Diagnosis & Grading of Intracerebral Hemorrhage

The diagnostic workup of intracerebral hemorrhage (ICH) answers three sequential questions: Is it hemorrhage? (non-contrast CT), Why did it bleed? (CTA and MRI screening for a macrovascular or secondary cause), and What is the prognosis and who is at risk of deterioration? (grading and expansion-prediction scores). This topic walks through imaging — NCCT, CTA, and MRI — and then the validated scores used for prognostication, secondary-cause screening, and hematoma-expansion risk.

1. Clinical Picture

ICH classically presents with a sudden focal neurological deficit that progresses over minutes to hours, frequently accompanied by headache, vomiting, depressed level of consciousness, and markedly elevated blood pressure — features that, together, favor hemorrhage over ischemia but are not specific enough to defer imaging. The deficit reflects hematoma location:

• Putaminal / basal ganglia (most common, hypertensive): contralateral hemiparesis, gaze deviation toward the lesion.
• Thalamic: contralateral sensorimotor loss, downward/medial gaze, miosis; IVH and hydrocephalus common.
• Lobar: deficits by lobe (aphasia, neglect, hemianopia, seizures); raises suspicion for CAA or a secondary cause.
• Cerebellar: ataxia, vomiting, headache — a neurosurgical emergency given risk of brainstem compression and hydrocephalus.
• Pontine: coma, pinpoint pupils, quadriparesis — typically devastating.

2. Non-Contrast CT (NCCT)

NCCT is the universal first-line study: it is fast, available, and highly sensitive for acute blood, which appears hyperdense (≈ 50–70 Hounsfield units) against lower-density brain. Beyond simple detection, the initial scan defines the parameters that drive grading and management:

• Volume by the bedside ABC/2 method (A = greatest diameter, B = diameter perpendicular to A, C = number of slices with hemorrhage × slice thickness, divided by 2) — the volume term used in nearly every ICH score.
• Intraventricular extension (IVH) and hydrocephalus — independent predictors of poor outcome.
• Mass effect, midline shift, and herniation.
• Early NCCT markers of expansion (blend, swirl, black-hole, island, satellite signs) — detailed in Section 8.

3. CT Angiography (CTA)

CTA is increasingly obtained acutely because it answers two distinct questions at once — will it grow? and why did it bleed?

The spot sign — predicting expansion

The spot sign is one or more foci of contrast enhancement within the hematoma on CTA, representing active contrast extravasation from ongoing bleeding. It is the strongest imaging predictor of hematoma expansion and of in-hospital mortality, and it underlies CTA-based expansion scores. (Despite its predictive power, spot-sign–targeted hemostatic therapy has not improved outcomes in trials such as SPOTLIGHT/STOP-IT.)

Screening for a secondary cause

CTA also screens for a macrovascular or structural cause, which changes management entirely:

• Vascular malformations — arteriovenous malformation (AVM), saccular aneurysm, dural arteriovenous fistula, developmental venous anomaly, or cavernoma.
• Enhancing tumor — a hemorrhagic neoplasm (primary or metastatic), suggested by disproportionate edema, an enhancing nodule, or atypical location.
• Cerebral venous thrombosis — particularly with lobar/parasagittal hemorrhage and a non-arterial pattern.

Suspicion for a secondary cause rises with younger age, lobar or intraventricular location, absence of hypertension or small-vessel disease, and atypical imaging features. The DIAGRAM and MACRO scores (Sections 6–7) formalize this judgment.

4. MRI: Staging Blood by Age

MRI signal of hemorrhage evolves predictably as hemoglobin degrades from oxyhemoglobin → deoxyhemoglobin → intracellular methemoglobin → extracellular methemoglobin → hemosiderin. Recognizing the stage helps date the bleed and is central to detecting old or multiple hemorrhages.

Stage	Approx. age	Hemoglobin form	T1	T2
Hyperacute	< 24 h	Oxyhemoglobin (intracellular)	Iso / dark	Bright
Acute	1–3 days	Deoxyhemoglobin (intracellular)	Iso / dark	Dark
Early subacute	3–7 days	Methemoglobin (intracellular)	Bright	Dark
Late subacute	1 week – 1 month	Methemoglobin (extracellular)	Bright	Bright
Chronic	> 1 month	Hemosiderin / ferritin	Dark	Dark

Memory aid (T1, T2 signal by stage): IB — ID — BD — BB — DD  (“Iso-Bright, Iso-Dark, Bright-Dark, Bright-Bright, Dark-Dark“), where I = isointense, B = bright, D = dark. A useful clinical anchor: the late subacute stage is bright on both T1 and T2, and the chronic stage is dark on both.

