Diagnostic accuracy of a revised computed tomography angiography score for brain death confirmation, combining supra-tentorial arteries and infra-tentorial veins

Abstract

Purpose: The 4-point score is the corner stone of brain death (BD) confirmation using computed tomography angiography (CTA). We hypothesized that considering the superior petrosal veins (SPVs) may improve CTA diagnosis performance in BD setting. We aimed at comparing the diagnosis performance of three revised CTA scores including SPVs and the 4-point score in the confirmation of BD.

Methods: In this retrospective study, 69 consecutive adult-patients admitted in a French University Hospital meeting clinical brain death criteria and receiving at least one CTA were included. CTA images were reviewed by two blinded neuroradiologists. A first analysis compared the 4-point score, considered as the reference and three non-opacification scores: a "Toulouse score" including SPVs and middle cerebral arteries, a "venous score" including SPVs and internal cerebral veins and a "7-score" including all these vessels and the basilar artery. Psychometric tools, observer agreement and misclassification rates were assessed. A second analysis considered clinical examination as the reference.

Results: Brain death was confirmed by the 4-score in 59 cases (89.4 %). When compared to the 4-score, the Toulouse score displayed a 100 % positive predictive value, a substantial observer agreement (0.77 [0.53; 1]) and the least misclassification rate (3.03 %). Results were similar in the craniectomy subgroup. The Toulouse score was the only revised test that combined a sensitivity close to that of the 4-score (86.4 % [75.7; 93.6] and 89.4 % [79.4; 95.6], p-value < 0.001, respectively) and a substantial observer agreement.

Conclusions: A score including SPVs and middle cerebral arteries is a valid method for BD confirmation using CTA even in patients receiving craniectomy.

Keywords: Brain death; Computed tomography angiography; Data accuracy; Psychometrics.