ABC/2 Haematoma Volume Calculator
ABC/2 Haematoma Volume Calculator
Estimate intracerebral haemorrhage volume as A × B × C ÷ 2. The method approximates the clot as an ellipsoid: it runs 4.5% below an exact ellipsoid and about 20% above CT planimetry on real haematomas, and the error grows with irregular and lobar clots.
ABC/2 haematoma volume
Three diameters to mLGreatest axial diameter 5.2 cm, perpendicular diameter on the same slice 3.4 cm, craniocaudal extent 3.0 cm
Formula
which approximates the ellipsoid π/6 × A × B × C, since π/6 = 0.5236
- A and B
- A is the greatest diameter of the haemorrhage in the axial plane, on the slice where it is largest — a diagonal in an obliquely oriented clot. B is the diameter at 90 degrees to A on the same slice. Measuring B on a different slice, or as a conventional anteroposterior diameter regardless of where A ran, is the commonest technical error and it inflates the volume
- C
- the craniocaudal extent: classically the number of slices containing haemorrhage multiplied by the slice thickness, with partial slices weighted 1 at or above 75 per cent of the largest slice’s haemorrhage area, 0.5 between 25 and 75 per cent, and 0 below. Entering a slice count as centimetres is the error that produces a volume several times too small
- the divisor 2
- a deliberate rounding of π/6, which is 0.5236, so ABC/2 sits 4.51 per cent below the ellipsoid it is modelling. The rounding was accepted because the error from treating a real haematoma as an ellipsoid at all is an order of magnitude larger
- the error against planimetry, and its direction
- Webb and colleagues compared ABC/2 with CT planimetry across 4,369 scans from 507 patients in three trials. ABC/2 overestimated: 15.2 cm³ on average against 12.7 cm³, roughly 20 per cent high. Only 84 per cent of scans fell within 5 mL of the reference and only 48 per cent within 20 per cent of it. Agreement was better for the mild/moderate/severe categorisation (kappa 0.75) than for the number, and a 5 mL change between consecutive scans was detected with sensitivity 0.76 and specificity 0.86
Worked example
Greatest axial diameter 5.2 cm, perpendicular diameter on the same slice 3.4 cm, craniocaudal extent 3.0 cm
5.2 × 3.4 = 17.68, × 3.0 = 53.04, ÷ 2 = 26.5 mL
The exact ellipsoid of the same three axes is π/6 × 53.04 = 27.8 mL, so the rounded divisor costs 4.51 per cent — at every size, because it is a constant ratio
Scaled by Webb’s mean ratio of planimetry to ABC/2 (12.7 to 15.2), the same clot would measure about 22.2 mL planimetrically. That is the error that matters: the 30 cm³ ICH score threshold sits between the two figures
Check the arithmetic is a plain product by swapping the inputs: 3.4 × 5.2 × 3.0 ÷ 2 and 3.0 × 3.4 × 5.2 ÷ 2 both give 26.5. Doubling any single diameter doubles the volume; doubling all three gives 212.2 mL, eight times the original
A 6 × 5 × 4 cm haematoma gives exactly 60.0 mL and a 4 × 3 × 2 cm one 12.0 mL — useful anchors, because a two-centimetre error in any one diameter moves the answer by tens of millilitres
What each letter is measured on
| Letter | Measurement | Where it goes wrong |
|---|---|---|
| A | Greatest diameter in the axial plane, on the slice where the haemorrhage is largest | Taking the left-to-right or anteroposterior width instead of the true greatest diameter, which in an oblique clot is a diagonal |
| B | Diameter at 90° to A, on the same slice as A | Measuring on a different slice, or taking a conventional anteroposterior diameter regardless of A’s direction |
| C | Slices containing haemorrhage × slice thickness, partial slices weighted 1 (≥75%), 0.5 (25–75%) or 0 (<25%) | Entering the slice COUNT rather than count × thickness, which divides the volume by the slice thickness in cm |
The two errors, in opposite directions
| Comparison | Direction | Magnitude | Source |
|---|---|---|---|
| ABC/2 against the exact ellipsoid of the same three axes | ABC/2 reads LOW | 4.51 per cent, at every size | Arithmetic: π/6 = 0.5236 and the method uses 0.5 |
| ABC/2 against CT planimetry on real haematomas | ABC/2 reads HIGH | 15.2 cm³ against 12.7 cm³, about 20 per cent. 84% of scans within 5 mL; only 48% within 20 per cent | Webb 2015, 4,369 scans from 507 patients |
An ellipsoid stood in for a clot, and it reads high
ABC/2 estimates the volume of an intracerebral haemorrhage by treating it as an ellipsoid and measuring three orthogonal axes off the CT. A is the greatest diameter in the axial plane, taken on the slice where the haemorrhage is largest; B is the diameter at 90 degrees to A on that same slice; C is the craniocaudal extent, classically the number of slices containing haemorrhage multiplied by the slice thickness, with partial slices weighted as a whole, a half or nothing according to how much of the largest slice’s haemorrhage area they contain. Three measurements and one division, which is why it has survived since 1996 as the volume that gets quoted down a telephone.
