Impact of attenuation correction strategies on the quantification of High Resolution Research Tomograph PET studies

van Velden, Floris H P, Kloet, Reina W, Van Berckel, Bart N M, Molthoff, Carla F M, de Jong, Hugo W A M, Lammertsma, Adriaan A, Boellaard, Ronald


Physics in Medicine and Biology 53 (1), p. 99-118


In this study, the quantitative accuracy of different attenuation correction strategies presently available for the High Resolution Research Tomograph (HRRT) was investigated. These attenuation correction methods differ in reconstruction and processing (segmentation) algorithms used for generating a micro-image from measured 2D transmission scans, an intermediate step in the generation of 3D attenuation correction factors. Available methods are maximum-a-posteriori reconstruction (MAP-TR), unweighted OSEM (UW-OSEM) and NEC-TR, which transforms sinogram values back to their noise equivalent counts (NEC) to restore Poisson distribution. All methods can be applied with or without micro-image segmentation. However, for MAP-TR a micro-histogram is a prior during reconstruction. All possible strategies were evaluated using phantoms of various sizes, simulating preclinical and clinical situations. Furthermore, effects of emission contamination of the transmission scan on the accuracy of various attenuation correction strategies were studied. Finally, the accuracy of various attenuation corrections strategies and its relative impact on the reconstructed activity concentration (AC) were evaluated using small animal and human brain studies. For small structures, MAP-TR with human brain priors showed smaller differences in micro-values for transmission scans with and without emission contamination (<8%) than the other methods (<26%). In addition, it showed best agreement with true AC (deviation <4.5%). A specific prior designed to take into account the presence of small animal fixation devices only very slightly improved AC precision to 4.3%. All methods scaled micro-values of a large homogeneous phantom to within 4% of the water peak, but MAP-TR provided most accurate AC after reconstruction. However, for clinical data MAP-TR using the default prior settings overestimated the thickness of the skull, resulting in overestimations of micro-values in regions near the skull and thus in incorrect AC for cortical regions. Using NEC-TR with segmentation or MAP-TR with an adjusted human brain prior showed less overestimation in both skull thickness and AC for these structures and are therefore the recommended methods for human brain studies.