A method was developed to calculate functional images of regional cerebral blood flow (rCBF) from a single scan using SPECT following intravenous 123I-N-isopropyl-p-iodoamphetamine (IMP) infusion. METHODS A two-compartment model that includes two parameters of rCBF and regional distribution volume of IMP (Vd) was employed to correct for clearance of IMP from the brain. Using a given input function and a fixed Vd value (30 ml/ml according to an analysis on dynamic SPECT data), a unique value of rCBF was calculated for each pixel of the SPECT image according to the table-look-up procedure. This technique was applied to 15 human subjects, and the calculated rCBF values were compared with those measured by PET. RESULTS A set of simulation studies demonstrated an optimal SPECT midscan time at 30 to 40 min postinjection of IMP, providing the minimal error sensitivity to the individual difference of the input function (rCBF values with an accuracy of +/- 10%). Another set of simulation suggested validity of fixing the Vd values, i.e., errors in calculated rCBF values were around +/- 7% for a change of Vd of +/- 10%. The measured rCBF values obtained from 15 human subjects were independent on the SPECT scan time. The calculated rCBF values also agreed well with those obtained by the nonlinear least-squares fitting analysis that were obtained from the dynamic SPECT scan and the frequent arterial blood sampling and measurement of lipophilic fraction for each sample (0.54 + 0.88x, r = 0.86), suggesting the validity of the simplified procedures in this method. CONCLUSIONS These observations suggested the validity of this method as a clinical tool for quantitative measurement of rCBF.