IC-P3-215: Automated image processing algorithms for evaluating novel 123-Iodine-labeled beta-amyloid imaging agents MNI-187 and MNI-308 in Alzheimer's patients

Abstract

There is intense interest in an optimal radiopharmaceutical for multicenter SPECT or PET evaluation in drug development trials of amyloid-targeting therapies. Our development of a series of novel investigational radioligands binding to beta-amyloid (a-beta) in Alzheimer's disease (AD) requires an objective method for analyzing brain regional uptake for critical comparison between candidate radioligands in AD patients and controls. A library of MATLAB scripts were written to automatically generate time-activity SUV curves for selected brain volumes of interest (VOI) in AD patients and healthy controls participating in dynamic clinical imaging evaluation of the novel 123-I labeled a-beta imaging agents MNI-187 and MNI-308. Briefly, the Objective, Generalized Radiotracer Analysis (OGRE) program accepts reconstructed image volumes, registers to a standardized brain template (AAL), applies an attenuation correction, extracts count density data from specified brain regions, develops decay-corrected time-activity curves (TAC), calculates VOI:cerebellar ratios, and exports the TAC for compartmental pixel-wise modeling. Comparison was made to the SUV ratios over time obtained by an experienced image processor. The TACs developed by OGRE were equivalent to those obtained by manual VOI placement by an experienced image processor. In addition, repeat analyses of the same dataset by OGRE produced identical TACs and SUV ratios. It is feasible to incorporate highly objective and standardized image analysis methods in the efficient evaluation of novel imaging agents being evaluated in humans as an imaging biomarker of brain a-beta. Once validated such methods will be of significant utility in facilitating multicenter imaging evaluation in AD therapeutic trials of amyloid-targeting treatments.