Abstract
In this work, the quantitative aspect of myocardial imaging is of main concern. The following quality control measurements have been assessed: planar and SPECT resolution and sensitivity, stability and response during rotation, mechanical alignment and rotational linearity, SPECT response to a uniform source and resolution. Statistical error, i.e. noise, as an image quality indicator was also measured. The assessment of propagation of noise in the reconstructed images was compared with the Budinger equation. In addition, the target-to-non-target ratios between myocardial agents were calculated from profiles, in terms of the maximum peak to the minimum valley. Both 99Tcm-labelled myocardial perfusion agents, tetrofosmin and methoxy isobutyl isonitrile (MIBI), have similar myocardial target-to-non-target ratio and both are superior to 201Tl-chloride. The system finite resolution at different depths and scattering and attenuation effects on the quantitation accuracy of myocardial perfusion was addressed. Compensation for these factors was proposed and assessment of the dual-energy window for scatter correction was performed. The proposed modified second derivative method shows a promising correction effect. Unlike non-gated techniques, the multi--gated Single Photon Emission Computed Tomography (MG SPECT) technique provides diagnostic images less affected by the heart motion artifact. The size of the defects are more clearly displayed and more easily identified as there is less superimposition of the systolic and diastolic slices of the raw data. The end systolic (ES) and end diastolic (ED) slices being separately acquired, ventricular ejection fraction can also be calculated especially when edge enhancement is applied for easy edge delineation and volume determination. A new method is proposed to measure the myocardial wall thickness. A horizontal profile passing through the middle of the short axis slice was generated and the FWHM was taken to be the wall thickness of each wall. The second derivative of each profile was used to correct for the system effect and compared with the count-based method. Both methods have shown a good agreement. As a result, wall thickness measurement is possible and therefore an accurate quantitative analysis of the thickening is also possible in order to predict myocardial viability.