Quantitative Measurements of Cerebral Blood Flow in Patients with Unilateral Carotid Artery Occlusion: a PET and MR Study

P15 It has been demonstrated that a quantitative measure of CBF can be obtained with MR when the acquired images are processed by the singular value decomposition (SVD) approach. However, two confounding factors have hampered the success of the SVD approach in clinical settings. First, a constant cardiac output is assumed, potentially leading to variability in the estimates of CBF among subjects. Second, the choice of the arterial input function could be contaminated by partial volume effect. In this study, normal volunteers (n=6) and patients with unilateral carotid artery occlusion (n=5) were studied. In the patient group, rCBF estimates were obtained from both PET and MR. In contrast, only MR estimated rCBF was obtained in the volunteer group. A mean rCBF of 73.6 ± 5.9 ml/100gm/min and 28.1 ± 4.8 ml/100gm/min was obtained for the GM and WM, respectively, in the volunteer group, in excellent agreement with the reported results in the literature via PET. However, when MR estimated rCBF was correlated with those obtained from PET in the patient group, no clear relationships were observed (slope=0.44 and r=0.42, unfilled circles in the figure). This finding is perhaps not surprising given the above mentioned confounding factors associated with the SVD approach. Therefore, a correction factor was derived based on the mean area of the concentration time curves of the superior sagittal sinus from all volunteers. Results are shown in the figure (filled diamonds) after the application of the correction factor. A slope of 1.1 and an r=0.8 was obtained, suggesting that the above mentioned confounding factors can be minimized and a highly linear relationship between the MR and PET estimates of rCBF can be obtained.