Imaging-based target expansion criteria (target expansion score ≥2) identified late-presenting ICH patients at high risk of hematoma expansion, which 'might inform future trials targeting HE and increase the pool of eligible patients.'
Key Findings
Results
A 4-point target expansion score based on three imaging predictors showed good discrimination for hematoma expansion in late-presenting ICH patients in both development and validation cohorts.
The score was derived from a multicenter, international, retrospective cohort of 492 patients presenting 2-12 hours from onset/last seen well
345 patients (70.1%) were in the development cohort and 147 (29.9%) in the validation cohort
AUC was 0.77 (95% CI 0.71-0.83) in the development cohort and 0.79 (95% CI 0.71-0.87) in the validation cohort
Patients with brainstem ICH, anticoagulation, and baseline ICH volume >100 mL were excluded to mirror a clinical trial population
Methods
The target expansion score was composed of three independent imaging predictors: high ultra-early hematoma growth (1 point), NCCT hypodensities (1 point), and CTA spot sign (2 points).
Ultra-early hematoma growth (uHG) was defined as baseline volume divided by onset-to-imaging time >10 mL/h
Hypodensities were assessed on noncontrast CT (NCCT)
Spot sign was assessed on CT angiography (CTA)
Point weights were derived from multivariable logistic regression estimates
The primary outcome was hematoma expansion defined as >33% or >6 mL growth
Results
There was a nonlinear increase in hematoma expansion frequency and mean absolute hematoma growth with higher target expansion scores.
Score 0: HE frequency 9.8%, mean absolute hematoma growth 0.9 mL
Score 1: HE frequency 17.1%, mean absolute hematoma growth 3.9 mL
Score 2: HE frequency 48.3%, mean absolute hematoma growth 11.5 mL
Score 3: HE frequency 70.8%, mean absolute hematoma growth 13.0 mL
Score 4: HE frequency 87.5%, mean absolute hematoma growth 26.9 mL
Results
A dichotomized target expansion score ≥2 predicted hematoma expansion with consistent sensitivity and specificity across development and validation cohorts.
In the development cohort: sensitivity 0.61, specificity 0.85
In the validation cohort: sensitivity 0.64, specificity 0.84
These findings were described as confirmed in the validation cohort
Background
The benefit of therapies targeting hematoma expansion in acute ICH is highly time dependent and greatest within the first 2 hours after onset, motivating the need for criteria applicable to late presenters.
The study focused on patients presenting 2-12 hours from onset or last seen well
Existing HE-targeting therapies have shown the greatest benefit within the first 2 hours
The authors aimed to identify a subset of late presenters who remain at sufficiently high HE risk to benefit from such therapies
The authors suggest these findings 'might inform future trials targeting HE and increase the pool of eligible patients'
What This Means
When someone has a brain bleed (intracerebral hemorrhage or ICH), the blood clot can continue to grow larger in the hours after it starts — a process called hematoma expansion. Treatments that try to stop this expansion work best when given very early, ideally within the first two hours. But many patients arrive at the hospital later than that, and it has been unclear which of these 'late presenters' are still at high enough risk of expansion to benefit from these treatments. This research suggests that three specific findings on brain imaging scans — how fast the bleed was growing before the patient arrived (ultra-early hematoma growth), certain low-density areas visible on a standard CT scan (hypodensities), and a marker seen on CT angiography called the 'spot sign' — can be combined into a simple 4-point score to identify late-arriving patients who remain at high risk.
The study found that patients with a score of 2 or higher had dramatically higher rates of hematoma expansion: nearly half of patients scoring 2 experienced expansion, rising to almost 9 in 10 patients with the maximum score of 4. By contrast, fewer than 1 in 10 patients scoring 0 experienced expansion. These results held up consistently in both the group used to develop the score and a separate group used to test it, across multiple hospitals and countries.
This research suggests that using these imaging-based 'target expansion criteria' could help doctors and clinical trial researchers identify which late-presenting ICH patients are most likely to benefit from treatments aimed at stopping the bleed from growing. This could potentially expand the window of time during which patients are eligible for such treatments, allowing more people to be included in future clinical trials and potentially receive benefit.
Morotti A, Pensato U, Nawabi J, Dell'Orco A, Mazzacane F, Li Q, et al.. (2026). Intracerebral Hemorrhage Expansion in Late Presenters: Proposed Target Expansion Criteria.. Neurology. https://doi.org/10.1212/WNL.0000000000218675