HAI Book 2025 - Flipbook - Page 650
Tiss, Amal
157
Findings from dynamic [18F]-MK6240 PET imaging after motion and
partial volume corrections
Amal Tiss1, Kuang Gong2, Cristina Lois1, Maeva Dhaynaut3, Nicolas J. Guehl3, Georges El
Fakhri3, Keith Johnson1, Jinsong Ouyang3
1
Department of Radiology, Massachusetts General Hospital and Harvard Medical School, Boston, MA, US
Department of Biomedical Engineering, University of Florida, Gainesville, FL, US
3
Department of Radiology and Biomedical Imaging, Yale University, New Haven, CT, US
2
Dynamic PET imaging of the tau tracer [18F]-MK6240 provides a comprehensive view of the tracer distribution.
However, the long acquisition time and small size of the regions involved in prodromal Alzheimer9s disease can
reduce the quantitative accuracy of PET imaging. We investigate the effects of inter- and intra- frame motion and
partial volume corrections (II-MC / PVC) on findings from cross-sectional and longitudinal studies.
We partitioned the PET data into frames according to established protocols and corrected intra-frame motion via
image-based registration. Additionally, we incorporated a patch-based kernel applied to the MR image into the
reconstruction scheme to perform PVC. We obtained a series of II-MC/PVC frames for our proposed method, IIMC frames, and the conventional inter-frame-only MC approach (IO-MC). We investigated the improvement of
each correction on a dataset of 55 scans from 33 subjects, including 22 longitudinal studies. Our measured
outcomes are DVR and R1 in 15 brain regions using the cerebellum grey matter as the reference region.
In cross-sectional studies, differences in DVR between the II-MC and IO-MC approaches exceeded 5% in 17% of
the datapoints. When analyzing the subject-specific data, we found that 51% of subjects exhibited better
quantitative results in at least one region. Similarly, when comparing II-MC/PVC to II-MC, 70% of subjects had at
least one region with a DVR difference greater than 5%. For R1, II-MC and II-MC/PVC improved results in 15% and
45% of subjects compared to IO-MC and II-MC, respectively. Figure 1 shows the comparisons between the three
methods. In longitudinal studies, no statistically significant differences were observed between the methods in
the surveyed regions due to the inter-subject variability (Figure 2). Parametric maps for one longitudinal study are
shown in Figure 3.
While improvements with II-MC/PVC are subtle, they can impact most of the enrolled subjects.
HAI2025 - 650
