[A study of rCBF measurement with autoradiography (ARG) method using N-isopropyl-p-[123I]iodoamphetamine (IMP) and SPECT--comparison of rCBF values between look-up table (TLU) and ARG methods, and evaluation of venous blood samplings as a substitute for arterial blood sampling].

: Regional cerebral blood flow (rCBF) values obtained by the TLU method with two 123I-IMP SPECT scans and one point arterial blood sampling and rCBF obtained by the ARG method with one 123I-IMP SPECT scan and fixed distribution volume (Vd) values were compared in 17 cases. A case with post ischemic hyperperfusion or luxury perfusion was not observed in our cases. The correlation coefficients between rCBF values and Vd values obtained by the TLU method were 0.49 (p < 0.001) in 184 ROI without hypoactive areas on the early image, and 0.61 (p < 0.001) in 207 ROI with hypoactive areas, respectively. A high rCBF value with a low Vd value was not observed in any region. Mean Vd value was 44.0 +/- 7.0 (mean +/- SD) in all regions. The correlation coefficients between rCBF values using the TLU method and those using the ARG method with Vd fixed at 44 and 50 were also 0.98. Error of the rCBF value was larger in the region of high rCBF, however, noticeable error of the rCBF value was not observed in the ARG method. The ARG method is more convenient for quantifying rCBF. Venous blood radioactivity at 10 min after 123I-IMP infusion was smaller than arterial blood radioactivity, and the blood activity in the distal vein was larger than that in the proximal vein. The ratio of venous blood activity to arterial blood activity was 0.92 +/- 0.04 (mean +/- SD) at the back of the hand, however, the ratio was a variant in each case. Arterial sampling was thought to be a reliable method to obtain more stable and precise rCBF.