Abstract
1960
Objectives The use of exogenous CO2 inhalation (hypercapnea) has been proposed as an alternative method to induce hyperemia for stress myocardial perfusion imaging (MPI). Breathing rate and tidal volume are increased dramatically in response to CO2, inducing a mismatch between PET and CT attenuation correction scans. Our objective is to determine the effect of respiratory gating and phase-matched attenuation correction on the uniformity of regional myocardial blood flow quantification during hypercapnia-induced stress imaging with dynamic rubidium-82 PET.
Methods List-mode data from rest/ hypercapnea-stress scans in healthy normal subjects (N=7) were gated into four phases. Dynamic images from each phase, as well as the full ungated series, were reconstructed and analyzed for blood flow quantification using FlowQuant. Polar-map uniformity was measured as SD/mean of the 460 sectors, and differences between the four phases were evaluated using one-way ANOVA and post-hoc t-testing. Two-factor ANOVA was used to evaluate the regional variation in blood flow values using a 5-segment model.
Results No significant differences in blood flow uniformity were found between respiratory phases at rest (p=0.75). For stress scans, significant differences in uniformity were found between the phases (p = 0.017). Post-hoc testing revealed no difference between the two inspiration phases, nor between the two expiration phases, but there were differences between the inspiration and expiration phases (46% vs 21%; p=0.002). Segmental ANOVA showed no effect due to inspiration vs expiration (p=0.49), but highly significant effects between segments (p < 0.001) and significant interaction of respiratory phase and segment (p < 0.001). The highest source of variability in blood flow values occurred in the posterior wall (over-estimation of 47%), as expected due to diaphragm motion-induced attenuation artifact. With phase-matched attenuation correction at end-expiration there was an average decrease in variability of blood flow values within segments by a factor 3 compared to end-inspiration, and an overall decrease in variability across all segments by a factor of 6.
Conclusions Respiratory gating and phase-matched attenuation correction decreases the variability of regional blood flow estimates, improving uniformity in healthy normals during hypercapnea -induced stress imaging with dynamic rubidium-82 PET.