An error in cerebral blood flow measured by 123I-IMP single-photon emission tomography and a microsphere model

We studied errors arising from back flow from brain to blood and blood-borne radioactivity of 123I-iodamphetamine (123I-IMP), both of which are neglected in the microsphere model. We used an analytical approach to compare cerebral blood flow (F), calculated at various-study times by microsphere analysis, and the unidirectional clearance from blood to brain (K 1) using two-compartment analysis, including parameters representing these assumptions explicitly. The values of back flow (K 2) and initial distribution volume (V 0) were very small, and did not seem to influence the estimation of F in the microsphere model. However, regression analysis between each F and K 1 revealed that F at 5 min was overrestimated because blood-borne radioactivity was neglected, while F at 53 min was underrestimated owing to the effect of k 2. The highest correlation of F with K 1 was observed at 13 min. We should be aware of regional overestimation on earlier images due to the blood-borne radioactivity, and underestimation on later images due to the back flow when analysing cerebral blood flow by 123I-IMP single photon emission tomography (SPECT) and the microsphere model.