Impact of acquisition time-window on clinical whole-body PET parametric imaging

Whole-body PET parametric imaging can combine the benefit of extended axial field-of-view (FOV) in multi-bed scans with that of generating time activity curves (TACs) in dynamic scans. We have recently proposed such a framework capable of delivering whole-body FDG Patlak images in clinically feasible scan times. The design of the acquisition protocol was limited to a single time-window and the standard Patlak graphical analysis method. However, the relatively long FDG half-life and uptake, compared to clinically acceptable acquisition time-windows, render the choice of this window critical. The major FDG kinetic components can be estimated from the early and intermediate TAC segments. On the contrary, at later time-windows, tumor contrast may be overall higher. In addition, the standard Patlak method does not account for tracer uptake reversibility, a property that becomes more apparent at later acquisition time-windows for certain tumors, thus increasing the probability for larger bias at later times. Consequently, the choice of the optimal time-window can be critical and should constitute an important design aspect of multi-bed dynamic protocols. In the present work we assessed the impact of a sliding acquisition time-window on whole-body FDG PET parametric images. This included incremental shift of a 6-pass acquisition time-window (~35min) along an extended scan period of 0-90min post injection, using both real patient kinetic data as well as realistic 4D simulations of the state-of-the-art Siemens Biograph mCT scanner. We also propose the selective application of a generalized Patlak method accounting for uptake reversibility. Our simulated and clinical results demonstrate that both Patlak methods (standard and generalized) result in enhanced tumor-to-background contrast as well as contrast-to-noise ratios with minimal bias at an early acquisition time window (10-45min post injection) with the generalized method exhibiting systematically superior performance.