2–3) risk. acetate as the first treatment. The MS was obtained after the first year of therapy as 2. Expanded Disability Status Scale (EDSS) [13] < 4.0 at follows: 1 point for patients with ≥ 3 new T2 lesions, 1 point treatment onset. for patients experiencing 1 relapse, and 2 points for patients 3. At least biannual follow-up visits from treatment ini- experiencing ≥ 2 relapses. A total score (0–2) was calculated tiation, including EDSS scoring performed by certified for each scoring system, and patients were classified into one neurologists. of two categories: low (score 0–1) or high (score 2) risk. 4. Complete data on brain MRI scans at baseline (within Taking into account that RoAD score is the only one that 1 month before starting treatment) and after 1 year considers baseline variables, the score was constructed as Fig. 1  Cohorts and study design 13 Journal of Neurology follows: Baseline factors; 1 point for patients ≥ 40 years, 1 Statistical analysis point for patients with a disease duration ≥ 2 years, 1 point for patients with a baseline EDSS ranging from 1.5 to 2 or Descriptive statistics were used to assess demographic and 2 points for patients with a baseline EDSS > 2, and 1-year clinical data. We calculated the diagnostic properties [sensi- factors; 1 point for patients with ≥ 3 new T2 lesions, 1 point tivity, specificity, positive predictive value (PPV), negative for patients with ≥ 2 CEL lesions, 1 point for patients expe- predictive value (NPV), and accuracy] of each one of the riencing 1 relapse or 2 points for patients experiencing ≥ 2 treatment response scoring systems based on the number of relapses. A total score (0–8) was calculated and patients patients who reached the pre-defined long-term outcomes. were classified into one of two categories: low (score 0–3) The 95% confidence intervals (CIs) of each of these indices or high (score 4–8) risk. for each parameter or scoring system were calculated. Kaplan–Meier survival analyses were used to estimate the cumulative risk of developing the EDSS worsening endpoint Long‑term disability outcomes according to the presence or absence of active disease based on the above-mentioned parameters and score systems after The study outcome was the time to reach the disability the beginning of therapy. We performed uni- and multivari- milestones of EDSS 4 and 6.0, corresponding to walking able logistic regression and Cox proportional hazard regres- restriction and the ability to walk only with unilateral sup- sion analyses to study the prognostic value of early clinical port and < 100 m without assistance and resting, confirmed disease activity for the prediction of long-term disability in at least two consecutive visits and sustained (stable or outcome with adjustment according baseline EDSS score. higher) at the end of follow-up period. The multivariable logistic regression and Cox regression We adopted such a combined outcome instead of the clas- models included the following covariates: sex, age at DMT sical 0.5-point or 1.0-point EDSS worsening [13] to set a onset, disease duration, time to first DMT, relapses before robust outcome based on a clinically significant milestone DMT onset, and baseline EDSS. for patients with MS. The statistical analyses were performed with SPSS 20.0 (SPSS Inc., Chicago, IL, USA), SAS (SAS Institute Inc., Cary, NC, USA), and G-Stat (GlaxoSmithKline S.A., MRI protocol and assessment Madrid, Spain) statistical software packages. The level of statistical significance was set at p < 0.05. All MRI scans were obtained on a 1.5 T magnet, using a standardized protocol that included at least the following sequences: 1/axial proton density-(PD) and T2-weighted Results turbo spin echo; 2/ axial T2-FLAIR (fluid attenuation inver- sion-recovery); and 3/axial T1-weighted spin echo before We studied 319 patients starting self-injectable DMTs who and after injection of 0.2 mmol/kg body weight (5–10 min met the inclusion criteria (Fig. 2) (233 female, 86 male) who delay). For each