# McDonald 2017 — McDonald criteria for multiple sclerosis (2017 revisions)

> Historical 2017 diagnostic algorithm for typical clinically isolated syndrome and progression from onset, integrating clinical attacks, dissemination in space, dissemination in time, and CSF-specific oligoclonal bands. A newer 2024 revision is published and must not be silently mixed into this map.

**Situação:** vigente · **Órgão:** Encéfalo · **Órgão emissor:** International Panel on Diagnosis of Multiple Sclerosis · **Versão:** 2017 historical version; 2024 revision published in 2025 · **Ano:** 2017

> ⚠️ Uma versão mais nova pode existir (em revisão).

## Procedência e vigência
- Família: algoritmo
- Tipo de lógica: flat
- Modalidade: Clinical, MRI, CSF
- Fonte primária: Thompson AJ, Banwell BL, Barkhof F, et al.. Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria (2018) — https://discovery.ucl.ac.uk/id/eprint/10041020
- Última verificação: 2026-07-24
- Última checagem: 2026-08-12

## Lógica de decisão
Apply only to an eligible clinical presentation, name the evidence for each logical gate, and keep the 2017 and 2024 rules separate. A matching MRI phrase is never enough to diagnose MS.

## Categorias

| Código | Rótulo | Critérios | Conduta | Risco | Localizador | Verificado |
| --- | --- | --- | --- | --- | --- | --- |
| periventricular | Periventricular lesion (2017 DIS region) | One of four 2017 MRI regions for dissemination in space. At least one T2-hyperintense lesion characteristic of MS in the periventricular region contributes one region; DIS requires at least one qualifying lesion in at least two of the four regions. | Report number, morphology and distribution rather than the word periventricular alone. In older patients or those with vascular risk factors, seek stronger corroboration and a higher periventricular lesion burden before attributing nonspecific lesions to MS. | A periventricular lesion is not specific for MS and supplies no individual conversion, disability or treatment-benefit probability. It cannot establish DIS or MS by itself. | Thompson et al. 2017 accepted manuscript, Panel 5 (periventricular as one of four regions and two-region DIS rule), Panel 5 footnote on older patients or vascular risk, and Panel 3 misdiagnosis safeguards. | ✓ |
| cortical-juxtacortical | Cortical or juxtacortical lesion (2017 DIS region) | One combined 2017 DIS region: a qualifying cortical or juxtacortical T2 lesion can contribute this region. Multiple cortical and juxtacortical lesions still represent one of the four regional compartments for the two-region DIS rule. | Describe whether the lesion is cortical, juxtacortical, or both and confirm that it is a true lesion rather than artifact. Use it only inside the complete clinical and regional DIS assessment. | This location supports typical distribution but is not independently diagnostic or prognostic. Standard MRI has limited sensitivity for cortical lesions and can confuse artifacts with true lesions. | Thompson et al. accepted manuscript, Panel 4 and cortical-lesion discussion (cortical added alongside juxtacortical; artifact caution), Panel 5 regional list. | ✓ |
| infratentorial | Infratentorial lesion (2017 DIS region) | One of four 2017 DIS regions, encompassing a qualifying MS-characteristic T2 lesion in the brainstem or cerebellar infratentorial compartment. Symptomatic and asymptomatic lesions can count in the 2017 framework. | Correlate the lesion with clinical findings and characterize morphology and alternative causes. Count the compartment once toward DIS and do not infer a clinical attack solely from an MRI lesion. | An infratentorial lesion does not by itself diagnose MS or quantify future relapse, disability or treatment response. Infectious, vascular, inflammatory and structural mimics remain relevant. | Thompson et al. accepted manuscript, Panels 4-5 (infratentorial region and inclusion of symptomatic lesions) and Panel 3 (integration and alternative diagnoses). | ✓ |
| spinal-cord | Spinal cord lesion (2017 DIS region) | One of four 2017 DIS regions. A qualifying T2-hyperintense spinal-cord lesion can contribute the spinal region to attack-onset DIS; the separate 2017 progression-from-onset pathway requires at least two spinal-cord T2 lesions for its spinal component. | Report level, length, morphology and acquisition adequacy, and distinguish the one-lesion attack-onset DIS rule from the two-lesion PPMS component. Do not infer neurologic deficits or a clinical attack from MRI alone. | A cord lesion is clinically important but not specific or a calibrated prognostic measure. Compressive, vascular, metabolic, infectious, NMOSD, MOGAD and other inflammatory causes require clinical exclusion as applicable. | Thompson et al. accepted manuscript, Panel 1 spinal-cord lesion definition, Panel 5 attack-onset DIS, and Panel 6 PPMS requirement of at least two cord lesions for that component. | ✓ |
| DIS | Dissemination in space demonstrated | For MRI DIS in a 2017-eligible attack-onset presentation, demonstrate at least one MS-characteristic T2 lesion in at least two of four regions: periventricular, cortical or juxtacortical, infratentorial, and spinal cord. A clinical attack at a different CNS site can also establish DIS in the applicable pathway. | Name every qualifying region and the evidence source, and continue to the correct attack-onset or progression-from-onset pathway. DIS alone is not a treatment indication and does not remove the requirement for DIT or an allowed substitute where that pathway requires it. | DIS is a binary diagnostic criterion, not a lesion-burden severity score or patient-specific prognosis. Its positive predictive value falls when applied outside the typical CIS population in which the criteria were validated. | Thompson et al. accepted manuscript, Panel 5 and Table 1 (MRI and clinical DIS routes), Panels 2-3 (validation population and misdiagnosis safeguards). | ✓ |
