INTRODUCTION Rest-stress 82Rubidium cardiac PET/CT is used widely in USA and it is becoming increasingly available in Europe in the last few years. It not only provides information about ischaemia and LV function but also absolute myocardial blood flow (MBF) and myocardial perfusion reserve (MPR) since dynamic imaging can be performed during the first pass of the tracer. Various software are commercially available to analyse the qualitative and quantitative data. We aim to compare 3 such software packages to assess the reproducibility and variations among them. PURPOSE This study compared the agreement between A. Corridor4DM v2017 alpha (INVIA), its predecessor B. Syngo Myocardial Blood Flow package (Siemens Medical Solutions) + Corridor4DM v2015, and C. Quantitative PET (QPET, Cedars-Sinai Cardiac Suite 2015). METHODS Rest-stress 82Rb cardiac PET/CT data from 55 patients (70% male) with mean age 66 years (range 34?85), scanned at our centre between August and December 2016 were processed using 3 software with minimal manual intervention by a single observer blinded to clinical information. The MBF, MPR (corrected with resting rate pressure product), summed scores and gated volumes were assessed using each method. Bland-Altman analysis was used to determine the level of agreement with 95% limits plotted as the mean ± coefficient of repeatability. ANOVA determined whether the mean values were statistically different, considering p<0.05. RESULTS There were no significant differences between method A and B in terms of EF, global MPR and MBF (p>0.05). However method C derived global MPR and stress MBF values differed significantly from method A and B. Pairwise differences of MPR values were no more than 0.37 and 0.42mL/min/g for stress MBF and MPR. Method C derived EF values were also different compared to A and B (p<0.01) and EF mean difference was under 6.34% among the 3 methods. The mean summed scores from all methods were not different as determined by one-way ANOVA (p=0.64, p=0.95 and p=0.49, for SSS, SRS and SDS respectively). Bland-Altman analysis showed very high level of agreement (91-98%) among the 3 methods for all variables. CONCLUSION The mean values of MPR, stress MBF and EF derived from QPET differed significantly from those obtained from both versions of Corridor 4DM and Syngo MBF. Users should therefore be cautious when using different software interchangeably as systematic differences amongst them may introduce wider quantitative variation which could be clinically significant.
Autores: JOAQUIM BARRETO FONSECA ANTUNES DE OLIVEIRA , Min Sen Yew , Kshama Wechalekar