sequence, 44 contiguous, 3 mm-thick slices started subcutaneous GA 20 mg OD (n = 24), intramuscu- were obtained. lar IFNB-1a 30 mcg OW (n = 120), subcutaneous IFNB- The number of active lesions (new T2 and CEL) on the 1b 250 mcg EOD (n = 72), or subcutaneous IFNB-1a 22 or 12 month MRI scan was visually assessed by two experi- 44 mcg TPW (n = 103) from 2001 to 2008. The patients enced neuroradiologists who were blinded to the patients’ had a mean age of 33.9 (SD 9.1; range18–69) years at the clinical data by direct comparison with the baseline scan, beginning of treatment and a mean disease duration of 4 according to previously published guidelines [14]. (SD 4.6) years. The median EDSS score at baseline was 1.5 Fig. 2  Flowchart of the patients included in the study 13 Journal of Neurology (IQR 1–2). The annualized relapse rate was 1.5 (SD 0.7). an EDSS ≥ 6.0 over the follow-up. 21 (11%) patients with an The mean number of CEL lesions was 3.4 (SD 6.6). There MRS = 0 (reference category), 14 (19%) with an MRS = 1 were no significant baseline differences in terms of age, sex, (HR = 2.2, 95% CI 1.2–4.1, p = 0.01), and 16 (34%) with a disease activity, and EDSS between patients lost to follow- high score (HR = 4.1, 95% CI 2.3–7,5, p < 0.001) reached up and those who underwent a full assessment (data not an EDSS ≥ 6.0 over the follow-up. 18 (10%) patients with shown). a MS = 0 (reference category), 15 (16%) with an MS = 1 A total of 86 (27%) patients required escalation to mono- (HR = 2.2, 95% CI 1.2–4.1, p = 0.015), and 18 (33%) with clonal antibodies (natalizumab = 41, rituximab = 13, alem- high score (HR = 4.3, 95% CI 2.3–7.9, p < 0.001) reached tuzumab = 3) or were changed to oral “high efficacy” agents an EDSS ≥ 6.0 over the follow-up. 20 (9%) patients with (fingolimod = 27, cladribine = 2) over the follow-up period. RoAD = 0–3 (reference category) and 31 (33%) with RoAD = 4–8 (HR = 6.2, 95% CI 3.5–10.8, p = 0.001) reached Early disease activity (first year) an EDSS ≥ 6.0 over the follow-up. As showed in Table 1, similar results were obtained for EDSS 4.0 as outcome Eighty patients (25%) presented one or more relapses with measure. an annualized relapse rate of 0.3 (SD 0.7), 195 (61%) pre- Survival curves showing time to EDSS ≥ 6.0 according to sented new T2 lesions with a mean number of 3.3 (SD 5.3), the level of disease activity after the first year of treatment 80 (25%) had one or more CEL lesions with a mean number are shown in Fig. 3. of 0.8 (SD 2.5), and 39 (12%) suffered a confirmed EDSS Finally, Table 2 shows the diagnostic properties of the worsening of ≥ 1 point. The number of patients scored as highest categories of different scores in identifying patients having low and high risk was 251 (79%) and 67 (21%) who reached (or not) an EDSS ≥ 6.0. The RoAD score according to the RS, 271 (85%) and 47 (15%) according showed the best sensitivity (61%, 95% CI 46–74), the MRS to the MRS, 264 (83%) and 55 (17%) according to the MS the best specificity (88%, 95% CI 84–92), and the RS the and 226 (71%), and 94 (29%) according to the RoAD score, best PPV (42%, 95% CI 33–51) and accuracy (81%, 95% respectively. CI 76–85). Long‑term disability by different scores During the follow-up period, 78 (24%) and 51 patients (16%) Discussion reached an EDSS score of 4.0 and 6.0, respectively. Table 1 shows the proportion of patients with an EDSS score of 4.0 The positive impact of treatments on the natural history of and 6.0 for the different categories of the different scores. MS is becoming increasingly clear [15, 16]. However, to After the first year of treatment, 18 (11%) patients optimize this influence is necessary to identify those patients with RS = 0 (reference category), 5 (6%) with a RS = 1 who will have an inadequate response to injectable treat- (HR = 0.5, 95% CI 0.2–1.1, p = 0.5), and 28 (42%) with a ment at an early stage [2]. Our data demonstrate that patients high score (HR = 5.0, 95% CI 3.2–7.9, p < 0.001) reached with clinical and radiological activity