| DIT | Dissemination in time demonstrated | 2017 DIT can be established by a second clinical attack, simultaneous gadolinium-enhancing and nonenhancing lesions at any time, or a new T2-hyperintense or gadolinium-enhancing lesion on follow-up compared with baseline regardless of baseline timing. Symptomatic and asymptomatic lesions can count. | State the exact DIT route and comparison date. If neither a valid clinical nor MRI route exists, evaluate whether the narrowly eligible CSF-OCB substitution applies; otherwise keep DIT unmet. | DIT establishes temporal dissemination but does not quantify relapse rate, disability trajectory or treatment benefit. Enhancing and nonenhancing lesions can have mimics and do not establish MS without the remaining gates. | Thompson et al. accepted manuscript, Panel 5 (two MRI DIT alternatives and symptomatic-lesion rule), Table 1 (clinical attack and OCB routes). | ✓ |
| CSF-OCB | CSF-specific oligoclonal bands (eligible 2017 DIT substitute) | At least two oligoclonal IgG bands restricted to CSF, demonstrated with paired serum and CSF using an appropriate standardized method. In a typical CIS with DIS, no better explanation and no atypical CSF findings, this can substitute for demonstration of DIT in the 2017 attack-onset pathway. | Verify that bands are CSF-specific and review the complete CSF profile before using the substitution. OCB positivity does not replace differential diagnosis, and it is not itself literal evidence of lesions arising at different times. | CSF-specific OCBs support intrathecal antibody synthesis but are not specific for MS and are not a numeric prognosis. Markedly atypical protein, cell count or cell types argue for alternate disease assessment. | Thompson et al. accepted manuscript, CSF discussion and Panel 4 (at least two CSF-specific bands, paired testing, atypical CSF warning), Table 1 footnote stating OCBs substitute for but do not demonstrate DIT. | ✓ |
| attack-onset-MS | 2017 attack-onset MS criteria fulfilled | The appropriate 2017 attack-onset row is fully satisfied: the number of clinical attacks and objective lesion sites supplies or is supplemented by the required DIS and DIT evidence, and no better explanation remains. One attack and one objective lesion requires both DIS and DIT; one attack and at least two objective lesions requires DIT; two attacks with one objective lesion requires DIS. | Have a clinician with MS expertise integrate history, examination, MRI, CSF and differential diagnosis, document the exact pathway, and separately assign disease course and activity. The criterion result does not choose disease-modifying therapy. | Meeting the 2017 diagnostic framework does not predict an individual's relapse rate, disability, treatment response or prognosis. Misapplication outside typical presentations can reduce specificity. | Thompson et al. accepted manuscript, Table 1 complete attack-onset matrix and final no-better-explanation rule; Panel 3 clinician-integration and misdiagnosis safeguards. | ✓ |
| possible-MS | Possible MS under the 2017 framework | Use the 2017 term possible MS when a typical clinically isolated syndrome raises suspicion for MS but the complete McDonald 2017 requirements are not met and no better diagnosis has yet been established. | Identify the missing gate and use appropriate clinical follow-up, comparison MRI, spinal imaging, CSF assessment or specialist review rather than converting uncertainty into definite MS or starting a treatment solely from the label. | Possible MS is an uncertainty state, not a quantified probability of conversion and not a severity grade. The actual likelihood depends on evidence and population factors outside this label. | Thompson et al. accepted manuscript, Table 1 conclusion: incomplete criteria in suspected typical CIS are termed possible MS; Panel 3 recommends follow-up and caution for non-classical presentations. | ✓ |
| not-MS | Not MS because another diagnosis better explains the presentation | Use the 2017 not-MS state when another diagnosis arising during evaluation better explains the clinical presentation. Merely failing to meet McDonald criteria is not sufficient to declare not MS. | Name and pursue the better explanation and avoid anchoring on the MS framework. If evidence is simply insufficient and no better diagnosis is established, use the appropriate uncertainty state rather than not-MS. | This is a diagnostic disposition, not a risk stratum. It does not quantify the chance that later evidence will change the diagnosis. | Thompson et al. accepted manuscript, Table 1 final paragraph distinguishing MS, possible MS and not MS according to completion and better explanation. | ✓ |
| PPMS | 2017 primary progressive MS criteria fulfilled | 2017 progression-from-onset pathway: at least one year of disability progression independent of relapse plus at least two of three components: one or more characteristic brain T2 lesions in a periventricular, cortical or juxtacortical, or infratentorial region; at least two spinal-cord T2 lesions; or CSF-specific oligoclonal bands, with no better explanation. | Document progression duration and each of the two qualifying components, obtain strong clinical and differential-diagnosis review, and do not infer the progression history from imaging. The diagnostic code does not choose therapy. | PPMS is a disease-course diagnostic pathway, not a numeric severity or future-disability calculator. Other causes of progressive myelopathy or neurologic decline must be excluded. | Thompson et al. accepted manuscript, Panel 6 (one year progression plus two of three brain, spinal and CSF components) and Panel 3 recommendation for extra caution and CSF evaluation in progressive presentations. | ✓ |

