Dementia Atlas

Data dictionary

The corpus is a graph of typed records validated on every build. This page is generated from the schema itself: the base fields every record shares, then each of the eighteen kinds with its own fields and a real example. Relationship fields hold ids of other records; backlinks are derived, so a link is declared once.

Where the schema lives

Source of truth: src/lib/schema.ts (Zod). Machine-readable JSON Schema: /api/v1/schema.json, published with the open API. Ids match ^[a-z0-9]+(-[a-z0-9]+)*$; dates are YYYY-MM-DD.

Base fields (every kind)

Identity, the two-register text (TL;DR for anyone, summary for specialists), the date the facts were checked, sources, and the relationship arrays that make the graph.

FieldTypeRequiredDescription
identity idyesStable kebab-case identifier; also the URL slug.
namestringyesDisplay name.
akastring[]
default []
noOther names, brands and abbreviations; used by search and hover-linking.
tldrstringyesPlain-language explanation for someone with no background.
summarystringyesTechnical summary. Markdown-light: paragraphs separated by blank lines.
status"approved" | "phase-3" | "phase-2" | "phase-1" | "preclinical" | "concept" | "standard-of-care" | "established" | "emerging" | "historic" | "withdrawn" | "active" | "completed" | "recruiting" | "positive" | "negative" | "mixed" | "planned"noStage or state, from the shared status list.
asOfdate (YYYY-MM-DD)yesDate the facts were last checked.
wikipediaURLnoWikipedia article for the concept.
links{ label, url }[]
default []
noExternal sources: label and URL.
  • label: string
  • url: URL
tagsstring[]
default []
noFree-form tags (for example "spike", "failure").
relatedentity id[]
default []
noFree-form related entity ids (any kind).
cancersentity id[]
default []
noTyped relationship arrays. Each holds ids of the named kind.
sectionsentity id[]
default []
no
technologiesentity id[]
default []
no
targetsentity id[]
default []
no
drugsentity id[]
default []
no
companiesentity id[]
default []
no
institutionsentity id[]
default []
no
pathwaysentity id[]
default []
no
termsentity id[]
default []
no
trialsentity id[]
default []
no
peopleentity id[]
default []
no
bottlenecksentity id[]
default []
noBottlenecks of the war on cancer this object bears on (ideas attack them; technologies, trials, collections relieve them).
keyPapersentity id[]
default []
noKey papers (kind "paper") this object rests on or is discussed in.
journalsentity id[]
default []
noJournals (kind "journal") this object was published in or is tied to.
dependsOnentity id[]
default []
no
notesstring[]
default []
noWhy this matters / what is hard about it. Optional editorial notes.
simplestringnoEven simpler explanation (about a 12-year-old reading age). Optional.
provenance{ editedBy, editedOn, note }noWho last substantively edited this record (name or handle) and when.
  • editedBy: string
  • editedOn: date (YYYY-MM-DD)
  • note: string (optional)
confidence{ probability, by, on, note }noNamed probability estimate for speculative content.
  • probability: [number, number]
  • by: string
  • on: date (YYYY-MM-DD)
  • note: string (optional)

Diagnosis

Diagnosis8 diagnoses

One page per dementia type or related condition: what it is, how it is assessed, and what is known.