during the first year Table 1  Cox regression models for the risk of reaching the disability milestones of EDSS 4.0 and 6.0 over 10 years of follow-up by level of risk after the first year of treatment according to the different score systems Score Reaching EDSS 4 HR 95% CIs p value Reaching EDSS 6 HR 95% CIs p value RS 0 29/165 (18%) 1.00 18/165 (11%) 1.00 RS 1 11/86 (13%) 0.69 0.33–1.46 0.7 5/86 (6%) 0.51 0.24–1.11 0.5 RS 2–3 38/67 (57%) 6.45 3.43–12.14 < 0.001 28/67 (42%) 5.00 3.18–7.86 < 0.001 MRS 0 34/197 (17%) 1.00 21/197 (11%) 1.00 MRS 1 21/74 (28%) 1.83 0.98–3.42 0.6 14/74 (19%) 2.24 1.21–4.12 0.01 MRS 2–3 23/47 (49%) 4.59 2.33–9.08 < 0.001 16/47 (34%) 4.13 2.26–7.53 < 0.001 MS 0 30/173 (17%) 1.00 18/173 (10%) 1.00 MS 1 24/91 (26%) 1.72 0.93–3.16 0.8 15/91 (16%) 2.18 1.15–4.09 0.015 MS 2 24/55 (44%) 3.72 1.92–7.20 < 0.001 18/55 (33%) 4.25 2.28–7.90 < 0.001 RoAD 0–3 29/224 (13%) 1.00 20/224 (9%) 1.00 RoAD 4–8 49/94 (52%) 7.40 4.22–12.97 < 0.001 31/94 (33%) 4.48 2.6–7.9 < 0.001 RS Rio Score, MRS Modified Rio Score, MS MAGNIMS Score, RoAD Risk of ambulation disability 13 Journal of Neurology Fig. 3  Time to reach the disability milestone of EDSS ≥ 6.0 over Status Scale. A. Rio score. B. Modified Rio score. C. MAGNIMS 10 years of follow-up by risk category after the first year of treatment score. D. RoAD score according to the different score systems. EDSS, Expanded Disability Table 2  Diagnostic properties Sensitivity (%) Specificity (%) NPV (%) PPV (%) Accuracy (%) of the scoring systems after 1 year of treatment for the Rio score (2–3) 55 (40–69) 85 (81–89) 91 (88–93) 42 (33–51) 81 (76–85) risk of reaching the disability Modified Rio score (2–3) 31 (19–46) 88 (84–92) 87 (85–89) 34 (23–47) 79 (74–84) milestone of EDSS ≥ 6.0 over 10 years of follow-up MAGNIMS score (2) 34 (21–49) 86 (82–90) 87 (85–89) 33 (23–44) 78 (73–82) RoAD score (4–8) 61 (46–74) 77 (71–82) 91 (88–94) 33 (27–40) 74 (68–79) NPV Negative predictive value, PPV Positive predictive value of therapy have a significant risk of accruing disability in based on one point of EDSS worsening. By contrast, there the long term. is little information in the literature on the prognostic role The concepts of responder or non-responder patients and of these scores on severe long-term disability (i.e., the mile- when the response is established are not well defined and stone of reaching EDSS ≥ 6.0 after longer periods of follow- nowadays are still arbitrary [17]. In clinical practice, the up). In this regard, our results demonstrate that clinical and lack of response is defined based on the presence of clinical radiological activity (RS 2–3, HR = 5.0, 95% CI 3.2–7.9, (relapses and/or disability worsening) and/or MRI activity, p < 0.001; MRS = 2–3, HR = 4.1, 95% CI 2.3–7.5, p < 0.001; defined by the presence of new T2 or CEL lesions [4–7, 18, MS = 2–3, HR = 4.3, 95% CI 2.3–7.9, p < 0.001; and RoAD 19]. Nevertheless, there is no consensus about the number score = 4–8, HR = 4.5, 95% CI 2.6–7.9, p = 0.001) during the of relapses or active lesions needed to define a confirmed first year of immunomodulatory treatment is a clear indica- poor response. With the above-mentioned variables, differ- tor of poor long-term prognosis (EDSS 6.0 after 10 years of ent score systems of response have been created. The main follow-up) regardless of the score used during the first year objective of these scores is to facilitate the evidence-based of treatment. decision-making process to determine the best treatment The MEDA concept (minimal evidence of disease activ- option for a particular patient. ity) has aroused interest as a composite outcome measure Most scores have been created in cohorts with a relatively that is easily accessible in the clinical practice setting and short follow-up period and the outcomes were generally has a good predictive value in identifying patients without 13 Journal of Neurology a long-term unfavorable outcome [2, 9]. Nevertheless, the RS and RoAD score (a combination of activity and