Purpose: Using combined positron emission tomography (PET) and computerized tomography (CT) instrumentation, PET measurements of myocardial tracer uptake performed with CT attenuation correction may differ from estimates using 68Germanium transmission correction due to differences in respiratory motion during acquisition. The purpose of this study is to evaluate the effects of respiratory motion on the CT acquisition and emission corrected images, and to evaluate the correlation of diaphragm position with regional differences in myocardial 2-deoxy-2-[18F]fluoro-D-glucose (FDG) uptake in clinical studies.
Methods: A canine myocardial FDG-PET study was performed with controlled ventilation. Attenuation correction was performed with CT scans acquired at end expiration and end inspiration, and throughout multiple respiratory cycles with conventional 68Germanium transmission scan. The mean myocardial FDG activity was evaluated in multiple short axis regions of interest (n=40) using each of these three AC maps. Differences in emission during CT acquisitions were identified and expressed as bias (%) compared to 68Germanium corrected data. Ten patient studies with high myocardial FDG uptake were retrospectively selected from a clinical population referred for whole body oncology studies. All subjects had both CT and 68Germanium AC. After analysis for diaphragm misregistration defined by imaging and diaphragm position, subjects were divided into two groups: Group A controls (n=5) with no or mild misregistration, and Group B (n=5) with moderate or severe diaphragm misregistration. Regional emission bias (n=400 regions) from CT correction was defined by using the 68Germanium attenuation corrected emission as the standard.
Results: The canine study using end-expiration CT for attenuation correction showed regional overestimation of activity (1.8%+/-0.7% for inferior; 2.0%+/-0.5% for inferolateral) compared to the 68Germanium corrected images. Conversely, the study using end-inspiration CT attenuation correction showed underestimation (-3.9%+/-0.5% for inferior; -4.0%+/-0.6% for inferolateral) of myocardial activity compared to 68Germanium corrected images. In subjects, Group B showed significant relative underestimation of FDG myocardial activity compared to Group A in the regions adjacent to the diaphragm including the inferior (P=0.0003), inferoseptal (P=0.008), and inferolateral (P<0.0001) regions.
Conclusions: In canine myocardium, differences in respiration influenced CT attenuation maps and subsequent CT attenuation corrected PET images in the inferolateral and lateral regions. In clinical PET-CT studies, diaphragm misregistration is associated with relative decreased emission activity in inferior, inferoseptal, and inferolateral walls. Nonuniformity of bias in the emission data can affect quantitative accuracy, and therefore, the interpretation of myocardial viability. Further studies are required to determine if the frequency of these findings warrants the use of 68Germanium transmission attenuation correction in myocardial FDG-PET. The quantitative differences between these techniques were typically modest.