FieldTypeRequiredDescription
kind"cancer"yesDiscriminator; fixed per kind.
groupstringyesICD-O / organ system grouping, e.g. "breast", "lung", "blood".
burdenstringnoRoughly how common, in words.
subtypesstring[]
default []
noNamed subtypes a clinician distinguishes.
biomarkersstring[]
default []
noKey biomarkers a clinician tests for.
standardOfCare{ setting, approach, refs, guideline }[]
default []
noOne row per clinical setting: the approach, the products it references, and the guideline mapping.
  • setting: string
  • approach: string
  • refs: entity id[] (optional)
  • guideline: { nccn, esmoMcbs, version, url } (optional)
stateOfArtstring[]
default []
noWhat is state of the art today, in one paragraph per point.
history{ year, title, note, refs }[]
default []
noDated milestones for the timeline, each pointing at the records it concerns.
  • year: integer | string
  • title: string
  • note: string (optional)
  • refs: entity id[] (optional)
pipelineentity id[]
default []
noWhat is coming: ids of drugs/technologies/trials/ideas.
openProblemsstring[]
default []
noWhat is still unsolved, one paragraph each.
basics{ symptoms, diagnosis, staging, sources }no
  • symptoms: string[] (optional)
  • diagnosis: string[] (optional)
  • staging: string[] (optional)
  • sources: { label, url }[] (optional)
prognosis{ text, sources }noPlain-English outlook: what moves survival and the headline population figures, with sources. Shown behind the survival disclosure.
  • text: string
  • sources: { label, url }[] (optional)
parententity idnoThe broader cancer this record is a subtype of (pleural mesothelioma -> mesothelioma); the parent page lists its subtypes at the top.
Example: alzheimers-disease (arrays trimmed to three items, long strings shortened)
{
  "id": "alzheimers-disease",
  "kind": "cancer",
  "name": "Alzheimer’s disease",
  "tldr": "Alzheimer’s disease is a brain disease that commonly causes dementia. Early-stage anti-amyloid medicines exist for a carefully selected group, and several ol...",
  "summary": "Alzheimer’s disease is defined pathologically by amyloid plaques and tau tangles. Clinically it often begins with memory of recent events, then spreads to la...",
  "asOf": "2026-09-21",
  "wikipedia": "https://en.wikipedia.org/wiki/Alzheimer%27s_disease",
  "links": [
    {
      "label": "FDA: Leqembi traditional approval",
      "url": "https://www.fda.gov/news-events/press-announcements/fda-converts-novel-alzheimers-disease-treatment-traditional-approval"
    },
    {
      "label": "FDA: Kisunla approval",
      "url": "https://www.fda.gov/drugs/news-events-human-drugs/fda-approves-treatment-adults-alzheimers-disease"
    },
    {
      "label": "NICE NG97",
      "url": "https://www.nice.org.uk/guidance/ng97"
    }
  ],
  "tags": [
    "dementia-atlas"
  ],
  "related": [
    "mild-cognitive-impairment",
    "mixed-dementia",
    "da-roadmap-anti-amyloid"
  ],
  "sections": [
    "da-front-anti-amyloid",
    "da-front-anti-tau",
    "da-front-imaging"
  ],
  "technologies": [
    "da-tech-anti-amyloid-mab",
    "da-tech-amyloid-pet",
    "da-tech-csf-biomarkers"
  ],
  "targets": [
    "da-target-amyloid-beta",
    "da-target-tau",
    "da-target-ache"
  ],
  "drugs": [
    "da-lecanemab",
    "da-donanemab",
    "da-donepezil"
  ],
  "companies": [
    "da-eisai",
    "da-biogen",
    "da-lilly"
  ],
  "institutions": [
    "da-nia",
    "da-ucl-dementia"
  ],
  "pathways": [
    "da-pathway-amyloid-tau"
  ],
  "terms": [
    "da-term-aria",
    "da-term-amyloid-pet"
  ],
  "trials": [
    "da-trial-clarity-ad",
    "da-trial-trailblazer-alz2",
    "da-trial-engage"
  ],
  "people": [
    "alois-alzheimer"
  ],
  "bottlenecks": [
    "da-bn-late-diagnosis",
    "da-bn-trial-exclusion",
    "da-bn-subtype-misclassification"
  ],
  "group": "dementia",
  "burden": "Alzheimer’s disease is the dementia type this atlas covers in the most depth, including early-stage anti-amyloid medicines and several symptom treatments.",
  "subtypes": [
    "Mild cognitive impairment due to Alzheimer’s disease",
    "Mild dementia due to Alzheimer’s disease",
    "Moderate to severe dementia of the Alzheimer’s type"
  ],
  "biomarkers": [
    "Amyloid PET",
    "CSF amyloid and tau",
    "Plasma amyloid and p-tau research assays"
  ],
  "standardOfCare": [
    {
      "setting": "Assessment",
      "approach": "Clinical diagnosis first. Confirm amyloid pathology before starting an anti-amyloid antibody. Discuss bleeding risk, MRI access, infusion logistics, and ApoE...",
      "refs": [
        "da-tech-amyloid-pet",
        "da-tech-csf-biomarkers",
        "da-lecanemab"
      ],
      "guideline": {
        "version": "NICE NG97; FDA labels",
        "url": "https://www.nice.org.uk/guidance/ng97"
      }
    },
    {
      "setting": "Early symptomatic Alzheimer’s disease with confirmed amyloid",
      "approach": "Where licensed and accessible, lecanemab or donanemab may be considered with MRI monitoring. These are not cures and do not restore lost function.",
      "refs": [
        "da-lecanemab",
        "da-donanemab",
        "da-trial-clarity-ad"
      ],
      "guideline": {
        "version": "NICE NG97; see FDA labels on the product pages",
        "url": "https://www.nice.org.uk/guidance/ng97"
      }
    },
    {
      "setting": "Cognitive symptoms across stages",
      "approach": "Donepezil, rivastigmine, or galantamine, and memantine in moderate to severe disease, as individually indicated. These are symptom treatments.",
      "refs": [
        "da-donepezil",
        "da-rivastigmine",
        "da-galantamine"
      ],
      "guideline": {
        "version": "NICE NG97",
        "url": "https://www.nice.org.uk/guidance/ng97"
      }
    }
  ],
  "stateOfArt": [
    "Two amyloid-lowering antibodies have traditional or full US approval for early symptomatic Alzheimer’s disease after amyloid confirmation.",
    "ARIA (imaging swelling or bleeding) is a boxed risk and needs a monitoring plan before the first infusion.",
    "Most people living with Alzheimer’s disease are not candidates for these antibodies because of stage, comorbidities, imaging, or access."
  ],
  "history": [
    {
      "year": 1984,
      "title": "Amyloid described as a core plaque protein",
      "note": "The amyloid cascade became a dominant research frame.",
      "refs": [
        "da-target-amyloid-beta",
        "da-pathway-amyloid-tau"
      ]
    },
    {
      "year": 1993,
      "title": "Tacrine, then safer cholinesterase inhibitors",
      "note": "Symptom treatments entered routine use.",
      "refs": [
        "da-donepezil"
      ]
    },
    {
      "year": 2021,
      "title": "Aducanumab accelerated approval",
      "note": "Later withdrawn. Included here so a past approval is not mistaken for a current option.",
      "refs": [
        "da-aducanumab",
        "da-trial-engage"
      ]
    }
  ],
  "pipeline": [
    "da-remternetug",
    "da-tech-tau-therapeutics",
    "da-idea-blood-first-triage"
  ],
  "openProblems": [
    "Who benefits enough to justify ARIA risk, infusion burden, and cost remains an individual clinical judgement.",
    "People with mixed or atypical presentations are under-represented in antibody trials.",
    "There is still no approved disease-modifying treatment once dementia is moderate or severe."
  ],
  "basics": {
    "symptoms": [
      "Recent-memory difficulty is common early. Language, wayfinding, and judgement may follow.",
      "Sudden new weakness, speech loss, or fluctuating alertness is not typical Alzheimer’s disease and needs urgent care."
    ],
    "diagnosis": [
      "A specialist may combine history, cognitive tests, MRI, and amyloid confirmation when an antibody is being considered.",
      "ApoE ε4 status changes the ARIA-risk conversation. It is not a diagnosis on its own."
    ],
    "staging": [
      "Antibody labels speak to mild cognitive impairment or mild dementia. Symptom medicines cover a wider range."
    ],
    "sources": [
      {
        "label": "FDA Leqembi",
        "url": "https://www.fda.gov/news-events/press-announcements/fda-converts-novel-alzheimers-disease-treatment-traditional-approval"
      },
      {
        "label": "NIA",
        "url": "https://www.nia.nih.gov/health/alzheimers-and-dementia"
      }
    ]
  },
  "parent": "dementia"
}

Front

Front11 fronts

The fronts of dementia research and care: imaging, biomarkers, treatments, and how care is delivered.

