Carotid Artery Image-Derived Blood Time-Activity Curves on the NeuroEXPLORER: Initial Multitracer Validation Against
Tommaso Volpi1, Jean-Dominique Gallezot2, Shannan Henry2
1Yale University, New Haven, Connecticut tommaso.volpi@yale.edu.
Abstract:
Image-derived input functions are noninvasive alternatives to arterial blood sampling for PET kinetic modeling, allowing one to measure the whole-blood time-activity curve (BTAC). However, partial-volume effects have limited the use of carotid arteries (CA) to generate image-derived BTACs (ID-BTAC) for brain PET. The NeuroEXPLORER is a next-generation scanner with unprecedented spatial resolution, possibly allowing better CA ID-BTAC extraction. Methods: Twelve individuals were scanned on the NeuroEXPLORER with 5 tracers (18F-FDG, 18F-SynVesT-1, 18F-flubatine, 11C-LSN3172176, and 11C-PHNO), and arterial input functions were measured from blood samples. Common carotid (CC) and internal carotid (IC) regions-of-interest (ROIs) were manually segmented on early summed images. ROI diameters were dilated around the CA centerline to evaluate the partial-volume effects. Bias in ID-BTAC area under the curve (AUC) against BTAC was calculated to estimate the ability to recover the gold standard at early (0-10 min) and late times (>10 min). Using ID-BTACs (1-mm ROI diameter), kinetic estimates were generated (K 1, K i, and V T), and their bias against arterial input function-based estimates was evaluated. Results: For 1-2-mm ROI diameters, early AUC recovery was good (bias: CC, -8% ± 5%; IC, -3 ± 7%). The late AUC was accurately recovered for 18F-FDG (CC, -9%; IC, -5%), 11C-PHNO (CC, 7%; IC, 4%), and 18F-SynVesT-1 (CC, -9%; IC, -13%) but not for 18F-flubatine (CC, 18%; IC, 45%) and 11C-LSN3172176 (CC, 46%; IC, 100%); this poorer agreement corresponded to higher late-time brain-to-blood/face-to-blood activity ratios. The IC outperformed the CC in most cases. ID-BTAC biases translated into small K 1 errors (CC, 5% ± 10%; IC, 3 ± 14%) and larger K i and V T errors (CC, -3% ± 20%; IC, -12 ± 25%). Conclusion: A simple ID-BTAC extraction approach provided accurate recovery of the early BTAC for all tracers, minimizing spill-out. Late BTAC recovery was more variable, especially with higher background activity. These results highlight how CA ID-BTAC extraction with minimal bias is feasible with ultra-high-resolution brain-dedicated PET scanners.