EDSS degree of early disease activity that confers a potential worsening), in predicting a future increase of disability [24]. risk for a poor long-term prognosis is currently unknown. Although sensitivity and specificity are the parameters of Recently, Prosperini et al. analyzed whether the MAGN- most interest from a statistical point of view, PPV and NPV IMS low score, that is, no relapses and < 3 new T2 lesions, are the parameters of most interest from a clinical point of can be considered as MEDA, i.e., minimal disease activ- view. In this case, the NPV should be prioritized, since it ity which does not imply a risk of future disability [20]. In maximizes the “true negatives” (i.e., “true” optimal respond- fact, patients with one relapse and < 3 new T2 lesions were ers), and minimizes the “false negatives” (i.e., suboptimal patients with a risk of long-term disability, therefore; not responders in whom treatment is not switched and there- fulfilling the MEDA definition. On the contrary, our data fore not receives therapy that might be helpful). Overall, show that patients with RS = 1, which includes patients with the predictive value of all criteria was limited. Overall, the one relapse without EDSS worsening and < 3 active lesions, PPV of the different scores is quite poor, since none of them are patients without a significant risk of long-term disability exceeds 50%; however, the NPV is quite good mainly for the (HR 0.5, 95% CI 0.24–1.11). Thus, we can include under RS and RoAD score with values above 90%. As previously the category MEDA those patients with one relapse without reported, the criterion with a best-balanced accuracy was EDSS worsening and no evidence of significant radiological an RS ≥ 2 [17]. activity (< 3 active lesions). Our paper has several limitations. First of all, the influ- Although patients with MRS and MS = 1 are classified ence of different treatment strategies as therapy switching as low risk, according to our data, these patients may have or escalation during follow-up has not been studied. How- a significant risk of disability during a follow-up period of ever, the number of patients who required therapeutic esca- 10 years [HR 2.2; (95% CI 1.2–4.1) for MRS and HR 2.2 lation in this cohort was relatively low probably due to the (95% CI 1.5–4.1) for MS]. These findings demonstrate one fact that therapeutic options were scarce during the study of the fundamental problems of MRS and MS, which is period. In fact, the proportion of patients who reached an their low ability to discriminate patients with high and low EDSS of 4 or 6 is higher in patients who escalate. Therefore, risk of long-term disability. Possibly, the explanation of this there is a modest impact of escalation in this cohort for the phenomenon is related to the fact that neither the MRS nor above-mentioned reasons. Second, other factors different the MS include EDSS worsening during the first year as a from relapses, MR activity, and EDSS worsening, such as variable of the score. Thus, these scores do not differenti- adherence, presence of neutralizing antibodies, oligoclonal ate relapses with or without EDSS worsening, while it may bands, or lesion topography, also can be related to long-term occur in the RS. As previously mentioned, the presence of an disability [25–28]. Third, our findings are limited to patients isolated relapse with no EDSS worsening or MRI activity is initially treated with GA and IFNs. Regardless of the fact not associated with long-term disability. From the foregoing, that it is likely that these scores can be used in patients it is clear that the use of EDSS worsening during the first treated with other first-line drugs, studies to validate this year of treatment helps to discriminate those patients at risk aspect are needed. of long-term disability. The practical issue is that there are By contrast, our study has several strengths. First, because protocols in which the possibility of a therapeutic switch or long-term data are so