Spuriously High Estimates of Cerebral Blood Flow by SPECT

W e want to report an error in interpreting d,l-hexamethylpropylene amine oxime (HMPAO) tomograms as cerebral blood flow (CBF) in subacute ischemic stroke cases. By simultaneously studying CBF by `33Xe dynamic single-photon emission computed tomography (SPECT), an obvious discrepancy was noted in that the HMPAO tomograms grossly exaggerated the reflow hyperemia following spontaneous reperfusion. To interpret 'Tc-HMPAO images in terms of CBF distribution, it is essential that the same or almost the same fraction of locally supplied tracer is fixed in the various regions, healthy as well as diseased, at the time imaging is performed. This fixation is due to extraction across the blood-brain barrier (BBB) and to retention of the extracted tracer molecules, a retention caused by chemical conversion to a hydrophilic metabolite that is unable to cross cell membranes as well as the BBB. A constant degree of fractional fixation in all brain regions of 40-50% has been found in humans.1 A few series comparing 'Tc-HMPAO images with CBF tomography obtained by `33Xe as well as with C1502 positron emission tomography images have consistently shown a good agreement in normal humans and in various disease states including acute ischemic stroke,2-4 but the number of such comparisons is rather limited. Nevertheless, the impression gained from this factual basis is that 9Tc-HMPAO can be trusted to image CBF distribution after correction for a minor nonlinearity due to back-diffusion in the first few minutes.5 The fidelity of 9`Tc-HMPAO to map CBF has been taken for granted to such an extent that it has been used as a "gold standard" of CBF distribution in comparative studies with other SPECT tracers of CBF such as N-isopropyl-p-(1231)iodoamphetamine (123I-IMP)6 or 99mTc-N,N'-1,2-ethylenediylbis-L-cysteine diethyl ester (`9Tc-ECD).7 We report here three cases of focal hyperfixation of 9`Tc-HMPAO in patients with subacute stroke (2-3 weeks) with evidence of infarct reperfusion. By hyperfixation we mean a higher counting rate over the infarct (relative to that of the opposite side) than can be explained by CBF as measured by 133Xe tomography at the same time. This can only occur if the fractional fixation exceeds the normal level of 40-50%. In case 1, on October 26, 1992, a 62-year-old man suddenly developed massive right-sided hemiparesis and aphasia. On day 3, SPECT with 'Tc-HMPAO showed massive reduction of tracer uptake in the anterior half of the left middle cerebral artery (MCA) distribution territory. When restudied on day 15, we first measured CBF by 133Xe tomography. As shown in Figure 1, the infarct