Kinetic Modeling of Dynamic PET-¹⁸F-FDG Atherosclerosis Without Blood Sampling

Atherosclerosis is a disease affecting the peripheral arteries characterized by a deposit of lipids and immune cells forming a plaque. The plaque can be calcified and be measured with CT, and it can be inflamed and be measured with PET and
¹⁸
F-FDG. Since artery walls are too small and are dominated by blood activity, the conventional compartmental model becomes inadequate to provide artery metabolism especially when using an image derived input function from an artery. The purpose of this study was to quantitatively evaluate glucose metabolism in arteries with a more appropriate compartmental modeling. Kinetic modeling of arteries was conducted using the classical
¹⁸
F-FDG compartmental model and a modification of this model without the need of blood sampling. The metabolic rate of glucose (MRG) was computed per artery segment in each image slice with both models. The ratio of calcification area (RCA) and Agatston calcification scores (ACSs) were automatically clustered with the automatic hierarchical K-means algorithm. The rate constants showed significant variation between the two models (P <; 0.05), however, Ki was with good agreement in both models (P > 0.05). Both models found that MRG values in nonmedication group were statistically significantly different from those in under-medication group.