Effect of reconstruction conditions on quantification of 18F-FDOPA and 11C-Raclopride PET as a biomarker of Parkinson’s disease

Objectives PET imaging of striatal dopaminergic function is useful for diagnosis of Parkinson’s disease and related disorders. Furthermore, 18F-FDOPA is valuable for therapeutic response monitoring, and 11C-raclopride (11C-RAC) is also used for evaluation of endogenous dopamine release by various stimulations. While quantitative measurement is essential in these studies, quantified uptake values are subject to spatial resolution and noise and depend on the PET camera model as well as the reconstruction parameters. The purpose of the present study is to clarify the effect of reconstruction parameters on the quantified value of striatal uptake using a phantom simulating 18F-FDOPA and 11C-RAC scans for optimization of the scanning conditions.

Methods We used the striatal phantom (Radiology Support Devices, Inc.) with 18F or 11C solutions of two types of striatal-to-background radioactivity ratio (SBR) simulating 18F-FDOPA and 11C-RAC uptake. The SBRs of the phantom models were 4.5 (right striatum) : 2.1 (left striatum) : 1.0 (background) and 3.0 : 2.0 : 1.0, respectively. The background activities of both phantom models were 3.0 kBq/mL. PET scans were performed for 30 minutes with a Discovery 690 PET/CT and were reconstructed by OSEM with variable reconstruction parameters ranging iterations 2-4, subsets 4-24, and Gaussian post-filter FWHMs of 2-6 mm. Regions-of-interest (ROIs) were manually placed to contour striatal and background regions on PET/CT fusion images. We evaluated striatum-to-background contrast and coefficient of variation of the background region (CV) as a function of iterative update (product of iteration and subset) and strength of Gaussian filter. We also exploratory evaluated the striatal-to-occipital ratio (SOR) and striatal-cerebellar-cortex ratio (SCR) of the human brain PET images acquired with 18F-FDOPA and 11C-RAC as a function of iterative update.

Results In the phantom study, the striatal contrast increased as the iterative update increased, and approached plateau after approximately 50 updates for both phantom models. The CV increased as the iterative update increased as a trade-off between contrast and noise, but was below 10% at the iterative update of 48. Based on the results, the optimal reconstruction conditions were empirically determined as iteration 4, subset 12, and Gaussian filter 2 mm FWHM. In the representative normal subject, the SOR of the 18F-FDOPA PET was 3.62 and the SCR of the 11C-RAC PET was 4.32, at the determined reconstruction conditions. The SOR and SCR of the human images converged at approximately 50 updates in the same way as the phantom study.

Conclusions We empirically determined the optimal reconstruction conditions for 18F-FDOPA and 11C-RAC PET. The quantified values of the human images were almost stable to the reconstruction condition with sufficient iterative updates. This supports the feasibility and reliability of 18F-FDOPA and 11C-RAC PET as an imaging biomarker of Parkinson’s disease.

• Download figure
• Open in new tab
• Download powerpoint