### Citações por categoria
- **periventricular**: Thompson AJ, Banwell BL, Barkhof F, et al.. Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria (2018) — https://discovery.ucl.ac.uk/id/eprint/10041020 · Thompson et al. 2017 accepted manuscript, Panel 5 (periventricular as one of four regions and two-region DIS rule), Panel 5 footnote on older patients or vascular risk, and Panel 3 misdiagnosis safeguards.
- **cortical-juxtacortical**: Thompson AJ, Banwell BL, Barkhof F, et al.. Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria (2018) — https://discovery.ucl.ac.uk/id/eprint/10041020 · Thompson et al. accepted manuscript, Panel 4 and cortical-lesion discussion (cortical added alongside juxtacortical; artifact caution), Panel 5 regional list.
- **infratentorial**: Thompson AJ, Banwell BL, Barkhof F, et al.. Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria (2018) — https://discovery.ucl.ac.uk/id/eprint/10041020 · Thompson et al. accepted manuscript, Panels 4-5 (infratentorial region and inclusion of symptomatic lesions) and Panel 3 (integration and alternative diagnoses).
- **spinal-cord**: Thompson AJ, Banwell BL, Barkhof F, et al.. Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria (2018) — https://discovery.ucl.ac.uk/id/eprint/10041020 · Thompson et al. accepted manuscript, Panel 1 spinal-cord lesion definition, Panel 5 attack-onset DIS, and Panel 6 PPMS requirement of at least two cord lesions for that component.
- **DIS**: Thompson AJ, Banwell BL, Barkhof F, et al.. Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria (2018) — https://discovery.ucl.ac.uk/id/eprint/10041020 · Thompson et al. accepted manuscript, Panel 5 and Table 1 (MRI and clinical DIS routes), Panels 2-3 (validation population and misdiagnosis safeguards).
- **DIT**: Thompson AJ, Banwell BL, Barkhof F, et al.. Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria (2018) — https://discovery.ucl.ac.uk/id/eprint/10041020 · Thompson et al. accepted manuscript, Panel 5 (two MRI DIT alternatives and symptomatic-lesion rule), Table 1 (clinical attack and OCB routes).
- **CSF-OCB**: Thompson AJ, Banwell BL, Barkhof F, et al.. Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria (2018) — https://discovery.ucl.ac.uk/id/eprint/10041020 · Thompson et al. accepted manuscript, CSF discussion and Panel 4 (at least two CSF-specific bands, paired testing, atypical CSF warning), Table 1 footnote stating OCBs substitute for but do not demonstrate DIT.
- **attack-onset-MS**: Thompson AJ, Banwell BL, Barkhof F, et al.. Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria (2018) — https://discovery.ucl.ac.uk/id/eprint/10041020 · Thompson et al. accepted manuscript, Table 1 complete attack-onset matrix and final no-better-explanation rule; Panel 3 clinician-integration and misdiagnosis safeguards.
- **possible-MS**: Thompson AJ, Banwell BL, Barkhof F, et al.. Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria (2018) — https://discovery.ucl.ac.uk/id/eprint/10041020 · Thompson et al. accepted manuscript, Table 1 conclusion: incomplete criteria in suspected typical CIS are termed possible MS; Panel 3 recommends follow-up and caution for non-classical presentations.
- **not-MS**: Thompson AJ, Banwell BL, Barkhof F, et al.. Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria (2018) — https://discovery.ucl.ac.uk/id/eprint/10041020 · Thompson et al. accepted manuscript, Table 1 final paragraph distinguishing MS, possible MS and not MS according to completion and better explanation.
- **PPMS**: Thompson AJ, Banwell BL, Barkhof F, et al.. Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria (2018) — https://discovery.ucl.ac.uk/id/eprint/10041020 · Thompson et al. accepted manuscript, Panel 6 (one year progression plus two of three brain, spinal and CSF components) and Panel 3 recommendation for extra caution and CSF evaluation in progressive presentations.

