Quantification of amyloid beta and tau PET without a structural MRI

Abstract

Introduction: Relying on magnetic resonance imaging (MRI) for quantification of positron emission tomography (PET) images may limit generalizability of the results. We evaluated several MRI-free approaches for amyloid beta (Aβ) and tau PET quantification relative to MRI-dependent quantification cross-sectionally and longitudinally. Methods: We compared baseline MRI-free and MRI-dependent measurements of Aβ PET ([18F]florbetapir [FBP], N = 1290, [18F]florbetaben [FBB], N = 290) and tau PET ([18F]flortaucipir [FTP], N = 768) images with respect to continuous and dichotomous agreement, effect sizes of Aβ+ impaired versus Aβ– unimpaired groups, and longitudinal standardized uptake value ratio (SUVR) slopes in a subset of individuals. Results: The best-performing MRI-free approaches had high continuous and dichotomous agreement with MRI-dependent SUVRs for Aβ PET and temporal flortaucipir (R2 ≥0.95; ± agreement ≥92%) and for Alzheimer's disease–related effect sizes; agreement was slightly lower for entorhinal flortaucipir and longitudinal slopes. Discussion: There is no consistent loss of baseline or longitudinal AD-related signal with MRI-free Aβ and tau PET image quantification.