Abstract
1594
Objectives: Characterization of novel radioligands as imaging biomarkers in clinical studies requires objective and reproducible regional brain analyses applied to various brain regions for many potential radioligands in groups of research subjects. To meet this image processing burden, we have developed a MATLAB library of scripts which automatically creates time activity curves for selected regions of interest (ROIs) extracted from human PET and SPECT images of new radioligands for input into the PMOD modeling package.
Methods: Our Objective and Generalized new Radioligand Evaluation (OGRE): 1) accepts the camera manufacturer supplied reconstructed images (attenuation un-corrected, for SPECT), 2) registers, in 3D, the PET or SPECT transverse slices to the AAL template available from MRIcro, 3) applies an attenuation correction for SPECT data, and 4) applies template's ROIs, and extracts the PET or SPECT intensities over all time points of the dynamic acquisitions, 5) corrects for radioactive decay, 6) computes ROI ratios to a selected reference ROI, 7) displays the TACs, and finally, 8) forwards the TAC data to the commercially available PMOD software package for compartmental pixel-wise modeling.
Results: The TACs obtained from the automated OGRE analysis are equivalent to those obtained by manual ROI placement. OGRE facilitates the required input to the PMOD modeling program.
Conclusions: For just one new radioligand, as many as 8-12 dynamic PET or SPECT scans need to be analyzed; under a FDA approved eIND, typically tens of human subjects are studied. Hence, using OGRE reduces the total required analysis time from hours to minutes. Furthermore, a very high degree of reproducibility and objectivity in evaluating new ligands is achieved by using this automated method.
- Society of Nuclear Medicine, Inc.