FieldTypeRequiredDescription
kind"section"yesDiscriminator; fixed per kind.
orderintegeryesPosition of the front in the navigation.
iconstringnoIcon id for the front.
Example: da-front-symptomatic-care (arrays trimmed to three items, long strings shortened)
{
  "id": "da-front-symptomatic-care",
  "kind": "section",
  "name": "Symptomatic care",
  "tldr": "Medicines and non-drug support aimed at symptoms, not at reversing the disease.",
  "summary": "Cholinesterase inhibitors and memantine can help some people. Communication, routine, pain, sleep, and carer support often move quality of life more than a p...",
  "asOf": "2026-09-21",
  "links": [
    {
      "label": "NICE NG97",
      "url": "https://www.nice.org.uk/guidance/ng97"
    }
  ],
  "tags": [
    "dementia-atlas"
  ],
  "related": [
    "da-tech-cholinesterase-inhibitors"
  ],
  "order": 4
}

Technology

Technology10 technologies

Every way we see, measure, or support the brain, explained.

FieldTypeRequiredDescription
kind"technology"yesDiscriminator; fixed per kind.
principlestringyesHow it works, for a technical reader.
generationstringnoWhich generation of the idea this is, if the field talks in generations.
strengthsstring[]
default []
noWhat it does well.
limitationsstring[]
default []
noWhere it falls short.
sinceinteger | stringnoYear first used in humans or first approved, if meaningful.
Example: da-tech-cholinesterase-inhibitors (arrays trimmed to three items, long strings shortened)
{
  "id": "da-tech-cholinesterase-inhibitors",
  "kind": "technology",
  "name": "Cholinesterase inhibitors",
  "tldr": "These older medicines can support cognitive symptoms for some people. They are not disease-modifying.",
  "summary": "Donepezil, rivastigmine, and galantamine are the core examples. Gastrointestinal effects and slow heart rate need review.",
  "asOf": "2026-09-21",
  "links": [
    {
      "label": "NICE NG97",
      "url": "https://www.nice.org.uk/guidance/ng97"
    }
  ],
  "tags": [
    "dementia-atlas"
  ],
  "sections": [
    "da-front-symptomatic-care"
  ],
  "drugs": [
    "da-donepezil",
    "da-rivastigmine",
    "da-galantamine"
  ],
  "principle": "Increase acetylcholine signalling at synapses."
}

Target

Target5 targets

The molecules drugs and tracers aim at.

FieldTypeRequiredDescription
kind"target"yesDiscriminator; fixed per kind.
symbolstringnoGene symbol / protein name.
hgncstringno
ensemblstringno
uniprotstringno
entrezstringno
biologystringyesWhat the target does in normal and cancer cells.
whereFoundstring[]
default []
noExpression or alteration by cancer, free text.
targetClass"surface-antigen" | "kinase" | "checkpoint" | "nuclear-receptor" | "enzyme" | "transcription" | "oncogene" | "tumor-suppressor" | "stroma" | "other"
default "other"
noBroad class used for filtering.
prevalence{ cancerId, pct, measure, source, note }[]
default []
noFraction of each cancer that expresses or carries the alteration, sourced. `pct` is 0-100 or a range string like "15-20".
  • cancerId: entity id
  • pct: number | string
  • measure: string (optional)
  • source: URL (optional)
  • note: string (optional)
Example: da-target-ache (arrays trimmed to three items, long strings shortened)
{
  "id": "da-target-ache",
  "kind": "target",
  "name": "Acetylcholinesterase",
  "tldr": "Blocking this enzyme is how donepezil, rivastigmine, and galantamine try to support signalling.",
  "summary": "The approach is symptomatic. It does not clear plaques or tangles.",
  "asOf": "2026-09-21",
  "links": [
    {
      "label": "NICE NG97",
      "url": "https://www.nice.org.uk/guidance/ng97"
    }
  ],
  "tags": [
    "dementia-atlas"
  ],
  "drugs": [
    "da-donepezil",
    "da-rivastigmine",
    "da-galantamine"
  ],
  "symbol": "ACHE",
  "biology": "Enzyme that breaks down acetylcholine in the synaptic cleft.",
  "targetClass": "enzyme"
}

Treatment

Treatment9 drugs

Approved and pipeline products, with what they hit and who makes them.

