Objective: The purpose of our study was to quantify arterial and portal hepatic arterial blood flows.
Materials and methods: Four models were developed using PET. The first model consisted of the components of the liver and the portal system. The second applied "curve analysis" to this model. The third model introduced a portosystemic shunt factor, whereas the last model introduced a coefficient for circulation time within the portal organs. In 51 patients (34 men and 17 women), PET scans of the liver were performed using the H2 15O dynamic method.
Results: Under all four models, the arterial and portal hepatic arterial blood flows of 504 regions of interest were calculated using the nonlinear least-squares method, and results were compared by the sum of the squares of errors. Additionally, results from the H2 15O dynamic method were compared by results from the C15O2 steady-state method.
Conclusion: Of the four models, the last model produced curves with the best fit. When hepatic blood flow was quantified using PET and the H2 15O dynamic method, a model applying "curve analysis" and components related to portosystemic shunting and circulation time was found to be most accurate.