Abstract
Background: LVEF is a reliable diagnostic and prognostic value in coronary artery disease. Several modalities including ECHO, GSPECT, cMRI and RNV can assess the LVEF. Filtering contributes to the variation of LVEF with influence on spatial resolution and statistical noise in the reconstructed images. The authors investigated the impact of Wiener and Butterworth filters variation for accuracy of LVEFs in CAD pts.
Methods: Scans of 21 patients; 14 males and 7 females, 25-76 years old, with suspected CAD who underwent the 99mTc Myoview GSPECT using QPS and QGS were reconstructed with Winer filter based on the outcome of an earlier SPECT-CT optimisation of two chest phantoms. The patient results compared to the previous reports using a standard Butterworth filter and the ECHO for comparison.
Results: 47%, 29% and 24% of pts reported normal, equivocal (EV) and having CAD based on LVEF respectively and prior to modified filtering. A good correlation between ECHO and the calculated LVEF with Winer and Butterworth (r2= 0.876 & r2=0.847) and for small heart (r2= 0.951 & r2=0.933) was observed. To achieve less variation in LVEF measurement, we established a mean cutoff value for Butterworth (0.26 +/-0.09) and a mean point value for Winer (4.9 +/-1.19) in correlation to ECHO. Patients with LVEF of <40 showed greater variation in points- setting for Winer, but less fluctuation in cutoff for Butterworth. No remarkable difference noticed for LVEF between two filter techniques for normal cases after adjusting parameters.
Conclusions: There was a good agreement between LVEF derived from GSPECT and ECHO after adjusting parameters with 10% higher accuracy in equivocal cases at this stage. Mean percentage LVEF error for Wiener and Butterworth against ECHO LVEF were 2% and 6% respectively post correction, together with more elevated qualitative acceptable images. The accuracy of intermediate CAD reporting for both rest and stress studies were elevated, though more patients’ images are needed to be assessed for equivocal and smaller size hearts to reach a more statically viable result.