critical for obtaining valid and accu- escalation is performed in the presence of an isolated relapse rate information on a therapeutic effect, we used very strict [21]. Thus, it is important to consider the radiological activ- long-term outcomes, that is, the disability milestones of ity and/or EDSS worsening before considering a treatment EDSS 4.0 and 6.0. Second, the study has been developed in change in these patients. a real-life setting following a standardized clinical and MRI The RoAD score differs from previous scores in that follow-up protocol, and finally, the study had a low rate of this score is the only one using baseline clinical and demo- patients lost to follow-up. Therefore, and taking into account graphic variables. It ranges from 0 to 8, and patients with that it is important to identify non-responders patients as a score of > 3 are at risk of long-term disability. It has the soon as possible to introduce high-efficacy therapies for advantage that a score of > 3 can be achieved only with minimizing early disability (29–33), our data may be useful baseline clinical and demographic variables. Thus, in these in the management of MS patients in daily clinical practice. patients with poor prognosis before the start of the DMD, In summary, the data reported in this study demonstrate the initiation of induction or intensive treatment could be that in RRMS patients treated with self-injectable DMTs, considered to minimize the risk of long-term disability [22]. the combined clinical–radiological activity measures inte- As previously described, the MRS and MS have the high- grated into different scores during the first year of treatment est specificity (88% and 86%) but with a very low sensitivity have an acceptable prognostic value for identifying patients compared to the RS and RoAD score [23]. The MRS and MS at risk of long-term disability. Furthermore, scores which (pure activity-based measure) do not perform as well as the include early EDSS worsening in the first year of treatment 13 Journal of Neurology seem to be better at discriminating patients with a greater Ethical standard statement The local ethical committee approved the risk of disability worsening. Thus, these data reinforce the study, and all patients provided their informed consent. concept of early treatment optimization to minimize the risk of long-term disability. Declarations References 1. Rotstein D, Montalban X (2019) Reaching an evidence-based Conflicts of interest J Río has received speaking honoraria and per- prognosis for personalized treatment of multiple sclerosis. Nat sonal compensation for participating on Advisory Boards from Almi- Rev Neurol 15(5):287–300 rall, Bayer-Schering Healthcare, Biogen-Idec, Genzyme, Merck- Se- 2. Gasperini C, Prosperini L, Tintoré M et al (2019) Unraveling treat- rono, Novartis, Teva, and Sanofi-Aventis. A Rovira serves on scientific ment response in multiple sclerosis: a clinical and MRI challenge. advisory boards for Novartis, Sanofi-Genzyme, SyntheticMR, Biogen Neurology 92:180–192 and OLEA Medical, and has received speaker honoraria from Bayer, 3. Sormani MP, De Stefano N (2013) Defining and scoring response Sanofi-Genzyme, Merck-Serono, Teva Pharmaceutical Industries Ltd, to IFN-β in multiple sclerosis. Nat Rev Neurol 9:504–512 Novartis, Roche and Biogen. C Gasperini has received fee as speaker 4. Río J, Castilló J, Rovira A et al (2009) Measures in the first year of or advisory board by Merck, Bayer, Biogen, Novartis, Teva, Genzyme. therapy predict the response to interferon beta in MS. Mult Scler M Tintore has received compensation for consulting services and 15(7):848–853 speaking honoraria from Almirall, Bayer Schering Pharma, Biogen- 5. Sormani M, Signori A, Stromillo M, De Stefano N (2013) Refin- Idec, Genzyme, Merck-Serono, Novartis, Roche, Sanofi-Aventis, Viela ing response to treatment as defined by the Modified Rio Score. Bio and Teva Pharmaceuticals. MT is co-editor of Multiple Sclerosis Mult Scler 19(9):1246–1247 