The aim of this study was to validate the estimation of left ventricular end-diastolic and end-systolic volumes (EDV, ESV) and ejection fraction (LVEF) as well as wall motion analysis from gated fluorine-18 fluorodeoxyglucose (FDG) positron emission tomography (PET) in patients with severe coronary artery disease (CAD) using software originally designed for gated single-photon emission tomography (SPET). Thirty patients with severe CAD referred for myocardial viability diagnostics were investigated using a standard FDG PET protocol enhanced with gated acquisition (8 gates per cardiac cycle). EDV, ESV and LVEF were calculated using standard software designed for gated SPET (QGS). Wall motion was analysed using a visual four-point wall motion score on a 17-segment model. As a reference, all patients were also examined within a median of 3 days with cardiovascular cine magnetic resonance imaging (cMRI) (20 gates per cardiac cycle). Furthermore, all gated FDG PET data sets were reoriented in a second run with deliberately misaligned axes to test the quantification procedure for robustness. Correlation between the results of gated FDG PET and cMRI was very high for EDV and ESV ( R=0.96 and R=0.97) and for LVEF ( R=0.95). With gated FDG PET, there was a non-significant tendency to underestimate EDV (174+/-61 ml vs 179+/-59 ml, P=0.21) and to overestimate ESV (124+/-58 ml vs 122+/-60 ml, P=0.65), resulting in underestimated LVEF values (31.5%+/-9.4% vs 34.2%+/-12.4%, P<0.003). The results of reorientations 1 and 2 showed very high correlations (for all R>/=0.99). Segmental wall motion analysis revealed good agreement between gated FDG PET data and cMRI (kappa =0.62+/-0.03). In conclusion, despite small systematic differences which contributed mainly to the lower temporal resolution of gated FDG PET, agreement between gated FDG PET and cMRI was good across a wide range of volumes and LVEF values as well as for wall motion analysis. Therefore, gated FDG PET provides clinically relevant information on function and volumes, using the commercially available software package QGS.