[Quantitation of regional cerebral blood flow using 123I-IMP from a single SPECT scan and a single blood sampling--analysis on statistical error source and optimal scan time].

Recently, a method has been proposed to measure quantitative regional cerebral blood flow (rCBF) in man using N-isopropyl-p-[123I]-iodoamphetamine (IMP) and single photon emission tomography (SPECT). In this method (IMP-ARG method), a functional map of rCBF was calculated from a single SPECT scan data set, in which effects of the clearance of IMP from the brain was taken into account by employing the 2 compartment model. To avoid the procedures of frequent and/or continuous sampling of the arterial blood that was required in determination of the input function, use of a standard input function and calibrating it by one blood sampling were validated. The present study was intended to investigate the sensitivity to various sources of errors in the IMP-ARG method such as (1) effects of individual difference of the arterial input function, (2) effects of ambiguity of the regional distribution volume (Vd) of IMP, and (3) effects of inaccuracy of the SPECT measurement. It was shown in the present simulation study that errors in the calculated rCBF values were dependent on the SPECT mid-scan time (MST), and that the minimum error corresponded to the MST of approximately 30 min after the IMP infusion. With this optimal MST, errors in rCBF due to the individual difference of the input function was approximately 10%, and the errors due to the ambiguity of the Vd were approximately 8%. The total accuracy of the calculated rCBF in the IMP-ARG method was, therefore, estimated as approximately 13%. Although a relatively large error was expected at a high rCBF area, this study strongly suggested the IMP-ARG method being useful and accurate for providing the quantitative rCBF map for a clinical use.