Comparison of MRI-based and PET-based image pre-processing for quantification of ¹¹C-PBB3 uptake in human brain

Purpose: Quantification of tau load using ¹¹C-PBB3-PET has the potential to improve diagnosis of neurodegenerative diseases. Although MRI-based pre-processing is used as a reference method, not all patients have MRI. The feasibility of a PET-based pre-processing for the quantification of ¹¹C-PBB3 tracer was evaluated and compared with the MRI-based method. Materials and methods: Fourteen patients with decreased recent memory were examined with ¹¹C-PBB3-PET and MRI. The PET scans were visually assessed and rated as either PBB3(+) or PBB3(−). The image processing based on the PET-based method was validated against the MRI-based approach. The regional uptakes were quantified using the Mesial-temporal/Temporoparietal/Rest of neocortex (MeTeR) regions. SUVR values were calculated by normalizing to the cerebellar reference region to compare both methods within the patient groups. Results: Significant correlations were observed between the SUVRs of the MRI-based and the PET-based methods in the MeTeR regions (r_Me = 0.91; r_Te = 0.98; r_R = 0.96; p < 0.0001). However, the Bland–Altman plot showed a significant bias between both methods in the subcortical Me region (bias: −0.041; 95% CI: −0.061 to −0.024; p = 0.003). As in the MRI-based method, the ¹¹C-PBB3 uptake obtained with the PET-based method was higher for the PBB3(+) group in each of the cortical regions and for the whole brain than for the PBB3(−) group (PET-based_Global: 1.11 vs. 0.96; Cliff's Delta (d) = 0.68; p = 0.04; MRI-based_Global: 1.11 vs. 0.97; d = 0.70; p = 0.03). To differentiate between positive and negative scans, Youden's index estimated the best cut-off of 0.99 from the ROC curve with good accuracy (AUC: 0.88 ± 0.10; 95% CI: 0.67–1.00) and the same sensitivity (83%) and specificity (88%) for both methods. Conclusion: The PET-based pre-processing method developed to quantify the tau burden with ¹¹C-PBB3 provided comparable SUVR values and effect sizes as the MRI-based reference method. Furthermore, both methods have a comparable discrimination accuracy between PBB3(+) and PBB3(−) groups as assessed by visual rating. Therefore, the presented PET-based method can be used for clinical diagnosis if no MRI image is available.