FieldTypeRequiredDescription
kind"drug"yesDiscriminator; fixed per kind.
brandstringnoBrand name where marketed.
codestringnoDevelopment code(s), comma separated.
modalitystringyesModality, e.g. "ADC", "small molecule", "mAb", "radioligand", "cell therapy", "vaccine", "device".
payloadstringnoCytotoxic payload, for ADCs.
linkerstringnoLinker chemistry, for ADCs.
mechanismstringyesOne or two sentences on how it works.
approvals{ region, year, indication, note }[]
default []
noRegulatory approvals: region, year, indication.
  • region: string
  • year: integer
  • indication: string
  • note: string (optional)
mechanismStepsstring[]
default []
noStep-by-step mechanism for the animated mechanism card.
dosing{ route, schedule, modifications, monitoring, source }noRoute, schedule, modifications and monitoring, from the label.
  • route: string
  • schedule: string
  • modifications: string (optional)
  • monitoring: string (optional)
  • source: URL (optional)
toxicity{ event, anyGradePct, grade3PlusPct, source, note }[]
default []
noAdverse events with rates in percent, from the label or pivotal trial.
  • event: string
  • anyGradePct: number (optional)
  • grade3PlusPct: number (optional)
  • source: URL (optional)
  • note: string (optional)
access{ country, listPrice, reimbursement, assistance, generic, source, asOf }[]
default []
noCost and access by country.
  • country: string
  • listPrice: string (optional)
  • reimbursement: string (optional)
  • assistance: string (optional)
  • generic: boolean (optional)
  • source: URL (optional)
  • asOf: date (YYYY-MM-DD) (optional)
regulatoryEvents{ date, type, region, note, source }[]
default []
noDated designations, filings, decisions and label changes.
  • date: string
  • type: "designation" | "filing" | "pdufa" | "approval" | "crl" | "withdrawal" | "label-change" | "advisory-committee"
  • region: string
  • note: string
  • source: URL (optional)
Example: da-lecanemab (arrays trimmed to three items, long strings shortened)
{
  "id": "da-lecanemab",
  "kind": "drug",
  "name": "Lecanemab",
  "tldr": "Lecanemab is an infused antibody for early Alzheimer’s disease after amyloid is confirmed. It can cause brain swelling or bleeding on MRI and is not a cure.",
  "summary": "US labeling supports starting in mild cognitive impairment or mild dementia of the Alzheimer’s type after amyloid confirmation. Infusions are typically every...",
  "status": "approved",
  "asOf": "2026-09-21",
  "links": [
    {
      "label": "FDA traditional approval",
      "url": "https://www.fda.gov/news-events/press-announcements/fda-converts-novel-alzheimers-disease-treatment-traditional-approval"
    }
  ],
  "tags": [
    "dementia-atlas"
  ],
  "cancers": [
    "alzheimers-disease",
    "mild-cognitive-impairment"
  ],
  "targets": [
    "da-target-amyloid-beta"
  ],
  "companies": [
    "da-eisai",
    "da-biogen"
  ],
  "terms": [
    "da-term-aria"
  ],
  "trials": [
    "da-trial-clarity-ad"
  ],
  "brand": "Leqembi",
  "code": "BAN2401",
  "modality": "mAb",
  "mechanism": "Amyloid-beta–directed antibody that reduces plaques.",
  "approvals": [
    {
      "region": "US",
      "year": 2023,
      "indication": "Early Alzheimer’s disease with confirmed amyloid"
    },
    {
      "region": "JP",
      "year": 2023,
      "indication": "Early Alzheimer’s disease"
    },
    {
      "region": "CN",
      "year": 2024,
      "indication": "Early Alzheimer’s disease"
    }
  ],
  "access": [
    {
      "country": "US",
      "reimbursement": "Coverage depends on plan, registry, and site capability",
      "source": "https://www.fda.gov/news-events/press-announcements/fda-converts-novel-alzheimers-disease-treatment-traditional-approval",
      "asOf": "2026-09-21"
    },
    {
      "country": "UK",
      "reimbursement": "A licence is not the same as NHS commissioning",
      "source": "https://www.nice.org.uk/guidance/ng97",
      "asOf": "2026-09-21"
    }
  ]
}

Company

Company3 companies

Who is building what.

FieldTypeRequiredDescription
kind"company"yesDiscriminator; fixed per kind.
hqstringyesHeadquarters city.
countryISO 3166 country codeyesISO 3166-1 alpha-2 country code.
companyType"pharma" | "biotech" | "diagnostics" | "imaging" | "devices" | "ai-software" | "radiopharma" | "cell-therapy" | "cro-services" | "nonprofit" | "investor"yesBroad type used for filtering and the landscape grid; `investor` marks a fund or corporate venture arm whose portfolio is derived from backlinks.
websiteURLnoOfficial website.
tickerstringnoStock ticker with exchange suffix where needed.
foundedintegernoYear founded.
stage"startup" | "growth" | "public" | "private-large" | "acquired" | "defunct"noBottlenecks: where in the pipeline it bites. Companies: where the company is in its life (startup, growth, public, private-large, acquired, defunct); listed companies default to public.
ycBatchstringnoY Combinator batch, e.g. "W21", "S24", "X26" (Spring), "F25" (Fall).
investorsentity id[]
default []
noIds of investor records (companies with `companyType: "investor"`) that have backed this company. Backlinks give each investor its portfolio.
funding{ round, year, amountUsd, source, note }[]
default []
noSourced financing rounds; omit rather than guess.
  • round: string
  • year: integer
  • amountUsd: number (optional)
  • source: URL
  • note: string (optional)
acquiredByentity idnoCompany id of the acquirer, when `stage` is "acquired" and the acquirer is in OnCo.
Example: da-lilly (arrays trimmed to three items, long strings shortened)
{
  "id": "da-lilly",
  "kind": "company",
  "name": "Eli Lilly",
  "tldr": "Lilly developed donanemab and is running further amyloid-antibody studies.",
  "summary": "Donanemab and remternetug sit on this company’s node so pipeline and approved products stay distinct.",
  "asOf": "2026-09-21",
  "links": [
    {
      "label": "Lilly",
      "url": "https://www.lilly.com/"
    }
  ],
  "tags": [
    "dementia-atlas"
  ],
  "drugs": [
    "da-donanemab",
    "da-remternetug"
  ],
  "trials": [
    "da-trial-trailblazer-alz2",
    "da-trial-expedition3"
  ],
  "hq": "Indianapolis",
  "country": "US",
  "companyType": "pharma",
  "website": "https://www.lilly.com/"
}

Institution

Institution4 institutions

The centres and universities that matter, mapped and ranked.

FieldTypeRequiredDescription
kind"institution"yesDiscriminator; fixed per kind.
citystringyesCity.
countryISO 3166 country codeyesISO 3166-1 alpha-2 country code.
latnumberyesLatitude, for the map.
lngnumberyesLongitude, for the map.
institutionType"cancer-center" | "university" | "hospital" | "research-institute" | "government" | "consortium"yesBroad type used for filtering.
websiteURLnoOfficial website.
nci"comprehensive" | "clinical" | "basic"noNCI designation, US only.
newsweekOncology2026integernoNewsweek / Statista World's Best Specialized Hospitals 2026, Oncology rank.
universitystringnoParent university, when the institution is a cancer centre inside one.
programsstring[]
default []
noNamed programmes or centres of excellence.
Example: da-nia (arrays trimmed to three items, long strings shortened)
{
  "id": "da-nia",
  "kind": "institution",
  "name": "National Institute on Aging",
  "tldr": "The US National Institute on Aging publishes public pages on Alzheimer’s disease, MCI, and caregiving.",
  "summary": "Used throughout this atlas as a public education source. It is not a clinic appointment service.",
  "asOf": "2026-09-21",
  "links": [
    {
      "label": "NIA",
      "url": "https://www.nia.nih.gov/"
    }
  ],
  "tags": [
    "dementia-atlas"
  ],
  "city": "Bethesda",
  "country": "US",
  "lat": 39.002,
  "lng": -77.104,
  "institutionType": "government",
  "website": "https://www.nia.nih.gov/"
}

Pathway

Pathway1 pathways

The classic signalling circuits, drawn and explained.

