Rasmussen Encephalitis — Comprehensive Research Report

Disease: Rasmussen Encephalitis (RE) · also "Rasmussen syndrome" (the field is drifting toward "syndrome") MONDO: MONDO:0016019 (label: Rasmussen subacute encephalitis, verified via local OAK) Category: Complex (immune-mediated, one-hemisphere neuroinflammatory epilepsy)

The one-sentence version: RE is a rare, sneaky, one-sided brain fire. A T-cell–driven inflammatory process sets up shop in a single cerebral hemisphere of a previously healthy kid, and over months to a couple of years it burns that hemisphere down — relentless focal seizures, a slowly failing arm and leg, a shrinking half-brain on MRI. It's the rare case in neurology where the cure for the seizures is to functionally disconnect half the brain, because the other half is fine and the sick half won't stop.

1. Disease Information

RE is a rare, progressive, chronic encephalitis of unknown trigger that is almost always confined to one cerebral hemisphere. The 2024 comprehensive review frames it plainly: "Rasmussen encephalitis (RE) is a rare and progressive form of chronic encephalitis that typically affects one hemisphere of the brain and primarily occurs in pediatric individuals" and is "characterized by recurrent seizures, delayed developmental milestones, progressive cognitive decline, neuroimaging with chronic inflammatory changes, and progressive hemispheric atrophy" (Medicina 2024, PMID:39597043 — verify).

First described by Theodore Rasmussen and colleagues in 1958 (Rasmussen, Olszewski, Lloyd-Smith, Neurology 1958).

Key identifiers: | Resource | ID | |---|---| | MONDO | MONDO:0016019 (Rasmussen subacute encephalitis) | | Orphanet | ORPHA:511 | | OMIM | None — RE is not a Mendelian disorder, so it has no gene-anchored OMIM entry (this is itself informative for curation) | | ICD-10 | G04.8 (other encephalitis) is the closest; often coded under G40.x epilepsy in practice | | ICD-11 | 8A45 region (autoimmune/inflammatory CNS) — closest bucket, no dedicated code | | MeSH | No dedicated descriptor; "Rasmussen" appears as an entry term under Encephalitis / Epilepsy |

Synonyms / alternative names: Rasmussen syndrome, Rasmussen's encephalitis, chronic focal encephalitis, chronic progressive epilepsia partialis continua of childhood (Kozhevnikov epilepsy is a historically overlapping term), Rasmussen subacute encephalitis.

Data provenance: RE knowledge is disease-level / aggregated — case series, small cohorts, surgical registries, and consensus statements. There is no large EHR or biobank cohort; the disease is too rare (see §9).

2. Etiology

Bottom line: the trigger is unknown; the effector is the immune system. RE is best understood as a T-cell–mediated (cell-mediated) autoimmune/neuroinflammatory process. The 2022 pathogenesis review summarizes five historically-proposed mechanisms — "virus infection, antibody-mediated degeneration, cell-mediated immunity, microglia-induced degeneration, and genetic mutations" (PMID:36189924 — verify) — but current consensus places cytotoxic CD8⁺ T-cell–mediated cell death at the center.

Causal factors: - Autoimmune / cell-mediated (primary): Clonally expanded, antigen-driven cytotoxic CD8⁺ T cells attack neurons and astrocytes (see §6). The autoantigen has never been definitively identified. - Antibody-mediated (historical, now largely discredited as primary): Anti-GluR3 (GluA3, AMPA-receptor subunit) antibodies were reported in the 1990s but failed replication ("Absence of antibodies to GluR3 in Rasmussen encephalitis"). Other antibodies (anti-NMDAR GluNε2/NR2A, anti-Munc18) have been reported sporadically but are not diagnostic. - Viral (unproven trigger hypothesis): No consistent pathogen. A "hit-and-run" viral trigger or molecular mimicry is hypothesized; one study found RE tissue shows relatively lower IFN-β production and enhanced cytotoxic T-cell activation upon herpesvirus infection (PMC8957159 — verify), keeping a viral-trigger idea alive but unproven.

