Impact of μ-map Processing and Transmission Scan Count Statistics on Quantification of PET Pig Brain Scans - and Temporal Variation of Scatter Correction Induced by μ-map Mismatch

Aim: In this work, we evaluate the two μ-map methods, total variation transmission scan processing and reconstruction (TXTV) and maximum a posteriori TX reconstruction (MAPTR), at two different transmission scan acquisition speeds (with different count statistics) against gold standard CT-based μ-maps on PET pig brain scans. Material and methods: Nine dynamic High Resolution Research Tomograph (HRRT) PET pig brain scans with different tracers reconstructed in 5 versions using different μmaps were registered to a pig brain atlas. The 5 different μ-maps were speed 50 or speed 10 transmission scan (TX), each reconstructed and processed using either TXTV or MAP-TR, and a rigidly co-registered gold standard μ-map from a same day CT scan. From time activity curves (TACs) on relevant volumesof-interest (VOIs) relative differences from the CT-based PET to the 4 HRRT TX-based PET and areas under the curves (AUC) were generated along with μ-map profile plots. Results: AUCs showed 3-16% (TXTV) and 2-10% (MAP-TR) difference from CT-based μ-maps on 11 VOIs with almost no difference between using speed 50 or 10 transmission scan acquisitions. Similar average differences were found in relative difference, but with an unexpected time dependant variability, which we found to correspond with a similar change in scatter fractions over time. This correspondence was likely explained by PET-to-μ-map mismatches. We primarily saw this when using CT-based μ-maps: The fixation of the pig's head and neck was changed when moving it non-rigidly to the CT scanner, which again caused a considerable mismatch to HRRT PET. Conclusion: Our results suggest that MAP-TR μ-maps made from speed 50 transmission scans is so far the best method for attenuation correction of HRRT PET pig brain images. However, difficulties with relocation between imaging modalities because of different head/neck fixation on the HRRT and (PET/)CT scanners may have hampered the evaluation of HRRT TX methods for pig brain scans. In particular, a non-rigid mapping would have been required to match the CT-based u-map with HRRT data, thus increasing the risk of residual error in the reference (gold standard) data.