Fazekas visual rating of periventricular and deep white matter hyperintensities

{
  "categories": [
    {
      "outcome_code": "0",
      "periventricular_WMH": "absent",
      "deep_WMH": "absent"
    },
    {
      "outcome_code": "1",
      "periventricular_WMH": "caps_or_pencil_thin_lining",
      "deep_WMH": "punctate_foci"
    },
    {
      "outcome_code": "2",
      "periventricular_WMH": "smooth_halo",
      "deep_WMH": "beginning_confluence"
    },
    {
      "outcome_code": "3",
      "periventricular_WMH": "irregular_signal_extending_into_deep_white_matter",
      "deep_WMH": "large_confluent_areas"
    }
  ],
  "applicability": {
    "use_for": "Visual MRI description of white matter hyperintensity burden after confirming that the observed lesions are appropriately characterized as WMH.",
    "original_population": "The 1987 derivation compared 12 patients with Alzheimer dementia, 4 with multi-infarct dementia and 9 controls on 1.5-T MRI; it is not a modern diagnostic or prognostic calibration cohort.",
    "required_context": [
      "patient_age",
      "clinical_indication",
      "MRI_sequence_and_quality",
      "lesion_distribution",
      "neurologic_history",
      "vascular_risk_factors"
    ],
    "do_not_apply_as": [
      "single_global_etiology_label",
      "dementia_diagnosis",
      "multiple_sclerosis_exclusion",
      "individual_event_probability",
      "treatment_score"
    ]
  },
  "rating_architecture": {
    "independent_axes": [
      {
        "id": "periventricular_WMH",
        "codes": [
          "0",
          "1",
          "2",
          "3"
        ],
        "report_separately": true
      },
      {
        "id": "deep_WMH",
        "codes": [
          "0",
          "1",
          "2",
          "3"
        ],
        "report_separately": true
      }
    ],
    "original_output": "Two raw scores, one periventricular and one deep; there is no single original total or mandated combined score.",
    "aggregation_guard": "If a downstream protocol requires one summary value, name the protocol and aggregation rule explicitly, commonly the maximum of the two axes, and retain both raw component scores. Never silently collapse them."
  },
  "decisive_boundaries": {
    "periventricular_0_1": "No periventricular WMH is 0; caps or a pencil-thin ventricular lining are 1.",
    "periventricular_1_2": "A smooth periventricular halo is 2 rather than the thin lining or caps of 1.",
    "periventricular_2_3": "Irregular periventricular signal extending into adjacent deep white matter is 3 rather than a smooth halo.",
    "deep_0_1": "No deep WMH is 0; punctate deep foci are 1.",
    "deep_1_2": "Beginning or early confluence moves deep WMH from punctate grade 1 to grade 2.",
    "deep_2_3": "Large confluent deep-white-matter areas are grade 3."
  },
  "imaging_and_terminology": {
    "sequence_rule": "Rate on an MRI sequence that demonstrates WMH reliably, typically T2-weighted and FLAIR, and document material motion, incomplete coverage or other quality limitations.",
    "STRIVE_term": "Use white matter hyperintensity of presumed vascular origin only when the imaging pattern and context support that terminology; STRIVE terminology does not prove vascular etiology in an individual.",
    "separate_findings": [
      "lacunes",
      "recent_small_subcortical_infarcts",
      "enlarged_perivascular_spaces",
      "microbleeds",
      "superficial_siderosis",
      "cortical_infarcts",
      "atrophy"
    ]
  },
  "etiology_and_differential_boundary": {
    "descriptive_not_etiologic": "Fazekas grades lesion burden and cannot by itself distinguish chronic small-vessel ischemic change from demyelinating, inflammatory, infectious, toxic-metabolic, traumatic, migraine-associated, neoplastic or other white-matter processes.",
    "atypical_pattern_rule": "When age, morphology, location, enhancement, diffusion, mass effect or clinical course is atypical for presumed vascular WMH, state the differential and withhold a vascular Fazekas interpretation if applicability is not established."
  },
  "current_incidental_WMH_care_context": {
    "initial_assessment": "For incidentally identified WMH, review history and examination for prior neurologic events and evaluate cardiovascular risk, especially blood pressure, diabetes, lipids and smoking, in the appropriate clinical setting.",
    "risk_factor_management": "Treat identified cardiovascular risks according to established primary-care or specialty guidance; a Fazekas code alone does not create a drug, dose or blood-pressure prescription.",
    "antithrombotic_guard": "Do not start aspirin, another antiplatelet agent or anticoagulation solely because incidental WMH or a Fazekas grade is present; a separate clinical indication is required.",
    "statin_guard": "Do not start a statin solely because incidental WMH or a Fazekas grade is present; use the person's usual lipid and cardiovascular indication.",
    "referral_context": "Neurology assessment is appropriate when symptoms are rapidly progressive or unexplained, or when the imaging pattern is atypical; routine specialist referral is not encoded by grade alone."
  },
  "risk_and_interpretation": {
    "population_context": "Greater WMH burden is associated at population level with stroke, cognitive decline, gait impairment, falls, depression and mortality, but the scale does not yield an individual probability for any of these outcomes.",
    "original_study_guard": "The original small derivation reported punctate and early confluent deep WMH in both Alzheimer and control participants; do not convert the 1987 sample into current per-grade risk estimates or etiologic certainty.",
    "age_and_pattern": "A small burden may be common with increasing age, whereas confluent burden deserves clinical context; age, distribution, symptoms, longitudinal change and coexisting imaging findings remain necessary for interpretation."
  },
  "agent_output_contract": [
    "periventricular_WMH_raw_score_0_to_3_and_descriptor",
    "deep_WMH_raw_score_0_to_3_and_descriptor",
    "any_named_summary_protocol_and_aggregation_rule",
    "MRI_sequence_quality_and_comparison_date",
    "distribution_and_atypical_features",
    "separate_STRIVE_small_vessel_disease_markers",
    "clinical_and_vascular_context_if_available",
    "uncertainty_and_missing_inputs",
    "burden_not_etiology_diagnosis_or_treatment_warning"
  ],
  "missing_input_behavior": [
    "If only one axis can be rated, return that raw score and mark the other axis unassessable rather than copying the observed score to both.",
    "If a report supplies one undifferentiated Fazekas number, preserve it as a reported external value and state that the periventricular and deep components cannot be reconstructed.",
    "If lesion pattern is atypical or the MRI sequence is inadequate, do not force a grade; request the relevant sequence, distribution or clinical context."
  ],
  "evidence_map": [
    {
      "role": "original_scale",
      "citation": "Fazekas et al., AJR 1987",
      "doi": "10.2214/ajr.149.2.351"
    },
    {
      "role": "current_terminology",
      "citation": "Wardlaw et al., STRIVE 2013 and Duering et al., STRIVE-2 2023",
      "doi": "10.1016/S1474-4422(23)00131-X"
    },
    {
      "role": "incidental_WMH_care_context",
      "citation": "Ottavi et al., Medical Journal of Australia 2023",
      "doi": "10.5694/mja2.52079"
    }
  ],
  "source_locator": "Fazekas et al., AJR 1987;149:351-356, PMID 3496763, abstract and original two-location rating definitions; Wardlaw et al., Lancet Neurol 2013, PMC3714437, STRIVE definitions; Duering et al., Lancet Neurol 2023, DOI 10.1016/S1474-4422(23)00131-X, STRIVE-2; Ottavi et al., Med J Aust 2023;219:278-284, DOI 10.5694/mja2.52079, Box 3 and consensus recommendations."
}