Applying a [C]Raclopride Template to Automated Binding Potential Estimation in HRRT Brain PET

Performing kinetic analysis of [C]raclopride PET data typically involves user intervention to identify key volumes of interest, such as the cerebellum for the reference region and the caudate and putamen for regions where the binding potential (BP) needs to be estimated. In many PET centres, this process is neither automated nor standardized, possibly producing discrepancies between centres. Conventionally, MR anatomical images are used to identify the key volumes of interest, but this is difficult to automate robustly, and user intervention can sometimes be required. This work considers instead the use of an anatomically labeled [C]raclopride template, formed from multiple subjects, which has the key advantages of low noise, good resolution and having a highly similar spatiotemporal intensity distribution to any given single subject raclopride scan. This makes the template an excellent target for automated image registration and segmentation. We present a methodology which works on post-reconstruction images, demonstrating an automated and consistent way of identifying key regions of interest (ROIs) and determining binding potential without any MR image or user intervention. The performance of the methodology is evaluated using simulated and real [C]raclopride data. The simplified reference tissue model with the basis function method (SRTM-BFM) was used for the kinetic modeling.