Parametric imaging of myocardial blood flow with ¹⁵O-water and PET using the basis function method

Regional myocardial blood flow (MBF) can be measured with ¹⁵O-water and PET using the 1-tissue-compartment model with perfusable tissue fraction, which provides an MBF value that is ftree from the partial-volume effect. Studies with ¹⁵O-water have several advantages, such as the ability to repeat a scan. However, because of the short scanning time and the small distribution volume of ¹⁵O-water in the myocardium, the image quality of ¹⁵O-water is limited, impeding the computation of MBF and perfusable tissue fraction at the voxel level. We implemented the basis function method for generating parametric images of MBF, perfusable tissue fraction, and arterial blood volume (V_a) with ¹⁵O-water and PET. The basis function method linearizes the solution of the 1-tissue-compartment model, Which results in a computationally much faster method than the conventional nonlinear least-squares fitting method in estimating the parameters. Methods: To validate the basis function method, we performed a series of PET studies on miniature pigs (n = 7). After acquisition of the transmission scan for attenuation correction and the ¹⁵O-CO scan for obtaining the blood-pool image, repeated PET scans with ¹⁵O-water were obtained with varying doses of adenosine or CGS-21680 (selective adenosine A_2a receptor agonist). MBF, perfusable tissue fraction, and V_a values of the myocardial region for each scan were computed using the basis function method and the nonlinear least-squares method, and the parameters estimated by the 2 methods were compared. Results: MBF images generated by the basis function method demonstrated an increase in blood flow after administration of adenosine or CGS-21680. The MBF values estimated by the basis function method and by the nonlinear least squares method correlated strongly. Conclusion: The basis function method produces parametric images of MBF, perfusable tissue fraction, and V_a with ¹⁵O-water and PET. These images will be useful in detecting regional myocardial perfusion abnormalities.