Brady John McKee MD, Presenter: Nothing to Disclose

Shanker Raja MD, Abstract Co-Author: Nothing to Disclose

Lamia Jamjoom MD, Abstract Co-Author: Nothing to Disclose

Anne Morris MD, Abstract Co-Author: Nothing to Disclose

Sebastian Flacke MD, Abstract Co-Author: Nothing to Disclose

To assess the correlation and agreement of Emory Cardiac Toolbox (ECTb) and 4D-MSPECT (4DM) software packages in the automatic quantification of left ventricular ejection fraction (EF), end-diastolic volume (EDV), myocardial mass (MM), and transient ischemic dilatation (TID) from rest and stress Rubidium-82 PET/CT myocardial perfusion imaging studies (Rb82MPI) .

46 patients (29 males) aged 40-90 years (mean 68y) with known or suspected coronary artery disease underwent consecutive Rb82MPI at both rest and stress on a 64 slice GE Discovery PET/CT scanner and were included in this study. Following low dose CT for attenuation correction, Rb82MPI images of the heart obtained at rest and at stress during adenosine infusion were retrospectively processed with ECTb and 4DM software by a single observer for automatic quantification of EF, EDV, and MM (92 observations) and TID (46 observations).  Correlation and agreement of the respective estimates were evaluated using regression analysis and Bland-Altman plots.

The correlation coefficient between methods was high for EF (.93), EDV (.99), and MM(.98) but low for TID (0.62).  Bland Altman analysis (Mean +/- SD) yielded relatively positive bias for ECTb vs 4DM in quantification of EF (3.55% +/- 5.8%) and TID (0.4 +/-.09) and relatively negative bias in quantification of EDV (-8.5ml +/- 9.3ml) and MM (-19.2g +/- 7.6g).  There was a slightly increased negative bias of ECTb vs 4DM for large values of MM.

While correlation between ECTb and 4DM automatic quantification of EF, EDV, and MM for Rb82MPI was high, the limits of agreement appear sufficiently wide to preclude the interchangeable use of these software packages for clinical purpose despite accounting for calculated bias.  Both correlation and agreement was poor for TID suggesting the need for further validity testing of these algorithms to quantifying this parameter in Rb82MPI.

In Rb82MPI as previously noted in gated SPECT, poor agreement between ECTb and 4DM estimates of EF, EDV, and MM precludes interchangeable use.  Poor TID correlation warrants further validity testing.

McKee, B, Raja, S, Jamjoom, L, Morris, A, Flacke, S, Agreement of Emory Cardiac Toolbox and 4D-MSPECT for Quantification of Left Ventricular Function in Rubidium-82 PET/CT Myocardial Perfusion Imaging.  Radiological Society of North America 2008 Scientific Assembly and Annual Meeting, February 18 - February 20, 2008 ,Chicago IL. http://archive.rsna.org/2008/6005968.html