Risk factors: - Genetic: No causal Mendelian gene. Whole-exome sequencing has looked for susceptibility contributors — a WES study reported candidate variants in immune-related genes, but nothing replicated or diagnostic (Frontiers in Neuroscience 2021, PMC8523672 — verify). RE is sporadic; familial recurrence is essentially not seen. - Environmental / demographic: The dominant non-genetic "risk factor" is simply young age (childhood, peak ~6 yrs). No confirmed toxin, occupational, or lifestyle exposure. No sex predilection in most series. - Prior insult: A subset report a preceding infection/inflammatory event weeks–months before onset, feeding the trigger hypothesis, but this is not consistent.

Protective factors: None established (genetic or environmental). This is a section to mark not available.

Gene–environment interactions: Speculative only — the leading model is that an environmental trigger (?viral) initiates an antigen-specific T-cell response in a susceptible immune background, but no concrete GxE interaction is documented. Mark hypothesized / not established.

3. Phenotypes

RE's phenotype is a hemispheric syndrome that marches: it starts focal and escalates into progressive loss of every function that half-brain supports.

Characteristics: - Age of onset: childhood, "median age of 6 years"; "around 10% of the individuals affected by RE are young adults" (PMID:39597043 — verify). Largest series range: 14 months–14 yr. - Severity/progression: progressive and typically severe if untreated; "Granata et al. found that 30% of the patients were wheelchair-bound within three years of the diagnosis" (PMID:39597043 — verify). - EPC definition (quotable, ILAE via review): "recurrent focal motor seizures (typically affecting hand and face...) that occur every few seconds minutes for extended periods (days or years)".

QoL impact: severe — intractable seizures, permanent hemiplegia, hemianopia, language loss (dominant side), cognitive decline; most patients end up with a fixed major neurological disability even after successful seizure control. Formal EQ-5D/SF-36 data are essentially absent (rare disease); mark limited data.

4. Genetic / Molecular Information

RE is not a genetic disease in the Mendelian sense — this is a firm "not applicable" for most subfields.

• Causal genes: None. No OMIM gene entry.
• Pathogenic variants: None established as causal. WES surveys have proposed immune-gene candidates without replication (PMC8523672 — verify). No ClinVar/HGMD pathogenic variant set. Somatic mosaicism has been looked for (given the strict unilaterality, a somatic brain mutation is an attractive hypothesis) but not confirmed.
• Somatic vs germline: The striking one-hemisphere restriction has driven a somatic-mutation hypothesis, but no recurrent somatic driver has been demonstrated.
• Modifier genes / epigenetics / chromosomal abnormalities: No established data. Mark not available.

The molecular action is at the immune-effector / expression level, not the germline-variant level — see §6.

5. Environmental Information

• Environmental/toxic factors: None established.
• Lifestyle factors: None established (childhood disease).
• Infectious agents: No confirmed causative pathogen. The viral-trigger hypothesis persists — candidate/investigated agents historically include herpesviruses (HSV, HHV-6, EBV, CMV), enteroviruses — but PCR/serology are inconsistent and RE is not an active viral encephalitis. Best curated as hypothesized trigger, unconfirmed. NCBI Taxonomy anchors would be speculative; recommend omit rather than over-assert.

6. Mechanism / Pathophysiology

This is the load-bearing section. The causal chain, upstream → downstream:

Unknown trigger → antigen-driven CD8⁺ T-cell response → cytotoxic killing of neurons + astrocytes (granzyme B/perforin) → astrocyte loss + microglial activation + neuronophagia → cortical neuronal loss → progressive hemiatrophy → intractable focal epilepsy + progressive hemispheric deficits.

