Validation of Population-Based Arterial Input Function for Kinetic Analysis of C-11-PK11195 PET Images of the Brain

69 Objectives: Use of an arterial input function (AIF) is the gold standard for kinetic modeling of dynamic PET images. The aim of this study was to assess the accuracy of kinetic analysis of 11C-PK11195 PET of the brain using a population-based AIF (PBAIF). The final goal is to possibly eliminate the need for AIF.\nMaterials: Arterial blood sampling was performed in 19 healthy controls who were administered 11C-PK11195. Each subject underwent 60 min of dynamic PET imaging of the brain on a Siemens mCTTM PET/CT (Knoxville, TN) starting simultaneously with the administration of 10.33±1.76 mCi of 11C-PK11195. Arterial blood sampling started concurrently with PET data acquisition and at 25 sec intervals for the first 10 min. Additional blood samples were withdrawn at 15, 20, 30, 45, and 60 min post-injection. Activity concentration of each blood sample was measured using a gamma counter and then the corresponding AIF was deduced. Six of the subjects underwent a repeat PET scan on the same day to assess the limits of agreement (LOA) of 11C-PK11195 kinetics. The PBAIF was generated as follows: The individual AIFs were each normalized by the corresponding injected dose and their peaks aligned. AIF curves were then averaged on a time point-by-time point basis. Metabolite correction was applied to both the subject-based AIF (SBAIF) and PBAIF. PBAIF’s were rescaled using the subject-specific injected dose. Patients were imaged on an MR. Brain structures were segmented on the MR T1 images using FreeSurfer v8.0 (Martinos Center for Biomedical Imaging, Charlestown, MA). PET dynamic images were then rigidly co-registered to the MR-T1 dataset, the MR volumes of interest (VOI) were overlaid on the dynamic PET dataset, and finally the respective time-activity-curves were extracted. Kinetic analysis was performed using PMOD v3.9 for the thalamus (T), gray matter (GM), and white matter (WM) using both a 1-tissue-2-compartment model with restricted K1/k2 that was previously determined from the brain cortex (VT2), as well as Logan plots (VTL). Kinetic analysis was performed using SBAIF and then repeated with PBAIF.\n Results: In the test-retest studies, the limits of agreement (LOA) as determined by Bland-Altman analysis, for T, GM, and WM were [-68.90%,28.11%], [-79.38%,59.69%] and [-72.25%,40.39%] for VT2 and [-64.3%,26.29%], [-56.33%,21.73%] and [-55.33%,20.16%] for VTL. Comparison of PBAIF and the SBAIF showed 2.72±13.32% and 14.96±9.7% difference in the peaks amplitudes and areas under the curve respectively. The variabilities in PBAIF- vs SBAIF- based VT2 and VTL were within the corresponding LOA’s for the three VOI respectively.\n Conclusions: The variability in PBAIF-based kinetic parameters of 11C-PK11195 in brain were within the corresponding LOA’s. Our results showed a PBAIF may be used for kinetic analysis of 11C-PK11195 in the brain.