The divisor is where the geometry hides. An ellipsoid with axes A, B and C has a volume of π/6 × ABC, and π/6 is 0.5236, so dividing by 2 instead gives a figure 4.51 per cent below the ellipsoid being modelled — a constant ratio at every size. That rounding was accepted deliberately, because it is dwarfed by the much larger error from assuming a real haematoma is an ellipsoid at all. Webb and colleagues measured that larger error across 4,369 scans from 507 patients in three trials against CT planimetry: ABC/2 averaged 15.2 cm³ where planimetry gave 12.7, an overestimate of roughly 20 per cent, with only 84 per cent of scans within 5 mL of the reference and only 48 per cent within 20 per cent of it. The two errors run in opposite directions, they do not cancel, and the real-world one is four times the arithmetic one.
The overestimation is not uniform. Webb’s analysis found the method performed better with smaller, thalamic, homogeneous clots and that accuracy “decreases with large, irregular, or lobar clots” — and ABC/2 is known to overestimate anticoagulant-related haemorrhage, which is characteristically irregular. Categorical agreement held up better than the number itself (kappa 0.75 for a mild/moderate/severe split), so the method is serviceable for ranking and weaker as a measurement. Which brings the error up against a threshold: the ICH score awards a point for 30 cm³ or more — one point out of six, and in its derivation cohort the difference between the 26 per cent and 72 per cent mortality strata. That derivation measured volume by ABC/2, so the threshold is at least calibrated against the same method; but a haematoma near 30 cm³ can fall either side of the line depending on who measured it. Know which way the method errs and by how much, and say which method a volume came from. A grade is not a diagnosis and a cohort risk is not this patient’s probability: a stratum in which 72 per cent died tells you about that cohort, not which 72 per cent. Every threshold here comes from a named cohort, and cohorts differ in case mix, era and treatment; where your unit’s protocol differs, it takes precedence.
Frequently asked questions
What is the ABC/2 formula for haematoma volume?
Volume in millilitres equals A × B × C divided by 2, with all three diameters in centimetres. A is the greatest axial diameter on the slice where the haemorrhage is largest, B is the diameter at 90 degrees to A on that same slice, and C is the craniocaudal extent.
How is C measured for ABC/2?
Classically as the number of CT slices containing haemorrhage multiplied by the slice thickness, with partial slices weighted: at least 75 per cent of the largest slice’s haemorrhage area counts as a whole slice, 25 to 75 per cent as a half, under 25 per cent as none. With multiplanar reformats the craniocaudal diameter can be measured directly.
Does ABC/2 overestimate or underestimate haematoma volume?
Both, against different references. Against the exact ellipsoid it models it reads 4.51 per cent low, because π/6 is 0.5236 and the method rounds it to 0.5. Against CT planimetry on real haematomas it reads about 20 per cent high — 15.2 cm³ against 12.7 cm³ across 4,369 scans. The second effect is the larger and the clinically relevant one.
When is ABC/2 least reliable?
In large, irregular and lobar haematomas, where the ellipsoid assumption breaks down and the overestimation grows, and in anticoagulant-related haemorrhage, which is characteristically irregular. Only 48 per cent of scans in the largest validation fell within 20 per cent of the planimetric volume.
Is 30 mL a meaningful threshold given the error?
It is the threshold that scores a point on the ICH score, and that score’s derivation cohort measured volume by ABC/2, so the two are at least calibrated together. But a haematoma near 30 cm³ sits well inside the method’s measurement error, and that single point is a sixth of the ICH score’s range.
Related calculators
References
- Webb AJS, Ullman NL, Morgan TC, et al. Accuracy of the ABC/2 score for intracerebral hemorrhage: systematic review and analysis of MISTIE, CLEAR-IVH, and CLEAR III. Stroke. 2015;46(9):2470–2476.
- ABC/2. Radiopaedia.org (accessed 7 October 2026), citing Kothari RU, Brott T, Broderick JP, et al., Stroke 1996;27(8):1304–1305.
- ICH volume calculator. Omni Calculator (accessed 7 October 2026).
- Hemphill JC 3rd, Bonovich DC, Besmertis L, Manley GT, Johnston SC. The ICH score: a simple, reliable grading scale for intracerebral hemorrhage. Stroke. 2001;32(4):891–897.
Not medical advice. For healthcare professionals and education. Reference intervals vary by laboratory and assay — always use your own laboratory's. Never base a dose or a treatment decision on this page alone. Full disclaimer at calcengines.com/disclaimer/