Brain death in children: is computed tomography angiography reliable as an ancillary test?

Abstract

The diagnosis of brain death is primarily clinical. Sometimes ancillary tests are needed. This study compared sensitivity and interobserver agreement of the 10-, 7- and 4-point CT angiography scoring systems for the diagnosis of brain death in children. CT angiography examinations of 50 pediatric patients with a clinical diagnosis of brain death were evaluated according to 10-, 7- and 4-point scoring systems. Images were evaluated by two radiologists who considered the vessel opacification first in the arterial phase (A0-V50) and then in the venous phase (A0-V50). We evaluated interobserver agreement for the assessment of vessel opacification and diagnosis of brain death. We compared the differences among brain death diagnoses between children with craniotomy-craniectomy defects, open fontanelles and preserved bone integrity. We subdivided children into two groups according to age: ≤ 2years and > 2years. We calculated sensitivities according to age groups. Using the clinical exam as the reference standard, we found sensitivities for 10-, 7- and 4-point scoring systems to be 70%, 88% and 92% in the A0-V50 method and 40%, 82% and 82% in the A50-V50 method, respectively. Percentage agreement between readers was 78% for the 7-point scale using the A0-V50 method and more than 90% for other scoring systems for both the A0-V50 method and the A50-V50 method. The sensitivity was much lower in children with open anterior fontanelles compared to the groups with preserved bone integrity and with a craniotomy-craniectomy defect. Just as in adult age groups, in children the 4-point scale appears to be more sensitive than the 10- and 7-point scales for CT angiography-based assessment of brain death. Because the scoring systems have similar sensitivities, they could be used as ancillary tests in pediatric cases.