Acute spontaneous intracerebral haemorrhage represents a medical emergency with high early mortality and severe disability. Haematoma expansion occurs rapidly within the first several hours after symptom onset, and it serves as the most potent modifiable determinant of neurological deterioration. Consequently, accurate haematoma expansion prediction remains essential for emergency physicians, neurologists, and neurointensivists. Clinicians must identify high-risk patients quickly to guide blood pressure reduction, hemostatic therapies, and neurosurgical intervention. In addition, a comprehensive multicentre study led by Pensato and colleagues provides vital comparative insights into how advanced imaging modalities refine expansion risk stratification.

Clinical Significance of Haematoma Expansion Prediction

Haematoma enlargement affects approximately one-quarter of individuals presenting with acute spontaneous intracerebral haemorrhage. Most expansion events unfold during the earliest hours after bleeding begins. Consequently, early growth directly accelerates intracranial pressure, worsens cerebral perfusion, and increases mortality. Therefore, clinicians urgently require validated stratification systems to detect patients vulnerable to rapid haemorrhage growth. Reliable stratification tools help physicians deliver targeted interventions, including intensive blood pressure lowering and prompt reversal of anticoagulation.

Moreover, clinical risk scores allow investigators to enrich clinical trials with patients most likely to benefit from experimental hemostatic drugs. In this comparative investigation, researchers evaluated 217 adult acute stroke patients admitted to a tertiary comprehensive stroke centre in Southern Alberta, Canada. The primary outcome assessed significant haematoma growth, defined as volume expansion of at least 6 mL or 33 percent. In addition, the team tracked severe expansion, characterized as volume growth reaching at least 12.5 mL or 66 percent. Overall, 51 patients developed significant enlargement, whereas 35 experienced severe growth. These rigorous outcome measures provide a standard framework to test prognostic scoring systems across different neuroimaging protocols.

Predictive Capacity of Non-Contrast Computed Tomography

Non-contrast computed tomography remains the universal first-line neuroimaging modality for acute stroke in emergency units worldwide. Emergency physicians value non-contrast scans because of rapid acquisition and widespread availability. Researchers evaluated eight distinct non-contrast computed tomography scores designed to forecast haemorrhage enlargement. These algorithms typically incorporate baseline haematoma volume, time from symptom onset, and subtle morphological signs. Specifically, radiologists evaluate imaging markers such as the blend sign, black hole sign, hypodensities, and irregular haematoma margins.

However, non-contrast computed tomography scores demonstrated only modest predictive accuracy in this rigorous comparative analysis. The c-statistic values for predicting haematoma enlargement ranged from 0.516 to 0.674 across the eight non-contrast scoring systems. Furthermore, when researchers examined severe haematoma growth, predictive discrimination remained similarly limited, yielding c-statistics between 0.505 and 0.666. Thus, these findings indicate that relying solely on unenhanced scans provides insufficient diagnostic discrimination for acute risk assessment. While non-contrast markers certainly reflect internal density differences and active clot organization, subjective visual evaluation often introduces interobserver variability. Consequently, acute stroke protocols require more precise, objective vascular imaging tools to stratify patients accurately during emergency triage.

Diagnostic Gains with Single-Phase CT Angiography

To overcome the diagnostic limitations of unenhanced imaging, stroke teams increasingly implement single-phase computed tomography angiography. This technique visualizes active extravasation of iodinated contrast into the haematoma parenchyma, widely recognised as the CTA spot sign. In this study, investigators assessed six established single-phase computed tomography angiography prediction scores. These scoring models integrate contrast extravasation metrics alongside clinical variables and baseline haemorrhage dimensions.

Notably, single-phase vascular imaging delivered a noticeable improvement in predictive capability compared to unenhanced protocols. The c-statistic values for predicting haematoma enlargement rose to a range between 0.627 and 0.725 among the single-phase systems. Similarly, for severe haematoma expansion, single-phase scores generated c-statistics ranging from 0.651 to 0.740. Therefore, identifying contrast extravasation substantially enhances clinical risk assessment over baseline parenchymal markers. Nevertheless, single-phase acquisitions capture contrast dynamics at only a single snapshot in time. As a result, early scans may miss slow or delayed contrast leakage, leading to false-negative evaluations in patients with lower flow dynamics. Indeed, this temporal limitation highlighted the clear necessity for dynamic vascular imaging protocols.

Superior Performance of Multiphase CT Angiography

Multiphase computed tomography angiography addresses the temporal blind spots of conventional single-phase angiography by acquiring three sequential phases across the cerebral circulation. This dynamic protocol tracks contrast entry, peak arterial flow, and delayed venous phases without requiring repeated contrast boluses. Researchers analyzed two dedicated multiphase scoring algorithms within this consecutive patient cohort. These models evaluate the timing of spot sign appearance and follow its morphological evolution across consecutive time frames.