FieldTypeRequiredDescription
kind"pathway"yesDiscriminator; fixed per kind.
analogystringyesLayman explanation using an analogy.
nodes{ id, label, x, y, targetId }[]yesDiagram nodes with positions; a node may point at a target record.
  • id: string
  • label: string
  • x: number
  • y: number
  • targetId: entity id (optional)
edges{ from, to, type }[]yesDiagram edges: activates or inhibits.
  • from: string
  • to: string
  • type: "activates" | "inhibits" (optional)
interventionsstring[]
default []
noHow drugs attack it.
Example: da-pathway-amyloid-tau (arrays trimmed to three items, long strings shortened)
{
  "id": "da-pathway-amyloid-tau",
  "kind": "pathway",
  "name": "Amyloid and tau cascade (simplified)",
  "tldr": "A simplified drawing of how amyloid and tau are often described in Alzheimer’s disease. Real brains are messier, and mixed pathology is common.",
  "summary": "The cascade is a research frame: amyloid accumulation, then tau spread, then synaptic failure. Antibodies act on amyloid. Tau drugs are mostly still in trial...",
  "asOf": "2026-09-21",
  "links": [
    {
      "label": "NIA",
      "url": "https://www.nia.nih.gov/health/alzheimers-and-dementia"
    }
  ],
  "tags": [
    "dementia-atlas"
  ],
  "cancers": [
    "alzheimers-disease"
  ],
  "analogy": "Think of amyloid as clutter that builds up first, and tau as the wiring that fails later. The drawing is a teaching sketch, not a complete biology textbook.",
  "nodes": [
    {
      "id": "ab",
      "label": "Amyloid-beta",
      "x": 20,
      "y": 40,
      "targetId": "da-target-amyloid-beta"
    },
    {
      "id": "plaque",
      "label": "Plaques",
      "x": 50,
      "y": 40
    },
    {
      "id": "tau",
      "label": "Tau",
      "x": 50,
      "y": 70,
      "targetId": "da-target-tau"
    }
  ],
  "edges": [
    {
      "from": "ab",
      "to": "plaque",
      "type": "activates"
    },
    {
      "from": "plaque",
      "to": "tau",
      "type": "activates"
    },
    {
      "from": "tau",
      "to": "sym",
      "type": "activates"
    }
  ],
  "interventions": [
    "Anti-amyloid antibodies act at plaques.",
    "Cholinesterase inhibitors act later, on signalling, not on this cascade."
  ]
}

Term

Term4 terms

Glossary with plain-English TL;DRs and Wikipedia links.

FieldTypeRequiredDescription
kind"term"yesDiscriminator; fixed per kind.
categorystringyesGlossary category (Endpoints, Biomarkers, Trials, ...).
Example: da-term-aria (arrays trimmed to three items, long strings shortened)
{
  "id": "da-term-aria",
  "kind": "term",
  "name": "ARIA",
  "aka": [
    "amyloid-related imaging abnormalities"
  ],
  "tldr": "ARIA is swelling or bleeding seen on MRI in people receiving amyloid antibodies. It can be silent or, less often, symptomatic.",
  "summary": "ARIA-E is oedema. ARIA-H is haemorrhage. Risk is higher in ApoE ε4 carriers. Monitoring schedules live on the product label, not on this glossary card.",
  "asOf": "2026-09-21",
  "links": [
    {
      "label": "FDA Leqembi",
      "url": "https://www.fda.gov/news-events/press-announcements/fda-converts-novel-alzheimers-disease-treatment-traditional-approval"
    }
  ],
  "tags": [
    "dementia-atlas"
  ],
  "drugs": [
    "da-lecanemab",
    "da-donanemab"
  ],
  "category": "Side effects"
}

Trial

Trial4 trials

Landmark and current trials that define the standard of care.

FieldTypeRequiredDescription
kind"trial"yesDiscriminator; fixed per kind.
nctstringnoPrimary registry id: ClinicalTrials.gov NCT number, or ISRCTN, ACTRN or NTR.
phase"1" | "2" | "3" | "4" | "1/2" | "2/3" | "observational" | "platform"yesTrial phase.
settingstringyesDisease setting and line of therapy.
sponsorstringnoLead sponsor.
resultstringnoHeadline result in one or two sentences, with numbers only if sourced.
yearReportedintegernoYear the primary result was reported.
enrolledintegernoParticipants enrolled.
outcomes{ endpoint, primary, unit, arms, hr, ci, p, source }[]
default []
noStructured outcomes; values in the arm's unit (months, percent). Enables pictograms and comparisons.
  • endpoint: string
  • primary: boolean (optional)
  • unit: string (optional)
  • arms: { name, n, value, note }[]
  • hr: number (optional)
  • ci: [number, number] (optional)
  • p: string (optional)
  • source: URL (optional)
replicationstringnoHas an independent trial or real-world study confirmed the effect?
Example: da-trial-clarity-ad (arrays trimmed to three items, long strings shortened)
{
  "id": "da-trial-clarity-ad",
  "kind": "trial",
  "name": "CLARITY AD",
  "tldr": "CLARITY AD tested lecanemab in early Alzheimer’s disease. It is the main efficacy trial behind the US traditional approval.",
  "summary": "A randomised, placebo-controlled phase 3 study in early symptomatic Alzheimer’s disease with confirmed amyloid. Primary clinical scale results favoured lecan...",
  "status": "completed",
  "asOf": "2026-09-21",
  "links": [
    {
      "label": "ClinicalTrials.gov NCT03887455",
      "url": "https://clinicaltrials.gov/study/NCT03887455"
    },
    {
      "label": "FDA traditional approval",
      "url": "https://www.fda.gov/news-events/press-announcements/fda-converts-novel-alzheimers-disease-treatment-traditional-approval"
    }
  ],
  "tags": [
    "dementia-atlas"
  ],
  "cancers": [
    "alzheimers-disease"
  ],
  "drugs": [
    "da-lecanemab"
  ],
  "companies": [
    "da-eisai",
    "da-biogen"
  ],
  "nct": "NCT03887455",
  "phase": "3",
  "setting": "Early Alzheimer’s disease",
  "sponsor": "Eisai",
  "result": "The trial met its primary clinical endpoint versus placebo. ARIA was the principal safety theme. See the FDA approval communication for the regulator’s reading.",
  "yearReported": 2023,
  "enrolled": 1795,
  "outcomes": [
    {
      "endpoint": "Clinical scale change (primary)",
      "primary": true,
      "unit": "points",
      "arms": [
        {
          "name": "Lecanemab",
          "note": "Favoured versus placebo on the primary scale"
        },
        {
          "name": "Placebo"
        }
      ],
      "source": "https://clinicaltrials.gov/study/NCT03887455"
    }
  ]
}

