CT Scan Hematoma Volume: A Surrogate Endpoint for the Effect of Treatment in Intracerebral Hemorrhage

Abstract

Intracerebral hemorrhage (ICH) represents approximately 15% of all strokes, with no approved therapeutic interventions available. The 30-day mortality rate is 35 to 50% and only 10 to 20% of survivors regain functional independence. Previous studies have indicated that substantial increase in hematoma volume occurs within the first hours after ICH onset. These studies also indicated that hematoma volume expansion is associated with neurological deterioration and was a critical determinant of 30-day mortality. Early interventions directed at reducing hematoma expansion would be of considerable value. However, no validated measurements exist to directly translate the impact of hematoma volume on clinical outcomes. Changes in hematoma volume as determined by computed tomography (CT) scan is a potential surrogate endpoint for clinical outcomes. A recent randomized, double-blind, parallel-group, placebo-controlled clinical trial (399 patients) evaluated the use of rFVIIa (recombinant activated coagulation Factor VII) to reduce expansion of ICH hematoma volume, also determined the inter- and intrareader variability of CT scan measurements. In addition, assessments of clinical outcomes (modified Rankin Score, mRS) and mortality were conducted.

A baseline CT scan was performed within 3 hours after symptom onset. Placebo or rFVIIa was administered within 1 hour after the baseline scan. CT scans were repeated at 24 hours and 72 hours post-treatment. Inter-reader and intrareader variability was assessed by comparison of ICH volumes measured on CT images by two independent neuroradiologists, masked to treatment.

A meta-analysis conducted on a hemostatic treatment–naïve ICH patient population pooled from four studies (N = 218) demonstrated that hematoma volume was an independent determinant of mortality (p < 0.0001) and mRS (p = 0.0003). The results of a clinical trial using rFVIIa indicated that CT scan measurements of ICH volumes showed excellent intraclass correlations for interreader (0.9569) and intrareader variability (0.9844). Additionally, treatment with rFVIIa resulted in significant reduction in hematoma volume (p = 0.01), mortality (p = 0.02) as well as improved clinical outcome (p = 0.004, mRS) for rFVIIa-treated patients compared to placebo. Therefore, change in hematoma volume as determined by CT scan has potential as a clinically relevant surrogate measure, with value in the study of early hemostatic interventions in acute neurologic settings where brain hemorrhage is of significant concern.