Mw. Parsons et al., Perfusion magnetic resonance imaging maps in hyperacute stroke - Relative cerebral blood flow most accurately identifies tissue destined to infarct, STROKE, 32(7), 2001, pp. 1581-1587
Background and Purpose-In ischemic stroke, perfusion-weighted imaging (PWI)
and diffusion-weighted imaging (DWI) provide important pathophysiological
information. A PWI > DWI mismatch pattern suggests the presence of salvagea
ble tissue. However, improved methods for distinguishing PWI > DWI mismatch
tissue that is critically hypoperfused from benign oligemia are required.
Methods-We investigated the usefulness of maps of relative cerebral blood f
low (rCBF), volume (rCBV), and mean transit time (rMTT) to predict transiti
on to infarction in hyperacute (<6 hours) stroke patients with PWI > DWI mi
smatch patterns. Semiquantitative color-thresholded analysis was used to me
asure hypoperfusion volumes, including increasing color signal intensity th
resholds of rMTT delay, which were compared with infarct expansion, outcome
infarct size, and clinical status.
Results-Acute rCBF lesion volume had the strongest correlation with final i
nfarct size (r=0.91, P <0.001) and clinical outcome (r=0.67, P <0.01). Ther
e was a trend for acute rCBF > DWI mismatch volume to overestimate infarct
expansion between the acute and outcome study (P=0.06). Infarct expansion w
as underestimated by acute rCBV > DWI mismatch (P <0.001). When rMTT lesion
s included tissue with moderately prolonged transit times (mean delay 4.3 s
econds, signal intensity values 50% to 70%), infarct expansion was overesti
mated. In contrast, when rMTT lesions were restricted to more severely prol
onged transit times (mean delay 6.1 seconds, signal intensity > 70%), these
regions progressed to infarction in all except 1 patient, but infarct expa
nsion was underestimated (P <0.001).
Conclusions-The acute rCBF lesion most accurately identified tissue in the
PWI > DWI mismatch region at risk of infarction. Color-thresholded PWI maps
show potential for use in an acute clinical setting to prospectively predi
ct tissue outcome.