Pairing

Pairing1 pairings

Things that work better together, and things that do not.

FieldTypeRequiredDescription
kind"pairing"yesDiscriminator; fixed per kind.
aentity idyesFirst member of the pair (entity id).
bentity idyesSecond member of the pair (entity id).
rationalestringyesWhy the pair works better than either alone.
evidencestringyesThe evidence for the pairing, with trial names.
pairingType"combination" | "sequence" | "diagnostic-therapeutic" | "platform" | "caution"yesCombination, sequence, diagnostic-therapeutic, platform or caution.
Example: da-pair-lecanemab-mri (arrays trimmed to three items, long strings shortened)
{
  "id": "da-pair-lecanemab-mri",
  "kind": "pairing",
  "name": "Lecanemab with MRI monitoring",
  "tldr": "Amyloid antibodies are used with scheduled MRI checks for ARIA, not as a standalone infusion without imaging follow-up.",
  "summary": "This pairing is a teaching node: the medicine and the scan belong together. It is not a local protocol.",
  "asOf": "2026-09-21",
  "links": [
    {
      "label": "FDA Leqembi",
      "url": "https://www.fda.gov/news-events/press-announcements/fda-converts-novel-alzheimers-disease-treatment-traditional-approval"
    }
  ],
  "tags": [
    "dementia-atlas"
  ],
  "cancers": [
    "alzheimers-disease"
  ],
  "technologies": [
    "da-tech-mri"
  ],
  "drugs": [
    "da-lecanemab"
  ],
  "terms": [
    "da-term-aria"
  ],
  "a": "da-lecanemab",
  "b": "da-tech-mri",
  "rationale": "ARIA monitoring on MRI is part of using an amyloid antibody, not an optional extra.",
  "evidence": "The US lecanemab label requires MRI monitoring for ARIA. See the FDA approval communication.",
  "pairingType": "diagnostic-therapeutic"
}

Roadmap

Roadmap1 roadmaps

Where a technology has been and where it is heading.

FieldTypeRequiredDescription
kind"roadmap"yesDiscriminator; fixed per kind.
steps{ era, title, description, refs, status }[]yesOrdered eras from history to horizon, each with a status.
  • era: string
  • title: string
  • description: string
  • refs: entity id[] (optional)
  • status: "historic" | "current" | "emerging" | "speculative"
Example: da-roadmap-anti-amyloid (arrays trimmed to three items, long strings shortened)
{
  "id": "da-roadmap-anti-amyloid",
  "kind": "roadmap",
  "name": "Anti-amyloid medicines",
  "tldr": "From failed or withdrawn antibodies to two US-approved early-stage products, with access and ARIA still open.",
  "summary": "This roadmap is a teaching timeline for the Alzheimer’s antibody story on this atlas. It is not a prediction that more antibodies will help more stages.",
  "asOf": "2026-09-21",
  "links": [
    {
      "label": "FDA Leqembi",
      "url": "https://www.fda.gov/news-events/press-announcements/fda-converts-novel-alzheimers-disease-treatment-traditional-approval"
    }
  ],
  "tags": [
    "dementia-atlas"
  ],
  "cancers": [
    "alzheimers-disease"
  ],
  "steps": [
    {
      "era": "2016-2019",
      "title": "Large antibody failures",
      "description": "EXPEDITION 3 and ENGAGE show that amyloid lowering is not automatically clinical benefit.",
      "refs": [
        "da-trial-expedition3",
        "da-trial-engage"
      ],
      "status": "historic"
    },
    {
      "era": "2021-2024",
      "title": "Approvals and a withdrawal",
      "description": "Aducanumab is authorised then withdrawn. Lecanemab and donanemab reach US labels for early disease.",
      "refs": [
        "da-aducanumab",
        "da-lecanemab",
        "da-donanemab"
      ],
      "status": "current"
    },
    {
      "era": "2025-",
      "title": "Access, mixed disease, next targets",
      "description": "Who is treated, who is monitored, and whether tau or blood-first triage changes the funnel.",
      "refs": [
        "da-remternetug",
        "da-idea-blood-first-triage"
      ],
      "status": "emerging"
    }
  ]
}

Idea

Idea3 ideas

Hypotheses and new directions, linked to the evidence.