5. Blood-Sensitive MRI Sequences

Beyond conventional T1/T2, susceptibility-based sequences exploit the paramagnetic effect of blood-breakdown products and are essential for detecting acute blood, chronic microbleeds, and cortical superficial siderosis.

Sequence	What it detects	Notes
GRE (gradient echo)	“Blooming” of deoxyhemoglobin, methemoglobin, and hemosiderin	Detects acute hemorrhage and chronic microbleeds; widely available
T2*	Same susceptibility effect as GRE	Older susceptibility technique; superseded by SWI where available
SWI (susceptibility-weighted)	Most sensitive to blood products, microbleeds, and superficial siderosis	More sensitive than GRE/T2*; can overestimate hematoma size due to blooming
DWI	Co-existing acute ischemia; variable signal within clot	Useful when an underlying infarct or tumor is suspected
FLAIR / T2	Perihematomal edema, chronic cavity, cortical superficial siderosis	Supports CAA diagnosis and dating

Microbleed distribution is an etiologic clue: a strictly lobar / cortical pattern (± cortical superficial siderosis) supports cerebral amyloid angiopathy, whereas a deep / basal-ganglia / brainstem pattern supports hypertensive arteriopathy. This same small-vessel-disease burden is what the MACRO score uses to estimate how likely a macrovascular cause is.

6. Prognostic Grading Scores

ICH Score (Hemphill, 2001)

The original and most widely used grading scale, designed for rapid bedside risk stratification and standardized communication. It sums five variables (0–6 points) and correlates strongly with 30-day mortality.

Component	Finding	Points
GCS	3–4	2
	5–12	1
	13–15	0
ICH volume	≥ 30 mL	1
Intraventricular hemorrhage	Present	1
Infratentorial origin	Yes	1
Age	≥ 80 years	1

ICH Score	0	1	2	3	4	5
30-day mortality	0%	13%	26%	72%	97%	100%

(No patients scored 6 in the original cohort; a score of 6 is presumed near-uniformly fatal.) Subsequent validations reproduce this near-linear mortality gradient.
Important caveat: these mortality figures derive from eras with frequent early withdrawal of life-sustaining treatment, creating a self-fulfilling prophecy. Current AHA/ASA guidance recommends against using early prognostic scores to justify limiting care in the first 24–48 hours; default to full aggressive treatment.

FUNC Score (Rost, 2008)

Where the ICH Score predicts death, the FUNC score predicts functional independence at 90 days (0–11 points; higher = better outcome).

Component	Finding	Points
ICH volume	< 30 mL	4
	30–60 mL	2
	> 60 mL	0
Age	< 70	2
	70–79	1
	≥ 80	0
Location	Lobar	2
	Deep	1
	Infratentorial	0
GCS	≥ 9	2
Pre-ICH cognitive impairment	Absent	1

Functional independence (GOS ≥ 4) at 90 days rises steeply with score: roughly 0% at ≤ 4 points up to ~80% at 11 points, supporting goals-of-care discussions once the acute phase has stabilized.

7. Screening for a Macrovascular / Secondary Cause

A macrovascular cause is found in roughly 7–15% of spontaneous ICH and is far more common in younger, normotensive patients with lobar or posterior-fossa bleeds. Missing one (e.g., an AVM or aneurysm) risks catastrophic rebleeding, so structured scores help decide who needs vascular imaging up to digital subtraction angiography (DSA).

DIAGRAM score (Hilkens, 2018) — CT/CTA-based

Combines age (18–50, 51–70, >70), ICH location (lobar or posterior-fossa vs. deep), and small-vessel-disease markers on CT, with an optional CTA term. Predicted probability of a macrovascular cause ranges from ~1% (older patient, deep ICH, with SVD) to > 50% (young patient, lobar/posterior-fossa ICH, no SVD). Discrimination: c-statistic 0.83 with clinical + CT variables, rising to 0.91 when CTA is added (DIAGRAM-plus). It flags patients in whom a negative CTA should still prompt MRI/MRA or DSA.

MACRO score (Fandler-Höfler, 2024) — MRI-based

A newer, MRI-based complement (MRI Assessment of the Causes of intRacerebral haemOrrhage) that leverages small-vessel-disease burden — the opposite of macrovascular disease. It combines age (0–39 / 40–69 / ≥70), ICH location (lobar / deep / infratentorial), and MRI SVD markers (cerebral microbleeds, lacunes, cortical superficial siderosis, white-matter hyperintensities by Fazekas).