Immune effectors (the engine): - Cytotoxic CD8⁺ T lymphocytes dominate the infiltrate and are clonally expanded / antigen-driven: "Most infiltrating lymphocytes in RE are cytotoxic T-cells, and long-lived clonal populations of cytotoxic T-cells were found in immunohistochemistry analysis of patients with RE" (PMID:39597043 — verify). Large-scale TCR sequencing confirmed pathogenic CD8⁺ expansion (Schneider-Hohendorf et al., Nat Commun 2016, PMC4822013 — verify), and peripheral CD8⁺ T-cell expansion correlates with disease severity. - Granzyme B / perforin killing: infiltrating T cells polarize and release cytotoxic granules onto targets. - Astrocytes as a specific target (key mechanistic paper): Bauer et al. (Ann Neurol 2007, PMID:17503512) — "Astrocytic apoptosis and subsequent loss of these cells is a specific feature of RE"; "Astrocytes in these tissues showed major histocompatibility complex class I expression"; and "granzyme-B(+) lymphocytes were found in close apposition to astrocytes bordering astrocyte-deficient lesions." Astrocyte loss is now considered a distinctive feature, not just bystander gliosis. - Microglia: activated microglia form nodules and drive neuronophagia. Recent single-cell/immune-microenvironment work (2023–2025) describes a T-cell trajectory from expansion to exhaustion over the disease course (J Neuroinflammation 2025, 10.1186/s12974-025-03477-5 — verify). - Innate/inflammasome: IL-1 and inflammasome activation contribute — "activation of inflammatory pathways and release of cytokines such as IL-1 mediated by CD8+ T cells has also been shown in RE" (PMID:39597043 — verify).

Histopathology (staging correlate): "T-lymphocytic infiltrate, reactive astrocytosis, activated microglia, and neuronophagia leading to neuronal loss and cortical atrophy are seen in brain parenchyma" — with perivascular T-cell cuffing, microglial nodules, and, in late stages, cavitation/gliosis and near-complete neuronal dropout.

Ontology suggestions: - Biological processes (GO): GO:0001913 T cell mediated cytotoxicity; GO:0002418 immune response to tumor cell (no—skip); GO:0006954 inflammatory response; GO:0006915 apoptotic process; GO:0001774 microglial cell activation; GO:0002446 neutrophil-mediated (skip); consider GO:0050900 leukocyte migration (brain infiltration). Modifier: INCREASED for the inflammatory/cytotoxic ones. - Cell types (CL): CL:0000909 CD8-positive, alpha-beta memory T cell / CL:0000625 CD8-positive, alpha-beta T cell; CL:0000129 microglial cell; CL:0000127 astrocyte; CL:0000540 neuron. - Chemical/protein effectors (CHEBI/PR): granzyme B, perforin, IL-1β, IFN-γ (mostly PR, not CHEBI).

Molecular profiling: Transcriptomic and single-cell studies of resected RE cortex exist (recent immune-microenvironment work) but there is no established proteomic/metabolomic/lipidomic signature. Mark advanced-omics as emerging / limited.

7. Anatomical Structures Affected

• Primary organ: brain — one cerebral hemisphere (UBERON:0000955 brain; UBERON:0001869 cerebral cortex; UBERON:0001870 frontal cortex often earliest).
• Lateralization: unilateral / strikingly asymmetric — this is the defining anatomical feature. Perisylvian and frontoinsular cortex are often affected early; the process can spread to involve the whole hemisphere, basal ganglia (esp. caudate atrophy), and sometimes the ipsilateral thalamus.
• Tissue level: gray matter (cortical neurons) and the astrocytic compartment; secondary white-matter and volume loss.
• Cell populations targeted: neurons (CL:0000540), astrocytes (CL:0000127); effectors are CD8⁺ T cells and microglia.
• Subcellular (GO CC): apoptotic machinery (mitochondria GO:0005739), MHC-I at plasma membrane (GO:0042612 MHC class I protein complex).
• Secondary/system involvement: motor system (contralateral hemiplegia), visual pathway (hemianopia), language cortex (aphasia if dominant). No systemic organ involvement — RE stays in the CNS.