FieldTypeRequiredDescription
kind"idea"yesDiscriminator; fixed per kind.
hypothesisstringyesThe claim, in one sentence.
rationalestringyesWhy the pair works better than either alone, or why it is a caution.
teststringyesWhat experiment or trial would confirm or kill it.
maturity"speculative" | "preclinical-evidence" | "early-clinical" | "being-tested-at-scale"yesHow far the idea has been tested.
actor"research" | "clinic" | "industry" | "regulator" | "payer" | "policy" | "patients" | "data" | "philanthropy" | "engineering"noWho would have to act: research, clinic, industry, regulator, payer, policy, patients, data, philanthropy.
cost"small" | "medium" | "large"noRough cost to try: small (<$1M), medium ($1-50M), large (>$50M).
horizonYearsintegernoTime to first evidence of impact in years.
Example: da-idea-inclusive-anti-amyloid-trials (arrays trimmed to three items, long strings shortened)
{
  "id": "da-idea-inclusive-anti-amyloid-trials",
  "kind": "idea",
  "name": "Antibody trials that enrol mixed and comorbid disease",
  "tldr": "Test amyloid antibodies in the mixed, comorbid people clinics actually see, instead of only in tidy trial cohorts.",
  "summary": "Until that trial exists, labels and MRI rules should stay conservative. This idea is a proposed test, not a reason to ignore exclusions.",
  "asOf": "2026-09-21",
  "links": [
    {
      "label": "ClinicalTrials.gov",
      "url": "https://clinicaltrials.gov"
    }
  ],
  "tags": [
    "dementia-atlas"
  ],
  "cancers": [
    "alzheimers-disease",
    "mixed-dementia"
  ],
  "drugs": [
    "da-lecanemab",
    "da-donanemab"
  ],
  "bottlenecks": [
    "da-bn-trial-exclusion"
  ],
  "hypothesis": "A pragmatic trial in people with mixed pathology or common comorbidities can show whether antibody benefit and ARIA risk still justify treatment outside the ...",
  "rationale": "Current phase 3 cohorts are cleaner than clinic. Access decisions still hit mixed disease.",
  "test": "Registry-based randomised or well-controlled observational comparison with ARIA and function as co-primary, stratified by vascular burden and anticoagulation.",
  "maturity": "speculative",
  "actor": "research",
  "cost": "large",
  "horizonYears": 8
}

Collection

Collection1 collections

The open databases and registries the field runs on.

FieldTypeRequiredDescription
kind"collection"yesDiscriminator; fixed per kind.
urlURLyesOfficial URL.
holdsstringyesWhat the collection or database holds.
licensestringnoLicence the data are released under.
maintainerstringnoWho maintains it.
Example: da-collection-clinicaltrials (arrays trimmed to three items, long strings shortened)
{
  "id": "da-collection-clinicaltrials",
  "kind": "collection",
  "name": "ClinicalTrials.gov",
  "tldr": "ClinicalTrials.gov is the public US trial registry. A listing is not a recommendation to join a study.",
  "summary": "This atlas links NCT numbers here so readers can check status, sites, and eligibility at the source.",
  "asOf": "2026-09-21",
  "links": [
    {
      "label": "ClinicalTrials.gov",
      "url": "https://clinicaltrials.gov"
    }
  ],
  "tags": [
    "dementia-atlas"
  ],
  "url": "https://clinicaltrials.gov",
  "holds": "Public registrations of interventional and observational studies, including dementia trials.",
  "license": "Public domain (US government)"
}

Person

Person1 people

The clinicians and scientists doing the work: specialisms, bios, papers, and where to find them.

FieldTypeRequiredDescription
kind"person"yesDiscriminator; fixed per kind.
rolestringyesCurrent role and affiliation in one line, e.g. "Chief, Breast Medicine Service".
institutionIdentity idnoPrimary institution id (also list in `institutions`).
specialismsstring[]
default []
noSpecialisms, for filtering.
profiles{ label, url }[]
default []
noProfessional links: profile, lab, Google Scholar, ORCID, PubMed, X/LinkedIn.
  • label: string
  • url: URL
papers{ title, journal, year, url, doi, note }[]
default []
noSelected publications, most recent or most cited first.
  • title: string
  • journal: string (optional)
  • year: integer (optional)
  • url: URL (optional)
  • doi: string (optional)
  • note: string (optional)
orcidstringnoORCID id.
hIndexintegernoh-index, if known and sourced.
Example: alois-alzheimer (arrays trimmed to three items, long strings shortened)
{
  "id": "alois-alzheimer",
  "kind": "person",
  "name": "Alois Alzheimer",
  "tldr": "Alois Alzheimer was a physician whose 1906 case description is why Alzheimer’s disease has that name. This is a historical biography, not a doctor to contact.",
  "summary": "Public encyclopedic biography only. Dementia Atlas does not list living clinicians as a referral directory.",
  "asOf": "2026-09-21",
  "wikipedia": "https://en.wikipedia.org/wiki/Alois_Alzheimer",
  "links": [
    {
      "label": "Wikipedia",
      "url": "https://en.wikipedia.org/wiki/Alois_Alzheimer"
    }
  ],
  "tags": [
    "dementia-atlas"
  ],
  "cancers": [
    "alzheimers-disease"
  ],
  "role": "Physician who described the disease later named after him (historical figure, 1864-1915)",
  "specialisms": [
    "history of medicine"
  ]
}

Bottleneck

Bottleneck4 bottlenecks

The constraints slowing diagnosis, trials, and care, with the ideas that could break each one.

FieldTypeRequiredDescription
kind"bottleneck"yesDiscriminator; fixed per kind.
stage"biology" | "prevention-detection" | "trials" | "regulation-manufacturing" | "access-delivery" | "data-knowledge" | "funding-incentives" | "people-culture"yesWhere in the pipeline it bites.
severity"critical" | "major" | "moderate"yesHow much it slows the whole effort.
metrics{ label, value, source, url }[]
default []
noNumbers that show the size of the problem, each with a source.
  • label: string
  • value: string
  • source: string (optional)
  • url: URL (optional)
causesstring[]
default []
noRoot causes, plain English, one per item.
currentEffortsstring[]
default []
noWhat is already being tried and by whom.
successLooksLikestringnoWhat would count as the bottleneck being broken.
Example: da-bn-subtype-misclassification (arrays trimmed to three items, long strings shortened)
{
  "id": "da-bn-subtype-misclassification",
  "kind": "bottleneck",
  "name": "Wrong-subtype treatment",
  "tldr": "Lewy body, vascular, and frontotemporal dementia are still treated as if they were typical Alzheimer’s disease, which can add harm.",
  "summary": "Antipsychotics in Lewy body disease, antibodies without amyloid confirmation, and cholinesterase inhibitors pushed in FTD are examples of map failures, not j...",
  "asOf": "2026-09-21",
  "links": [
    {
      "label": "NICE NG97",
      "url": "https://www.nice.org.uk/guidance/ng97"
    }
  ],
  "tags": [
    "dementia-atlas"
  ],
  "cancers": [
    "dementia-with-lewy-bodies",
    "frontotemporal-dementia",
    "vascular-dementia"
  ],
  "stage": "biology",
  "severity": "major",
  "causes": [
    "Overlapping symptoms.",
    "Limited access to specialist assessment.",
    "Trial evidence dominated by Alzheimer’s disease."
  ],
  "currentEfforts": [
    "NICE NG97 pathway language.",
    "Lewy body awareness materials."
  ],
  "successLooksLike": "A medicine is offered only when the diagnosis and the label match, or the off-label use is explicit."
}