Total score (0–7)	Estimated macrovascular probability	Implication
≥ 6 (high SVD burden)	≈ 0.2%	Very low — extensive angiographic workup rarely needed
≤ 2 (little/no SVD)	≈ 48.9%	High — pursue DSA / dedicated vascular imaging

Note the direction: higher MACRO score = lower macrovascular risk (more small-vessel disease makes a “primary” cause likelier). In derivation (1,043 patients, 7.5% with a macrovascular cause) the c-statistic was 0.90 (95% CI 0.88–0.93) and 0.87 on external validation — outperforming CT-based scores. MACRO and DIAGRAM are complementary: DIAGRAM triages at the CT/CTA stage, MACRO refines the decision once MRI is available.

8. Hematoma-Expansion Prediction

Hematoma expansion — growth of ≥ 6 mL or ≥ 33% — occurs in roughly a third of patients, mostly within the first few hours, and is the strongest modifiable predictor of poor outcome. Identifying high-risk patients early targets monitoring and aggressive blood-pressure and reversal strategies (see INTERACT2 and the Acute Management → Blood Pressure topic).

Non-contrast CT markers of expansion

Marker	Description
Blend sign	Hypoattenuating region blended with adjacent hyperattenuating hematoma, with a clear demarcation
Black-hole sign	Encapsulated hypoattenuating area within the hyperdense hematoma, distinct densities
Swirl sign	Region of hypo- or isodensity within a hyperdense hematoma (active bleeding)
Island sign	≥ 3 scattered small hematomas separate from the main bleed, or ≥ 4 bubble-/budding-shaped foci
Satellite sign	Small hemorrhage separate from but within 20 mm of the main hematoma

BAT score (Morotti, 2019)

A simple 0–5 non-contrast score combining one imaging sign, one density feature, and timing:

Component (BAT)	Finding	Points
Blend sign	Present	1
Any hypodensity	Present	2
Timing (onset → NCCT)	< 2.5 hours	2

Higher scores predict greater expansion risk; a score ≥ 3 identifies a high-risk group. The BAT score requires only the initial non-contrast scan and the time of onset.

HEP score (Hematoma Expansion Prediction)

A more granular 0–18 nomogram-derived scale that integrates several NCCT features (hypodensities, blend sign, hematoma shape and density, fluid level) with clinical variables (including a history of dementia and smoking). It predicts substantial expansion (≥ 6 mL or ≥ 33%) with accuracy comparable to the CTA spot sign — a useful advantage where CTA is unavailable or contraindicated.

9. Putting It Together

• Everyone: NCCT → confirm, localize, volume (ABC/2), IVH, hydrocephalus; compute the ICH Score.
• Expansion risk: assess NCCT markers + BAT/HEP; add CTA for the spot sign when early intervention is being considered.
• Secondary cause: in younger, lobar, or normotensive patients, use DIAGRAM (CT/CTA) and, once MRI is done, MACRO to decide on DSA.
• Prognosis: ICH Score (mortality) and FUNC (function) inform — but do not dictate — goals of care; avoid early nihilism.

References

1. Hemphill JC 3rd, et al. The ICH Score: a simple, reliable grading scale for intracerebral hemorrhage. Stroke. 2001;32(4):891–897.
2. Rost NS, et al. Prediction of functional outcome in patients with primary intracerebral hemorrhage: the FUNC score. Stroke. 2008;39(8):2304–2309.
3. Hilkens NA, et al. Predicting the presence of macrovascular causes in non-traumatic intracerebral haemorrhage: the DIAGRAM prediction score. J Neurol Neurosurg Psychiatry. 2018;89(7):674–679.
4. Fandler-Höfler S, et al. MRI-based prediction of macrovascular causes of intracerebral hemorrhage: the MACRO score. Neurology. 2024;103(10):e209950.
5. Morotti A, et al. Predicting intracerebral hemorrhage expansion with noncontrast computed tomography: the BAT score. Stroke. 2018;49(5):1163–1169.
6. Yao X, et al. The HEP score: a nomogram-derived hematoma expansion prediction scale. Neurocrit Care. 2015;23(2):179–187.
7. Greenberg SM, et al. 2022 Guideline for the Management of Patients With Spontaneous Intracerebral Hemorrhage. Stroke. 2022;53(7):e282–e361.