Abstract
1159
Objectives Gated blood pool SPECT radionuclide angiography (GSPECT RNA) is interesting for the evaluation of cardiac function. We previously validated the use of QBS software (Cedars Sinai) for the quantification of left ventricular (LV) and right ventricular (RV) function. This was done with filtered backprojection reconstruction (FBP). We aimed to study the performance of QBS for the quantification LV and RV function when using FBP as compared to iterative reconstruction with resolution recovery (3D-Flash, Siemens).
Methods Our study included 37 patient addressed for LV and RV function evaluation with planar (planarLAO) and GSPECT RNA. Studies were acquired on a two-headed gamma-camera (Symbia, Siemens). PlanarLAO were processed with the NXT program (Vision, GEMS) and provided planarLAO LVEF and RVEF. GSPECT RNA were reconstructed using FBP and 3D-Flash and then processed with QBS. Results provided with the maximal activity threshold method (MAT) of QBS were noted: LV and RV end diastolic volumes, end systolic volumes and stroke volumes (SV), as well as LVEF and RVEF. For comparison of the performance of FBP versus 3D-Flash, planarLAO LVEF and RVEF were considered gold standard. And for RV and LV volumes, we hypothesized that the best reconstruction method would be the one providing the highest correlation between RV-SV and LV-SV.
Results LVEF provided by planarLAO (61±9%) is highly correlated to LVEF measured with QBS-FBP (72±17%, r=0.81; P <0.0001) and QBS-3D-Flash (72±16, r=0.83; P <0.0001). RVEF provided by planarLAO (46±7%) is moderately correlated to RVEF measured with QBS-3D-Flash (48±7%, r=0.27; P <0.0001) but not to QBS-FBP (45±9%, NS). Linear correlation between LV-SV and RV-SV measured is higher with QBS-3D-Flash (81±22 ml and 97±27 ml, r=0.74; P <0.0001) than QBS-FBP (85±28 ml and 123±37 ml, r=0.38; P <0.0001).
Conclusions QBS performances seem better when using 3D-Flash than FBP for both LV and RV EF and volume measurements