Key paper

Key paper1 key papers

The papers that changed practice or thinking, each explained: what it found, what it means, and what to be careful about.

FieldTypeRequiredDescription
kind"paper"yesDiscriminator; fixed per kind.
journalstringyesJournal name as printed.
yearintegeryesPublication year.
doistringnoDigital object identifier.
pmidstringnoPubMed id.
authorsstringyes"Surname A, Surname B, et al."
paperType"rct" | "meta-analysis" | "observational" | "real-world" | "basic" | "translational" | "review" | "guideline" | "methods"yesStudy design.
findingsstring[]
default []
noThe key results, one per item, with the numbers.
whatItMeansstringyesPlain English: what this changes for patients, clinicians or the field.
caveatsstring[]
default []
noLimitations, open questions, disputes.
changedPracticebooleannoDid it change guidelines or approvals?
participantsintegernoRoughly how many participants, for trials and cohorts.
Example: da-paper-clarity-ad (arrays trimmed to three items, long strings shortened)
{
  "id": "da-paper-clarity-ad",
  "kind": "paper",
  "name": "Lecanemab in Early Alzheimer’s Disease (CLARITY AD)",
  "tldr": "The published CLARITY AD paper reports the phase 3 lecanemab trial in early Alzheimer’s disease. Read the paper and the FDA communication together.",
  "summary": "Randomised, placebo-controlled trial in early symptomatic Alzheimer’s disease with confirmed amyloid. ARIA was the principal safety theme. This atlas records...",
  "asOf": "2026-09-21",
  "links": [
    {
      "label": "DOI",
      "url": "https://doi.org/10.1056/NEJMoa2212948"
    },
    {
      "label": "ClinicalTrials.gov NCT03887455",
      "url": "https://clinicaltrials.gov/study/NCT03887455"
    }
  ],
  "tags": [
    "dementia-atlas"
  ],
  "cancers": [
    "alzheimers-disease"
  ],
  "drugs": [
    "da-lecanemab"
  ],
  "trials": [
    "da-trial-clarity-ad"
  ],
  "journals": [
    "da-journal-nejm"
  ],
  "journal": "New England Journal of Medicine",
  "year": 2023,
  "doi": "10.1056/NEJMoa2212948",
  "authors": "van Dyck CH, et al.",
  "paperType": "rct",
  "findings": [
    "Primary clinical endpoint favoured lecanemab versus placebo in the enrolled population.",
    "ARIA occurred, more often in ApoE ε4 carriers."
  ],
  "whatItMeans": "This is the main published efficacy trial behind the US traditional approval of lecanemab. A mean scale change is not a promise for one person.",
  "caveats": [
    "Trial cohorts are not the same as mixed-pathology clinic populations.",
    "Individual benefit cannot be read from a group mean."
  ]
}

Journal

Journal1 journals

Where the evidence is published: the journals, their scope and access model, and the key papers each one carried.

FieldTypeRequiredDescription
kind"journal"yesDiscriminator; fixed per kind.
publisherstringyesPublisher.
urlURLyesOfficial URL.
issnstringnoISSN.
scopestringyesBroad scope: general medicine, oncology, haematology, basic science, radiology, etc.
access"subscription" | "hybrid" | "open-access" | "diamond"noOpen access model: subscription, hybrid, open access, diamond.
foundedintegernoYear founded.
impactFactor{ value, year, source }noLatest impact factor or citation metric, with the year it refers to. Optional and clearly labelled.
  • value: number
  • year: integer
  • source: string (optional)
matchNamesstring[]
default []
noNames as they appear in paper records, so papers can be matched to this journal.
societystringnoOwning or affiliated society.
Example: da-journal-nejm (arrays trimmed to three items, long strings shortened)
{
  "id": "da-journal-nejm",
  "kind": "journal",
  "name": "New England Journal of Medicine",
  "tldr": "A general medical journal. CLARITY AD was published here. Listing the journal is not an endorsement of every paper it carries.",
  "summary": "Used so the graph can point from the CLARITY AD paper to its venue.",
  "asOf": "2026-09-21",
  "links": [
    {
      "label": "NEJM",
      "url": "https://www.nejm.org/"
    }
  ],
  "tags": [
    "dementia-atlas"
  ],
  "publisher": "Massachusetts Medical Society",
  "url": "https://www.nejm.org/",
  "scope": "General medicine, including pivotal clinical trials",
  "access": "hybrid",
  "matchNames": [
    "N Engl J Med",
    "New England Journal of Medicine",
    "NEJM"
  ]
}

Statuses

One shared list across kinds. Products use the approval and phase values; trials use recruiting, active, completed and the outcome values; technologies use established, emerging and historic.

ValueShown as
approvedApproved
phase-3Phase 3
phase-2Phase 2
phase-1Phase 1
preclinicalPreclinical
conceptConcept
standard-of-careStandard of care
establishedEstablished
emergingEmerging
historicHistoric
withdrawnWithdrawn
activeActive
completedCompleted
recruitingRecruiting
positivePositive
negativeNegative
mixedMixed
plannedPlanned