P3357Reproducibility of myocardial flow reserve estimation using Rubidium-82 PET/CT scans in healthy, young volunteers: comparison of three commercially available software packages

Abstract

Abstract Background The use of myocardial blood flow (MBF) reserve assessed with cardiac positron emission computed tomography (PET/CT) is well established and contain important prognostic information e.g. in patients with ischaemic heart disease. However, perfusion data must be interpreted under consideration of the limitations of the method. It has previously been concluded that software specific age and sex matched normal values for MBF reserve should be used as reference. Furthermore, reproducibility between scans seems to vary more in persons with higher risk of cardiovascular disease than in those with no risk factors. Comparison between scans has previously been conducted using dipyridamole as a stressor. However, no former studies have addressed this using adenosine as stressor for 82Rb-PET/CT in a combined study of scan reproducibility and software comparison. Purpose The aim of this study was to investigate reproducibility of rest- and adenosine stress 82Rb-PET/CT estimated MBF reserve in healthy young volunteers between two scans analysed with three commercially available software packages. Methods Forty healthy young volunteers completed two 82Rb-PET/CT rest- and adenosine stress scans obtaining myocardial blood flow (MBF) and MBF reserve, syngo MBF, Quantitative Gated SPECT (QGS) and Corridor4DM (4DM) were used for analyses. Results Fifty percent were men and mean (±SD) age was 24±4 years. Varying reproducibility of MBF reserve between scans was found for the different software packages. MBF reserve with syngo.MBF (Figure 1A): mean difference (95% CI) 0.26 (−0.03 to 0.54), P=0.07, limits of agreement: −1.43 to 1.95, MBF reserve with QGS (Figure 1B): mean difference (95% CI) 0.19 (−0.08 to 0.46), P=0.15, limits of agreement: −1.38 to 1.76 and MBF reserve with 4DM (Figure 1C): mean difference (95% CI) 0.08 (−0.17 to 0.34), P=0.50, limits of agreement: −1.37 to 1.53. Significantly higher MBF reserve was measured with 4DM (mean±SD: 4.00±0.96) compared to syngo.MBF (mean±SD 3.75±0.91, P<0.001) and QGS (mean±SD 3.75±0.89, P<0.0001), whereas concordance between MBF reserve measured by syngo.MBF and QGS was high (P=0.87). Figure 1 Conclusions All software packages identified the participants with flow values in the high end of the normal range, as expected for healthy volunteers. Varying reproducibility of MBF reserve was found between scans for the different software packages and 4DM seemed to have the highest inter-scan reproducibility. Concordance between MBF reserve measures was high between syngo.MBF and QGS whereas higher MBF reserve measures was conducted from 4DM in comparison with the two other software packages. Our results suggest that MBF analysed with syngo.MBF and QGS may be mutually comparable but that 4DM may be preferred for analyses due to possibly higher scan to scan reproducibility.