Background: Changes in renal blood flow (RBF) may play a pathophysiological role in hypertension and kidney disease. However, RBF determination in humans has proven difficult. We aimed to confirm the feasibility of RBF estimation based on positron emission tomography/ computed tomography (PET/CT) and rubidium-82 (82Rb) using the abdominal aorta as input function in a 1-tissue compartment model.
Methods: Eighteen healthy subjects underwent two dynamic 82Rb PET/CT scans in two different fields of view (FOV). FOV-A included the left ventricular blood pool (LVBP), the abdominal aorta (AA) and the majority of the kidneys. FOV-B included AA and the kidneys in their entirety. In FOV-A, an input function was derived from LVBP and from AA; in FOV-B from AA. 1-tissue compartmental modeling was performed using tissue time activity curves generated from volumes of interest contouring the kidneys, where the renal clearance of 82Rb is represented by the K1 kinetic parameter. To investigate the correct interpretation of K1, we assumed to first estimate effective renal plasma flow (ERPF) by extrapolating clearance values (ml/min/cm3) to whole kidney values (ml/min) using the estimated total kidney volume. Thereafter, RPF was estimated from ERPF using an assumed extraction fraction (0.89). Lastly, RBF was estimated from RPF using measured haematocrit values. Intra-assay coefficients of variation and inter-observer variation were calculated.
Results: For both kidneys, K1 values derived from AA did not differ significantly from values obtained from LVBP, neither were significant differences seen between AA in FOV-A and AA in FOV-B, nor between the right and left kidneys. For both kidneys, the intra-assay coefficients of variation were low (~ 5%) for both input functions. The measured K1 of 2.04 ml/min/cm3 suggests an estimated total renal perfusion normalized to body surface area of 628 ± 95 ml/min/1.73 m2 and subsequently an estimated RBF of 1091 ± 162 ml/min/1.73 m2.