Journal- ETC. L Prosperini has received fee as speaker or advisory 6. Sormani MP, Gasperini C, Romeo M et al (2016) Assessing board by Merck, Bayer, Biogen, Novartis, Teva, Genzyme. S Otero- response to interferon-β in a multicenter dataset of patients with Romero has received speaking and consulting honoraria from Gen- MS. Neurology 87(2):134–140 zyme, Biogen-Idec, Novartis, Roche, Excemed and Merk; and research 7. Gasperini C, Prosperini L, Rovira A et al (2021) Scoring the support from Novartis. M Comabella has received compensation for 10 year risk of ambulatory disability in DMD-treated multiple consulting services and speaking honoraria from Bayer Schering Phar- sclerosis patients: the RoAD score. Eur J Neurol 28(8):2533–2542 ma, Merk Serono, Biogen-Idec, Teva Pharmaceuticals, Sanofi-Aventis, 8. Bermel RA, You X, Foulds P et al (2013) Predictors of long-term and Novartis. C Nos has received compensation as steering committee outcome in multiple sclerosis patients treated with interferon beta. member of clinical trials from Hoffmann-La Roche, and funding for Ann Neurol 73:95–103 registration in scientific meetings from Novartis. A Vidal-Jordana re- 9. Río J, Rovira A, Tintoré M et al (2018) Disability progression ceives support for contracts Juan Rodes (JR16/00024) from Fondo de markers over 6–12 years in interferon-β-treated multiple sclerosis Investigaciones Sanitarias, Instituto de Salud Carlos III, Spain; and has patients. Mult Scler 24:322–330 received speaking honoraria and travel expenses from Novartis, Roche, 10. Bergamaschi R, Berzuini C, Romani A, Cosi V (2001) Predicting Teva, Biogen and Sanofi-Genzyme. C Tur is currently being funded by secondary progression in relapsing–remitting multiple sclerosis: a Junior Leader La Caixa Fellowship. The project that gave rise to these a Bayesian analysis. J Neurol Sci 189(1–2):13–21 results received the support of a fellowship from “la Caixa” Founda- 11. Bergamaschi R, Quaglini S, Trojano M et al (2007) Early predic- tion (ID 100010434). The fellowship code is LCF/BQ/PI20/11760008. tion of the long-term evolution of multiple sclerosis: the Bayesian She has also received the 2021 Merck’s Award for the Investigation in Risk Estimate for Multiple Sclerosis (BREMS) score. J Neurol Multiple Sclerosis (Ayudas para la Investigación en Esclerosis Múlti- Neurosurg Psychiatry 78(7):757–759 ple, 2021), awarded by the Merck Foundation. In 2015, she received 12. Bergamaschi R, Montomoli C, Mallucci G et al (2015) BREMSO: an ECTRIMS Post-doctoral Research Fellowship and has received a simple score to predict early the natural course of multiple scle- funding from the UK MS Society. She has also received honoraria rosis. Eur J Neurol 22(6):981–989 and support for traveling from Merck Serono, Sanofi, Roche, TEVA 13. Kurtzke JF (1983) Rating neurologic impairment in multiple Pharmaceuticals, Novartis, Biogen, Bayer and Ismar Healthcare. G. sclerosis: an expanded disability status scale (EDSS). Neurology Arrambide has received compensation for consulting services or par- 33:1444–1452 ticipation in advisory boards from Sanofi and Merck; research support 14. Wattjes MP, Rovira À, Miller D et al (2015) Evidence-based from Novartis; travel expenses for scientific meetings from Novartis, guidelines: MAGNIMS consensus guidelines on the use of MRI Roche, and ECTRIMS; and speaking honoraria from Stendhal, Sanofi, in multiple sclerosis–establishing disease prognosis and monitor- Merck. G. Arrambide is a member of the executive committee of the ing patients. Nat Rev Neurol 11(10):597–606 International Women in Multiple Sclerosis (iWiMS) network. A. Co- 15. Brown JWL, Coles A, Horakova D et al (2019) Association of bo-Calvo has received grant from Instituto de Salud Carlos III, Spain; initial disease-modifying therapy with later conversion to second- JR19/00007. Breogán Rodríguez-Acevedo has received honoraria for ary progressive multiple sclerosis. JAMA 321(2):175–187 consulting services from Wellspect. C Auger has received speaking 