Abstract
2247
Introduction: Perfusion imaging is performed where disturbances in normal regional blood flow may impact the course of treatment, such as in coronary artery or peripheral vascular diseases, cerebral ischemia, or changes in tumor perfusion with therapy. Compared to commonly used perfusion tracers such as [15O]-water, [13N]-ammonia and [82Rb]-rubidium chloride, [11C]-butanol has superior extraction fraction at high flow rates. Therefore, it is a promising tracer for total-body perfusion mapping, such that blood flow to highly-perfused organs such as the kidneys and myocardium can be accurately assessed. Here, we present the initial methods, dosimetry, and kinetic analysis of an ongoing study assessing the use of [11C]-butanol for dynamic perfusion imaging across the entire body with the uEXPLORER total-body PET/CT system.
Methods: Three subjects were recruited and consented (2 male, 1 female, 54-66 years old) into this IRB-approved study. At each visit, the subjects received an intravenous bolus injection of 266-349 MBq of [11C]-butanol and underwent a 30-minute dynamic acquisition at rest. Two of the three subjects (1 male, 1 female) were healthy volunteers and underwent a test-retest study paradigm, where both acquisitions were performed under the same conditions no more than two weeks apart. The third subject had a history of peripheral artery disease (PAD) and underwent a single scan at rest. Dynamic images were reconstructed with an isotropic voxel size of 4.0 mm with a framing of 12x5s, 6x10s, 6x30s, and 5x300s for dosimetry estimates, and an isotropic voxel size of 2.3 mm with framing of 30x2s, 6x10s, 2x30s, 1x60s was used for total-body kinetic modeling. Whole-organ regions of interest (ROIs) were delineated in several tissues using PMOD. Dosimetry estimates were performed using OLINDA 2.0 and the MIRD method (Stabin and Farmer, 2012). In order to estimate regional perfusion (K1, ml/min/ml), the first 4 minutes were used to perform 1-tissue compartment modeling using the descending aorta as the image-derived input function and joint estimation of bolus delay. Parametric images were generated using the same approach voxel-by-voxel but with the Akaike Information Criterion used to perform model selection.
Results: Across all five scans, the three subjects had an average total effective dose of 1.31 ± 0.28 mSv per 370 MBq injection of [11C]-butanol. The organs with the highest average absorbed doses (µGy/MBq) were the liver (12.00 ± 2.94), pancreas (11.97 ± 3.60), and heart wall (8.89 ± 2.87). ROI-based estimates of K1 for one healthy test-retest subject are shown in Figure 1A, with the corresponding parametric K1 cross-sectional image of the brain shown in Figure 1B. As shown by the K1 maximum intensity projection images (MIPs) in Figure 1C, repeat scans demonstrated similar flow estimates. However, there were some differences between datasets, particularly between the PAD subject versus healthy volunteer studies. Parametric flow rates were high (>0.75 ml/min/ml) in the kidneys, myocardium, and spleen, while K1 ranged from 0.5-1.0 ml/min/ml in the brain, and from 0.01 to 0.1 ml/min/ml in skeletal muscle in the leg, depending on the subject.
Conclusions: Quantitative dynamic total-body perfusion studies with [11C]-butanol have been performed for the first time in humans. Dose estimates are similar to existing dosimetry estimates for [11C]-butanol from Jackson et al. 2020, who reported a total effective dose of 1.6 mSv per 370 MBq dose, as well as dosimetry from other carbon-11 tracers (Zanotti-Fregonara et al. 2021). Further subject recruitment and test-retest imaging will provide an estimate of the reliability of [11C]-butanol parametric imaging using high-sensitivity high-spatial resolution total-body PET. Subsequently, analyses of the differences between the healthy and PAD cohorts will be performed to assess the clinical potential of [11C]-butanol.