8. Temporal Development

Classic three-stage natural history (Bien staging):

1. Prodromal stage — low seizure frequency, mild hemiparesis; can last months (median ~7 months). "The first stage is characterized by mild hemiparesis and seizures, generally occurring in a low frequency."
2. Acute stage — the destructive phase: frequent focal motor seizures / EPC, progressive hemiparesis→hemiplegia, hemianopia, and (dominant side) aphasia, plus cognitive decline. "A few months later, RE individuals tend to exhibit a higher frequency of seizures, presenting as focal motor seizures or EPC... RE patients tend to develop worsening focal deficits, such as hemianopia, hemiplegia, behavioral changes, aphasia, and cognitive deficits." Typically lasts ~8–12 months.

3. Residual (burnt-out) stage — seizure frequency often decreases but fixed, permanent neurological deficits and hemiatrophy remain. "The 'residual stage' is the last stage and is characterized by a decrease in the frequency of seizures and persistent neurological deficits."

4. Onset pattern: subacute-to-chronic, progressive.

5. Course: progressive during the acute phase, then plateaus into a stable deficit-laden residual phase (over ~1–3 yr total to reach residual).
6. Duration: chronic/lifelong disability; the active inflammatory phase is self-limited-ish (burns out) but leaves permanent damage.
7. Critical window: the acute stage is the intervention window — the whole rationale for early immunotherapy and timely surgery is to stop hemispheric destruction before it completes. Adult-onset cases tend to progress more slowly.

9. Inheritance and Population

• Incidence: very rare — "The incidence of RE is estimated at 1.8 to 2.4 out of every 10 million people annually" in those under 18 (PMID:39597043; Bien et al. incidence/therapy study, PMID:23216622 — verify). Roughly ~0.18 per 100,000/yr in children.
• Prevalence: no reliable point-prevalence figure — appropriately an ultra-rare band. Prevalence-class curation: BELOW_1_IN_1000000 / ULTRA_RARE.
• Inheritance: not heritable — sporadic, non-Mendelian. All the genetics subfields (penetrance, anticipation, founder effect, consanguinity, carrier frequency) → not applicable.
• Demographics: primarily children (peak ~6 yr); ~10% adolescent/adult onset. No confirmed sex predominance (some series hint at slight variation, not robust). No ethnic/geographic clustering; worldwide distribution.

10. Diagnostics

Diagnosis is clinical-radiological-pathological, formalized by the Bien European consensus criteria (Bien et al., Brain 2005;128:454–471, PMID:15689357), recently updated by an international modified-Delphi consensus (Stredny et al., Epilepsia 2026, 10.1002/epi.70225 — verify).

Bien criteria (two-part): diagnosis if all three of Part A OR two of three Part B: - Part A: (1) clinical — focal seizures (± EPC) and unilateral cortical deficit; (2) EEG — unihemispheric slowing ± epileptiform activity, unilateral seizure onset; (3) MRI — unihemispheric focal cortical atrophy plus ≥1 of gray/white-matter T2/FLAIR hyperintensity or caudate head hyperintensity/atrophy. - Part B: (1) EPC or progressive unilateral cortical deficit; (2) progressive unihemispheric atrophy on serial MRI; (3) histopathology — T-cell–dominated encephalitis with activated microglia (classically nodules) and reactive astrogliosis. (Presence of numerous parenchymal macrophages, B cells, plasma cells, or viral inclusions argues against RE.)