## Histórico de versões

| Data | Evento | Detalhe | Situação |
| --- | --- | --- | --- |
| 2026-08-12 | revised | Monitored source changed (content_hash). Detected automatically; awaiting reviewer confirmation. | needs_review |
| 2026-07-30 | revised | Monitored source changed (content_hash). Detected automatically; awaiting reviewer confirmation. | needs_review |
| 2026-07-25 | revised | Monitored source changed (content_hash). Detected automatically; awaiting reviewer confirmation. | needs_review |
| 2025-09-18 | revised | A newer 2024 revision was published in The Lancet Neurology. The historical 2017 map must not be presented as the current diagnostic standard or mixed with the new optic-nerve, CSF and MRI biomarker rules. | needs_review |
| 2018-02-01 | revised | 2017 revisions of the McDonald criteria published in Lancet Neurology. | confirmed |
| 2017-12-21 | published | The 2017 revisions of the McDonald criteria were published online by the International Panel on Diagnosis of Multiple Sclerosis. | confirmed |


---
> Conteúdo de referência reescrito. Confira a publicação primária vigente. Não é dispositivo médico nem substitui o julgamento clínico. O radiologista responsável pelo laudo permanece o autor e o responsável.

Página HTML: https://radcommons.laudos.ai/systems/mcdonald-2017 · JSON da API: https://radcommons.laudos.ai/api/v1/systems/mcdonald-2017 · Índice para agentes: https://radcommons.laudos.ai/llms.txt