Consequently, multiphase computed tomography angiography yielded outstanding gains in prognostic discrimination. For predicting primary haematoma expansion, multiphase scores achieved c-statistics between 0.800 and 0.814. Furthermore, in forecasting severe haematoma growth, multiphase models generated c-statistic values ranging from 0.813 to 0.828. Statistical evaluations using the DeLong test demonstrated significant diagnostic superiority for multiphase computed tomography angiography over both unenhanced scans and single-phase angiography. By capturing delayed contrast pooling, multiphase imaging effectively unmasks slow, persistent bleeding sources that single-phase scans fail to detect. Thus, time-resolved imaging offers superior biological fidelity when assessing active intracranial extravasation.

Clinical Implications for Emergency Stroke Management

The findings of this benchmark study carry actionable implications for acute neurocritical care pathways and emergency departments. Specifically, incorporating multiphase vascular protocols directly sharpens clinical triage decisions. For example, clinicians can swiftly distinguish stable haematomas from those under active expansion pressure. Patients with dynamic spot signs on multiphase imaging require aggressive physiological management, including rapid blood pressure control and intensive care admission.

Conversely, patients with low expansion risk may avoid overtreatment or unnecessary transfer, optimizing critical care bed allocation. In Indian hospital settings, where stroke units face substantial logistical constraints and varying imaging resources, implementing standardized vascular protocols can substantially streamline clinical management. However, clinicians must recognize that no current score achieved an excellent c-statistic above 0.90 in this cohort. Therefore, risk stratification tools should complement clinical judgment rather than dictate management in isolation. Multidisciplinary teams must evaluate ongoing anticoagulation, renal parameters, and haemodynamic stability alongside imaging metrics. Through this integrated approach, acute stroke care teams can maximize safety and optimize long-term functional recovery.

Future Directions and Score Refinement

Although multiphase imaging offers marked prognostic advantages, existing prediction scores still require systematic refinement. Future predictive algorithms must incorporate quantitative computational metrics, automated lesion volumetry, and artificial intelligence-assisted spot sign detection. In particular, automated algorithms can minimize human interobserver variability and deliver instantaneous risk calculations at the scanner console.

Furthermore, integrating systemic biomarkers, such as admission glucose, platelet function, and inflammatory indices, could bridge the remaining diagnostic performance gap. Clinical researchers should also evaluate whether score-guided treatment pathways translate directly into improved functional outcomes in prospective randomized trials. While advanced imaging protocols consume additional scanner time and modest radiation doses, modern scanners execute multiphase acquisitions within seconds. Therefore, stroke centres worldwide should actively upgrade emergency imaging pathways to incorporate multiphase computed tomography angiography. By refining prognostic scores and standardizing acquisition protocols, physicians can better tailor neuroprotective interventions and minimize secondary brain injury after spontaneous haemorrhage.

What is the clinical definition of haematoma expansion in acute intracerebral haemorrhage?

In acute intracerebral haemorrhage, clinical studies define primary haematoma expansion as an absolute volume increase of at least 6 mL or a relative enlargement of 33 percent or more from baseline. Additionally, researchers define severe expansion as haematoma growth exceeding 12.5 mL or a 66 percent volumetric increase.

Why does multiphase CT angiography outperform single-phase imaging in predicting expansion?

Multiphase CT angiography captures arterial, peak venous, and late venous vascular phases. Consequently, it detects active iodinated contrast extravasation that develops gradually over several seconds. Single-phase scans frequently miss this delayed contrast leakage. Therefore, multiphase imaging demonstrates superior biological sensitivity and yields significantly higher c-statistic scores for risk prediction.

Can non-contrast CT markers reliably substitute for vascular imaging when assessing expansion risk?

Non-contrast CT markers, such as the blend sign and black hole sign, provide valuable preliminary clues when contrast imaging is unavailable. However, non-contrast scores demonstrated lower discrimination, yielding c-statistics between 0.516 and 0.674. Consequently, clinicians should not rely solely on unenhanced scans when advanced vascular angiography protocols are accessible.

Disclaimer: This content is for informational and educational purposes only and is not intended to serve as medical advice, diagnosis, or treatment. Healthcare professionals should rely on their clinical judgment and reference validated institutional protocols. The therapeutic landscape evolves continuously; therefore, readers should verify findings with current peer-reviewed evidence and regulatory alerts. Refer to the latest local and national guidelines for clinical practice.

References

1. Pensato U et al. Stepwise improvement in intracerebral haematoma expansion prediction with advanced imaging: a comprehensive comparison of existing scores. Stroke Vasc Neurol. 2025 Dec 23. doi: 10.1136/svn-2024-003988. PMID: 40194847.
2. Demchuk AM, Dowlatshahi D, Rodriguez-Luna D, et al. Prediction of haematoma growth and outcome in patients with intracerebral haemorrhage using the CT-angiography spot sign (PREDICT): a prospective observational study. Lancet Neurol. 2012;11(4):307-314. doi: 10.1016/S1474-4422(12)70038-8.
3. Morotti A, Boulouis G, Dowlatshahi D, et al. Standards for detecting, interpreting, and reporting noncontrast computed tomographic markers of intracerebral hemorrhage expansion. Ann Neurol. 2019;86(4):480-492. doi: 10.1002/ana.25563.