Low-dose 3D PET rubidium ARMI (Alternative Radiopharmaceutical for Myocardial Imaging) multicentre trial standards and quality assurance

1805 Objectives The unstable Tc-99m supply requires alternative tracers for myocardial perfusion imaging (MPI). Rb-82 PET MPI has low radiation dose and may have superior accuracy, but requires further validation using 3D PET-CT. Rb-ARMI is a multicentre trial with an initial objective of standardizing Rb-82 PET MPI with highly repeatable interpretation in Canadian centers using 3D PET-CT technology. Methods Rest & stress phantom scans were performed at all sites to standardize image reconstruction and quantitative scoring with 4DM-PET. Patients underwent low-dose (10 MBq/kg) rest & dipyridamole stress Rb-82 MPI. Sum stress, rest (SSS, SRS) and difference scores (SDS=SSS-SRS) were visually assessed using a 17-segment model. QA cases (n=25) from all sites were co-read to assess variability of scoring and interpretation. Cases with SDS differences > 3 underwent a 3rd review to reach consensus. Results Qualifying phantom scans resulted in expected scores: SSS, SDS = 2 at all sites. Comparison of patient scores between core & recruiting sites showed very good agreement: intraclass correlation coefficient (ICC)-r = 0.91 for SSS, 0.86 for SDS. 81% of SSS & 87% of SDS scores had differences (site-core) ≤ 3. Most discrepancies occurred in large defects spanning multiple segments; however, these cases were all correctly interpreted as abnormal by recruiting and core sites. Following consensus review, scoring agreement improved to: r = 0.98 for SSS, 0.96 for SDS (p Conclusions With effective standardization and training, there was good agreement in scoring of Rb-82 MPI scans at core and recruiting sites. Standardized & repeatable interpretation is achievable across imaging centers using different 3D PET-CT scanners. Research Support Financial support from CIHR grant MIS100935