Objectives: We sought to evaluate the accuracy of the perfusion computed tomography (PCT) deconvolution-based brain perfusion measurements and the lesions' (infarct and penumbra) volumetric with regard to arterial input function (AIF) selection in patients with acute stroke.
Materials and methods: Eighteen consecutive patients with symptoms of acute stroke underwent PCT at admission. Follow-up magnetic resonance imaging was obtained in all patients after 3.6 +/- 1.7 days (range, 1.5-6 days). PCT maps were generated focusing on the anterior cerebral artery (ACA) and branches of the middle cerebral artery (MCA) ipsilateral and contralateral to the ischemic lesion as AIFs. Infarct, penumbra, and total ischemic lesion were delineated on cerebral blood flow (CBF) maps. CBF, cerebral blood volume (CBV), and mean transit time (MTT) were calculated in the ischemic regions as provided by the 3 different AIFs, the normality test was applied for the obtained parameters, and the values were correlated (Pearson's correlation coefficient). Volumes of the ischemic regions (as obtained by the different AIFs) also were correlated and compared (paired t test) to the follow-up infarct volume.
Results: The CBF and CBV values obtained by the different AIFs in the infarct, penumbra, and total ischemic lesion were significantly correlated (r=0.94-0.96, P<or=0.01). Only in the infarct region calculated MTT values were correlated (r=0.88-0.91, P<0.05) between the different AIFs groups. High correlation coefficients (r=0.79-0.91, P<0.001) were observed between the admission PCT infarct and total ischemic volume and the MRI follow-up infarct volume. ACA as AIF provided the best correlations (r=0.91, P=0.0002) with the follow-up measurements. No statistically significant difference was found between the 3 different AIF-estimated admission total ischemic volumes and the follow-up infarct volume.
Conclusions: The AIF selection in the ACA as well as in the ipsilateral (to the hypoperfused area) or contralateral branches of the MCA has no statistically significant impact on the calculation of the CBF, CBV values, and the volume estimation of the ischemic region in the acute stroke patients.