Automatic determination of brain perfusion index for measurement of cerebral blood flow using spectral analysis and 99mTc-HMPAO.

Cerebral blood flow (CBF) can be non-invasively quantified using the brain perfusion index (BPI), determined from radionuclide angiographic data generated by technetium-99m hexamethylpropylene amine oxime ((99m)Tc-HMPAO). We previously reported the use of a spectral analysis (SA) method using (99m)Tc-HMPAO to calculate the BPI. In this report, we demonstrate an automatic method for determining the optimal BPI value and compare the optimal BPI values with the absolute CBF values measured using H(2)(15)O positron emission tomography (PET). Bilateral cerebral hemispheres of 11 patients with various brain diseases were examined using (99m)Tc-HMPAO. In the automatic SA procedure, the radioactivity curve for the aortic arch ( C (a)) was shifted by 0-10 s. The radioactivity curve for the brain ( C (b)) was estimated using the shifted C (a), and the error value between the actually measured and the estimated C (b) (Err) was calculated. When the Err was at a minimum, the BPI value was defined as optimal BPI. The difference in BPI from the optimal BPI was calculated as |BPI - optimal BPI| / optimal BPIx100 (%). In all participants, an H(2)(15)O PET examination was also performed, and the BPI values were compared with the absolute CBF values measured using H(2)(15)O PET (mCBF(PET)). The difference between BPI and the optimal BPI increased significantly from 4.87%+/-1.69% to 18.38%+/-3.93% (mean+/-SD) when the Err value increased. The optimal BPI value ( y) was well correlated with the mCBF(PET) value ( x) ( y=0.21 x-0.0075, r=0.800). Our results suggest that this automatic SA method provides an accurate estimate of BPI that can be used for the quantification of CBF using (99m)Tc-HMPAO SA.