Blazer: A Versatile And Efficient Workflow For Analyzing Pet/Mr Neuroimaging Data In Alzheimer'S Disease

583 Objectives: The semi-automated Biomarker Localization, Analysis, Visualization, Extraction, and Registration (BLAzER) workflow allows for rapid evaluation of MRI and amyloid- and tau-PET images, combining features well-suited for both clinical and research workflow. The purpose of the study was to assess BLAzER for regional brain PET/MR quantification using participants with different cognitive statuses from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) dataset. Additionally, we determined whether different segmentation inputs, FreeSurfer and Neuroreader, can be used and provide similar results with our workflow. Methods: 127 amyloid-PET and 55 tau-PET studies along with corresponding volumetric MRI were selected from ADNI. Subjects were selected to represent the spectrum of ADNI participants in terms of cognitive status (cognitively normal (CN), mild cognitive impairment (MCI), and Alzheimer’s disease (AD)), and age (cohorts aged 55-59, 60-69, 70-79, and ≥80). The BLAzER workflow begins with segmentation of MR images by FreeSurfer v6.0.0 (Boston, MA) or Neuroreader (Horsens, Denmark). Segmented output files along with source MR and PET scans are then visualized and quantified using an automated workflow on FDA-approved software (MIM v6.6.13, Cleveland, OH), enabling quality control to ensure optimal registration (Figure 1). Results: For efficiency, Neuroreader was faster than FreeSurfer on a per case basis (15 min/case vs. 12 hours/case) yet slower for total processing time for batches of studies (45.5 vs. 12 hours) due to parallelizing on FreeSurfer. For reproducibility, all MRI volumes and amyloid- and tau-PET standardized uptake value ratio (SUVR) measurements showed strong agreement between operators (ICC>0.97). For volumetric accuracy, BLAzER showed strong agreement with ADNI for frontal, temporal, cingulate, and parietal cortical regions (r>0.95, p 0.97, p<0.001, Figure 4). Finally, upon comparison, Neuroreader and FreeSurfer gave similar results (r=0.9841, p<0.001) despite systematic bias (4.0%) as cases showed similar dichotomization results after correcting with logistic regression (83 vs. 84 amyloid+ participants). Conclusions: BLAzER provides an efficient workflow for regional brain PET quantification. FDA approved components and the ability to visualize registration reduces barriers between research and clinical applications.