16. Harding K, Williams O, Willis M et al (2019) Clinical outcomes honoraria from Novartis, Biogen and Stendhal. J Sastre-Garriga has of escalation vs early intensive disease-modifying therapy in received compensation for participating on Advisory Boards, speak- patients with multiple sclerosis. JAMA Neurol 76(5):536–541 ing honoraria and travel expenses for scientific meetings, consulting 17. Río J, Ruiz-Peña JL (2016) Short-term suboptimal response cri- services or research support from Celgene, Novartis, Biogen, Teva, teria for predicting long-term non-response to first-line disease Merck, Almirall, and Genzyme. X Montalban has received speaking modifying therapies in multiple sclerosis: a systematic review and honoraria and travel expenses for participation inscientific meetings, meta-analysis. J Neurol Sci 361:158–167 has been a steering committee member of clinical trials or participated 18. Río J, Comabella M, Montalban X (2009) Predicting responders in advisory boards of clinical trials in the past with Actelion, Amirall, to therapies for multiple sclerosis. Nat Rev Neurol 5(10):553–560 Bayer, Biogen, Celgene, Genzyme, Hoffmann-La Roche, Novartis, 19. Prosperini L, Gallo V, Petsas N, Borriello G, Pozzilli C (2009) Oryzon Genomics, Sanofi-Genzyme, and Teva Pharmaceutical. L Mi- One year MRI scan predicts clinical response to interferon beta daglia, I Galán, J Castilló, and A Zabalza report no disclosures. in multiple sclerosis. Eur J Neurol 16(11):1202–1209 13 Journal of Neurology 20. Prosperini L, Mancinelli C, Haggiag S et al (2020) Minimal evi- 28. Galassi S, Prosperini L, Logoteta A et al (2016) A lesion topog- dence of disease activity (MEDA) in relapsing–remitting multiple raphy-based approach to predict the outcomes of patients with sclerosis. J Neurol Neurosurg Psychiatry 91:271–277 multiple sclerosis treated with Interferon Beta. Mult Scler Relat 21. Rae-Grant A, Day GS, Marrie RA et al (2018) Practice guide- Disord 8:99–106 line recommendations summary: disease-modifying therapies for 29. Iaffaldano P, Lucisano G, Butzkueven H et al (2021) Early treat- adults with multiple sclerosis neurology. Neurology 90:777–788 ment delays long-term disability accrual in RRMS: results from 22. Rush CA, MacLean HJ, Freedman MS (2015) Aggressive multiple the BMSD network. Mult Scler. https://​doi.​org/​10.​1177/​13245​ sclerosis: proposed definition and treatment algorithm. Nat Rev 82110​10/​28 Neurol 11(7):379–389 30. Kalincik T, Diouf I, Sharmin S et al (2021) Effect of disease- 23. Romeo M, Martinelli V, Rodegher M et al (2015) Validation of modifying therapy on disability in relapsing–remitting multiple 1-year predictive score of long-term response to interferon-β in sclerosis over 15 years. Neurology 96:e783–e797 everyday clinical practice multiple sclerosis patients. Eur J Neurol 31. He A, Merkel B, Brown JWL et al (2020) Timing of high-efficacy 22:973–980 therapy for multiple sclerosis: a retrospective observational cohort 24. Rio J, Nos C, Tintoré M et al (2006) Defining the response to study. Lancet Neurol 19:307–316 interferon-beta in relapsing–remitting multiple sclerosis patients. 32. Buron MD, Chalmer TA, Sellebjerg F et al (2020) Initial high- Ann Neurol 59:344–352 efficacy disease-modifying therapy in multiple sclerosis: a nation- 25. Filippi M, Rocca MA (2010) Multiple sclerosis: monitoring long- wide cohort study. Neurology 95:e1041–e1051 term treatments in multiple sclerosis. Nat Rev Neurol 6:421–422 33. Iaffaldano P, Lucisano G, Caputo F et al (2021) Long-term dis- 26. Alenda R, Costa-Frossard L, Alvarez-Lafuente R et al (2018) ability trajectories in relapsing multiple sclerosis patients treated Blood lymphocyte subsets identify optimal responders to IFN- with early intensive or escalation treatment strategies. Ther Adv beta in MS. J Neurol 265:24–31 Neurol Disord 14:1–10 27. Tintoré M, Rovira A, Río J et al (2008) Do oligoclonal bands add information to MRI in first attacks of multiple sclerosis? Neurol- ogy 70(13 Pt 2):1079–1083 13