Key modalities: - MRI (RadLex/neuroimaging): serial MRI is the workhorse — progressive unilateral cortical/insular atrophy, T2/FLAIR hyperintensity, caudate atrophy. Recent multi-institutional work compared MRI and pathology staging to hemispherotomy outcome (Child's Nerv Syst 2024, 10.1007/s00381-024-06353-4 — verify). - EEG / electrophysiology (LOINC/clinical neurophysiology): unihemispheric slowing, multifocal ipsilateral epileptiform discharges, lateralized seizure onset; EPC often has poor EEG correlate. - CSF: may show mild pleocytosis, oligoclonal bands — nonspecific, supportive. - Biopsy/histopathology (SNOMED CT): confirmatory when imaging is atypical; shows the T-cell/microglial encephalitis pattern above. The consensus trend is to minimize biopsy when criteria are otherwise met. - Autoantibodies: anti-GluR3 is not recommended (poor specificity/reproducibility); a broader autoimmune-encephalitis antibody panel is done mainly to exclude mimics. - Genetic/omics testing: not diagnostic; used only to exclude genetic mimics (e.g., mitochondrial disease presenting with EPC, like POLG). - Differential dx: other causes of EPC/hemispheric epilepsy — MELAS/POLG mitochondrial disease, hemimegalencephaly/cortical dysplasia, Sturge-Weber, unihemispheric stroke/vasculitis (e.g., primary CNS angiitis), tumor, and other autoimmune encephalitides.

11. Outcome / Prognosis

• Mortality: low disease-specific mortality; deaths relate to status epilepticus or surgical complications rather than the disease directly. Not a classically "fatal" disease — the burden is disability.
• Morbidity (the real story): near-inevitable progression to permanent hemiplegia, hemianopia, cognitive decline, and (dominant hemisphere) aphasia. Untreated, ~30% wheelchair-bound within 3 years of diagnosis.
• Seizures: medically refractory — antiseizure drugs rarely control EPC.
• Surgical outcome (hemispherectomy/hemispherotomy = the definitive seizure cure): reported seizure-freedom ~81.5%, 63.6%, 55.6% at 1/5/10 yr in one cohort; UCLA cohort ~68%/48%/22% at 1/5/10 yr; meta-analytic 5-yr mean ~65% (range 17–100%) (PMC9514735, PMID:32679562 — verify). The trade-off is a guaranteed contralateral hemiplegia and hemianopia — accepted because that deficit is largely already present or inevitable.
• Prognostic factors: shorter preoperative hemiparesis duration predicted lower seizure-freedom and more reoperation; complete disconnection on postop MRI improved seizure freedom; reoperation for incomplete disconnection was frequently curative. Notably, imaging/pathology stage did not reliably predict individual seizure outcome (Child's Nerv Syst 2024 — verify).
• Recovery/plasticity: young age favors post-hemispherectomy functional (language, ambulation) reorganization — the younger the brain, the better it rewires.

12. Treatment

Two parallel goals: (a) immunotherapy to slow the inflammatory destruction, and (b) surgery to actually stop the seizures. Antiseizure meds are supportive but rarely sufficient.

Immunotherapy (MAXO: MAXO:0000917 immunosuppressive therapy / MAXO:0001211 immunomodulation — verify labels with OAK): - Acute/first-line: IV corticosteroids (methylprednisolone) ± IVIg — Stredny consensus: "Intravenous corticosteroids are recommended as first-line, acute immunotherapy for seizure exacerbations and status epilepticus, with or without the addition of intravenous immunoglobulin." CHEBI: methylprednisolone CHEBI:6888; corticosteroid class NCIT:C2322. - Maintenance: IVIg; tacrolimus (calcineurin inhibitor, T-cell targeted) — Bien RCT compared tacrolimus vs IVIg (PMID:23216622 — verify), both slowing progression; azathioprine (Immunomodulation with Azathioprine, Neurology 2021 — verify). - Targeted/experimental biologics (case series): - Rituximab (anti-CD20 B-cell depletion) — reduced seizure burden in 16/26 patients across pooled reports (PMC9058598; PMID:19657347 — verify). - Natalizumab (anti-α4-integrin, blocks lymphocyte CNS entry) — response in ~10/32 (Neurology 2013 — verify). - Tocilizumab (anti-IL-6R) — high response rate in refractory status epilepticus (16 patients pooled — verify). - Adalimumab (anti-TNF-α) — complete response ~45% in an 11-patient series; clinicaltrials:NCT04003922 (adalimumab efficacy/tolerance). - Cyclophosphamide has been used as an alternative T-cell–directed agent. - Mechanistic note: these biologics map cleanly onto the CD8-T-cell/microglia mechanism — good target_mechanisms candidates linking drug → the cytotoxic-T-cell node.

