Specific recognition of arteriovenous malformations using Xenon-133 RCBF technique.

With respect to the methodology of the atraumatic Xenon-133 technique the problem whether or not the proposed and introduced arterial artifact (AA) truely represents radiation from intravascular volume and to what extent it affects regional cerebral blood flow (rCBF) calculation is unresolved. We performed rCBF measurements in 22 patients with angiomas to clarify this issue in those patients known to have pathologically enlarged intracranial vessels. P4--the parameter suggested to represent the AA--as well as the conventional blood flow parameter for gray matter (F1) were compared to those of 50 volunteers using four criteria of abnormality: 1. intrahemispheric distribution, 2. interhemispheric differences of homologous detector pairs, 3. differences of mean hemispheric values, 4. visual evaluation of CBF maps. 19 of the 22 patients with angioma fulfilled at least two of the four criteria of abnormality, in comparison to 1 of 50 volunteers. P4's sensitivity for detecting angiomas proved to be higher (86%) than the perfusion parameters of gray matter. Focal increase of P4 proved to be highly specific for the presence of arteriovenous malformation (AVM, specifity 98%). A true arterial artifact exists in most instances in the presence of an AVM. Disregarding AA in the algorithm for calculation rCBF leads to an artificial overestimation of tissue flow in the region of the AVM.