BACKGROUND AND PURPOSE: Because cerebral perfusion imaging for acute stroke
is unavailable in most hospitals, we investigated the feasibility of a met
hod of perfusion scanning that can be performed rapidly during standard cra
nial CT. Our aim was to identify the scanning parameters best suited to ind
icate tissue at risk and to measure a perfusion limit to predict infarction
.
METHODS: Seventy patients who had suffered stroke and had undergone cranial
CT 0.5 to 12 hours (median, 3.75 hr) after the onset of symptoms participa
ted in the study. While undergoing conventional CT, each patient received a
bolus of iodinated contrast medium. Maps of time to peak (TTP), cerebral b
lood volume (CBV), and CBF were calculated from the resulting dynamically e
nhanced scans. These perfusion images were compared with follow-up CT scans
or MR images showing the final infarctions.
RESULTS: CBF maps predicted the extent of cerebral infarction with a sensit
ivity of 93% and a specificity of 98%. In contrast, CBV maps were less sens
itive and TTP maps were less specific and also showed areas of collateral f
low. Infarction occurred in all of the patients with CBF reduction of more
than 70% and in half of the patients with CBF reduction of 40% to 70%.
CONCLUSION: Dynamic CT perfusion imaging safely detects tissue at risk in c
ases of acute stroke and is a feasible method for any clinic with a third-g
eneration CT scanner.