Surgery (definitive, MAXO: MAXO:0000004 surgical procedure; NCIT hemispherectomy term): - Functional hemispherectomy / hemispherotomy — disconnects (rather than removes) the diseased hemisphere; the gold standard and only reliably seizure-freeing therapy: "the only gold-standard treatment for this disorder is hemispherectomy." Timing is the central clinical dilemma — do it early enough to stop cognitive/seizure damage, but the price is fixed hemiplegia/hemianopia.

Supportive: antiseizure medications (broad-spectrum), rehabilitation (PT/OT/speech — MAXO:0000011 physical therapy), and post-surgical neurorehabilitation to exploit plasticity.

Pharmacogenomics / personalized: none established for RE specifically (tacrolimus dosing follows general CYP3A5 pharmacogenetics, not RE-specific).

13. Prevention

• Primary prevention: none — trigger unknown, not heritable, no vaccine. Mark not applicable.
• Secondary prevention (early detection): the closest real "prevention" is early diagnosis + early immunotherapy to limit hemispheric atrophy during the critical acute window; and timely surgery to prevent further seizure-related cognitive decline. This is disease-modifying, not primary prevention.
• Tertiary prevention: manage refractory seizures/status, prevent injury, rehabilitation to limit disability.
• Screening / genetic counseling / immunization / public health: not applicable (sporadic ultra-rare, no genetic or infectious basis to screen for).

14. Other Species / Natural Disease

• Taxonomy: RE is essentially a human-only clinical entity (NCBITaxon:9606). There is no described naturally-occurring animal equivalent (no OMIA entry).
• Comparative biology: the mechanistic parallel is general CD8⁺ T-cell–mediated CNS autoimmunity, not a species-specific homolog.
• Transmission/zoonosis: not applicable — RE is not infectious or transmissible.

Mark this whole section not applicable / human-specific.

15. Model Organisms

RE has historically been hard to model — no spontaneous animal disease, no single gene to knock out, and the pathology is driven by human T cells against an unknown human antigen. The breakthrough was a humanized mouse model:

• Humanized PBMC-engraftment model (flagship): RE-patient peripheral blood mononuclear cells engrafted into immunodeficient NSG mice (NOD-scid IL2Rγ^null^) — the human CD4⁺/CD8⁺ T cells infiltrate the CNS and reproduce RE-like pathology and seizures. "Numerous granzyme B+CD8+ T lymphocytes were detected in the brains of NSG mice", with elevated IFN-γ⁺/IL-17⁺ human T cells vs controls (JCI 2018, "Humanized mouse model of Rasmussen's encephalitis supports the immune-mediated hypothesis," PMC5919802 — verify PMID). A follow-up showed blocking immune intrusion into the brain suppresses epilepsy in this model (JCI, article 120444 — verify) — strong causal support for the T-cell-effector mechanism and a therapeutic proof-of-concept for CNS-entry blockade (cf. natalizumab).
• Model type: mammalian, humanized/xenograft (immune-cell transfer), NCBITaxon:10090 Mus musculus host.
• Evidence source: MODEL_ORGANISM (with human immune cells → arguably a human-immune xenograft; still tag MODEL_ORGANISM for the mouse readout, and note the humanized design in the explanation).
• Recapitulation: reproduces CNS T-cell infiltration, granzyme-B⁺ CD8 cells, and seizures — good for the immune-effector arm.
• Limitations: does not reproduce the strict unilaterality/hemispheric-restriction of human RE, doesn't identify the autoantigen, and depends on donor-patient cells (not a stable genetic line). This is a solid HUMAN_MODEL_MISMATCH discussion candidate: the model supports the immune-mediated hypothesis but leaves the hemispheric-restriction and antigen questions unresolved.
• Prior to humanized models, RE relied on ex vivo human resected-tissue immunohistochemistry (Bauer 2007) and TCR-repertoire studies — in vitro / patient-tissue rather than true animal models.

