A systematic comparison of kinetic modelling methods generating parametric maps for [(11)C]-(R)-PK11195

[(11)C]-(R)-PK11195 is presently the most widely used radiotracer for the monitoring of microglia activity in the central nervous system (CNS). Microglia, the resident immune cells of the brain, play a critical role in acute and chronic diseases of the central nervous system and in host defence against neoplasia. The purpose of this investigation was to evaluate the reliability and sensitivity of five kinetic modelling methods for the formation of parametric maps from dynamic [(11)C]-(R)-PK11195 studies. The methods we tested were the simplified reference tissue model (SRTM), basis pursuit, a simple target-to-reference ratio, the Logan plot and a wavelet based Logan plot. For the reliability assessment, the test-retest data consisted of four Alzheimer's patients that were scanned twice at approximately a six-week interval. For the sensitivity assessment, comparison of [(11)C]-(R)-PK11195 binding in Huntington's disease (HD) patients and normal subjects was performed using a group contrast to localize significant increases in mean pixel volume of distribution (VD) in HD. In all instances, a reference region kinetic extracted by a supervised clustering technique was used as input function. Reliability was assessed by use of the intra-class correlation coefficient (ICC) across a wide set of anatomical regions and it was found that the wavelet-based Logan plot, basis pursuit and SRTM gave the highest ICC values on average. The same methods produced the highest z-scores resulting from increases in mean striatal VD in HD patients compared with controls. The reference-to-target ratio and the Logan graphical approach were significantly less reliable and less sensitive.