Total intracranial hemorrhage volume measurement summating all compartments best in traumatic and nontraumatic intracranial bleeding

Abstract

Abstract Background and purpose The ANNEXA‐4 trial measured hemostatic efficacy of andexanet alfa in patients with major bleeding taking factor Xa inhibitors. A proportion of this was traumatic and nontraumatic intracranial bleeding. Different measurements were applied in the trial including volumetrics to assess for intracranial bleeding depending on the compartment involved. We aimed to determine the most reliable way to measure intracranial hemorrhage (ICrH) volume by comparing individual brain compartment and total ICrH volume. Methods Thirty patients were randomly selected from the ANNEXA‐4 database to assess measurement of ICrH volume by compartment and in total. Total and compartmental hemorrhage volumes were measured by five readers using Quantomo software. Each reader measured baseline hemorrhage volumes twice separated by 1 week. Twenty‐eight different ANNEXA‐4 subjects were also randomly selected to assess intra‐rater reliability of total ICrH volume measurement change at baseline and 12‐h follow up, performed by three readers twice to assess hemostatic efficacy categories used in ANNEXA‐4. Results Compartmental minimal detectable change percentages (MDC%) ranged between 9.72 and 224.13, with the greatest measurement error occurring in patients with a subdural hemorrhage. Total ICrH volume measurements had the lowest MDC%, which ranged between 6.57 and 33.52 depending on the reader. Conclusion Measurement of total ICrH volumes is more accurate than volume by compartment with less measurement error. Determination of hemostatic efficacy was consistent across readers, and within the same reader, as well as when compared to consensus read. Volumetric analysis of intracranial hemostatic efficacy is feasible and reliable when using total ICrH volumes.

Keywords

• neuroimaging
• neuroscience
• neurology
• stroke