Quantitative measurement of regional cerebrovascular reactivity to acetazolamide using 123I-N-isopropyl-p-iodoamphetamine autoradiography with SPECT: validation study using H2 15O with PET.

UNLABELLED A simplified technique using (123)I-N-isopropyl-p-iodoamphetamine ((123)I-IMP) autoradiography (ARG) with SPECT has been proposed recently for quantifying regional cerebral blood flow (rCBF). To validate the accuracy of (123)I-IMP-ARG for quantifying regional cerebrovascular reactivity (rCVR) to acetazolamide, we compared rCVR determined using (123)I-IMP-ARG with that determined using H(2)(15)O PET. METHODS Thirty-nine patients with chronic stenoocclusive disease in a unilateral major cerebral artery underwent SPECT and PET studies before and after intravenous administration of acetazolamide. The rCBF images in the 4 conditions in each patient were calculated according to the ARG method. The same standard input function and the same distribution volume of 35 mL/mL were used in the calculation of rCBF images using the (123)I-IMP-ARG method at resting state and with acetazolamide challenge. One large cortical region of interest (ROI) for a unilateral middle cerebral artery territory was bilaterally determined on each standardized summed rCBF image. On the basis of the rCBF values in each ROI, rCVR to acetazolamide was calculated as follows: rCVR (%) = ([acetazolamide challenge rCBF - resting rCBF]/resting rCBF) x 100. RESULTS Significant correlation was observed between rCVR values obtained using (123)I-IMP-ARG and H(2)(15)O PET methods in the 78 ROIs examined in the 39 patients (r = 0.820; P < 0.0001). When a rCVR lower than the mean - 2 SD of values obtained in healthy volunteers (18.4% for (123)I-IMP-ARG and 18.2% for H(2)(15)O PET) was defined as reduced, and when the H(2)(15)O PET method was assumed to represent the true determinant of rCVR, (123)I-IMP-ARG was 90% sensitive and 92% specific and displayed an 87% positive predictive value for detecting patients with reduced rCVR. CONCLUSION These findings demonstrate that (123)I-IMP-ARG methods accurately quantify rCVR and can adequately define subgroups of patients with reduced rCVR.