Validation of Noninvasive Tracer Kinetic Analysis of ¹⁸F-Florbetaben PET Using a Dual–Time-Window Acquisition Protocol

Accurate amyloid PET quantification is necessary for monitoring amyloid-β accumulation and response to therapy. Currently, most of the studies are analyzed using the static SUV ratio (SUVR) approach because of its simplicity. However, this approach may be influenced by changes in cerebral blood flow (CBF) or radiotracer clearance. Full tracer kinetic models require arterial blood sampling and dynamic image acquisition. The objectives of this work were, first, to validate a noninvasive kinetic modeling approach for ¹⁸F-florbetaben PET using an acquisition protocol with the best compromise between quantification accuracy and simplicity and, second, to assess the impact of CBF changes and radiotracer clearance on SUVRs and noninvasive kinetic modeling data in ¹⁸F-florbetaben PET. <b>Methods:</b> Using data from 20 subjects (10 patients with probable Alzheimer dementia and 10 healthy volunteers), the nondisplaceable binding potential (BP_ND) obtained from the full kinetic analysis was compared with the SUVR and with noninvasive tracer kinetic methods (simplified reference tissue model and multilinear reference tissue model 2). Various approaches using shortened or interrupted acquisitions were compared with the results of the full acquisition (0-140 min). Simulations were performed to assess the effect of CBF and radiotracer clearance changes on SUVRs and noninvasive kinetic modeling outputs. <b>Results:</b> An acquisition protocol using time windows of 0-30 and 120-140 min with appropriate interpolation of the missing time points provided the best compromise between patient comfort and quantification accuracy. Excellent agreement was found between BP_ND obtained using the full protocol and BP_ND obtained using the dual-window protocol (for multilinear reference tissue model 2, BP_ND [dual-window] = 0.01 + 1.00·BP_ND [full], <i>R</i>² = 0.97; for simplified reference tissue model, BP_ND [dual-window] = 0.05 + 0.92·BP_ND [full], <i>R</i>² = 0.93). Simulations showed a limited impact of CBF and radiotracer clearance changes on multilinear reference tissue model parameters and SUVR. <b>Conclusion:</b> This study demonstrated accurate noninvasive kinetic modeling of ¹⁸F-florbetaben PET data using a dual-window acquisition, thus providing a good compromise between quantification accuracy, scan duration, and patient burden. The influence of CBF and radiotracer clearance changes on amyloid-β load estimates was small. For most clinical research applications, the SUVR approach is appropriate. However, for longitudinal studies in which maximum quantification accuracy is desired, this noninvasive dual-window acquisition with kinetic analysis is recommended.