Curation notes & flagged uncertainties

• MONDO MONDO:0016019 and ORPHA:511 are solid anchors. No OMIM (non-Mendelian) — expect the entry to lean heavily on pathophysiology + phenotypes + treatments, light on genetic.
• Strongest, most citable primary papers for evidence blocks: Bien 2005 European consensus (PMID:15689357) for definition/criteria/staging; Bauer 2007 (PMID:17503512) for astrocyte-target mechanism (I have verbatim abstract quotes above — those should validate cleanly); Bien 2013 (PMID:23216622) for incidence + tacrolimus/IVIg. Verify each with just fetch-reference before use.
• PMID:39597043 (Medicina 2024 review) is the best single narrative source but is a review — good for framing/synthesis quotes, less ideal than primaries for specific mechanistic claims. The PMID came from a web fetch; confirm it resolves to the right paper.
• Two ontology cautions: HPO's only close "encephalitis" leaf is HP:0002383 Infectious encephalitis, which is wrong for RE — don't force it; RE's "encephalitis" is better captured through the pathophysiology/imaging phenotypes (HP:0100308 cerebral cortical hemiatrophy) than a phenotype term. And confirm the focal-seizure and cognitive-decline HP IDs with OAK before committing (I verified EPC, hemiparesis, hemianopia, aphasia, status epilepticus, hemiatrophy directly).
• Good conforms_to target: epilepsy_excitation_inhibition_imbalance#Excitation-Inhibition Imbalance for the seizure arm — though RE's driver is upstream immune destruction of one hemisphere, so the disease-specific chain (CD8/granzyme-B → astrocyte+neuron loss → hemiatrophy → seizures) is the core content and doesn't fully reduce to that module.

Sources: - Medicina 2024 comprehensive review (PMC11596482) - Bien et al. 2005 European consensus, Brain (PMID:15689357) - Bauer et al. 2007, Astrocytes as immunological target (PMID:17503512) - Bien et al. 2013 incidence / tacrolimus vs IVIg (PMID:23216622) - Progress in pathogenesis and therapy of RE 2022 (PMID:36189924) - Stredny et al. 2026 international consensus, Epilepsia - Humanized mouse model of RE, JCI 2018 (PMC5919802) - Blocking immune intrusion suppresses epilepsy in RE model, JCI - CD8+ T-cell TCR sequencing in RE (PMC4822013) - Genetic factors in RE by WES, Front Neurosci 2021 (PMC8523672) - Dynamics of brain T-cell populations in RE, J Neuroinflammation 2025 - Rituximab in RE (PMC9058598) · Rituximab case (PMID:19657347) - Natalizumab in RE, Neurology 2013 · Azathioprine in RE, Neurology 2021 - Adalimumab trial NCT04003922 - Hemispherectomy seizure outcomes & reoperation (PMC9514735) · UCLA surgical series (PMID:32679562) - MRI/pathology vs hemispherotomy outcome 2024 - NORD — Rasmussen Encephalitis

That's the whole tour — one hemisphere, under siege by its own CD8 cells, and the strange mercy that a young brain can hand its whole job to the other half if you disconnect the sick side in time. Want me to go ahead and scaffold the actual kb/disorders/Rasmussen_Encephalitis.yaml entry from this (with the Bauer/Bien primaries as the first verified evidence blocks), or keep this as a research doc for now?