Simultaneous dual isotope imaging, absolute tracer quantification and image quality analysis for myocardial perfusion imaging

This thesis aims at improving the diagnosis of coronary artery disease using Single Photon Emission Computed Tomography (SPECT) imaging. Substantial contributions of this work are the introduction and evaluation of a new clinical protocol aiming at a clinical work-flow improvement, a more accurate method for absolute quantification of myocardial tracer uptake, and an improved procedure for image quality assessment. In the following, these three subtopics and related contributions of this thesis are described. In contrast to the established, serial clinical procedure, the developed clinical protocol is based on the injection of the two radioactive tracers, ^99mTc-sestamibi and ²⁰¹Tl-chloride. Subsequently, the perfusion of the myocardium under stress and rest conditions is determined simultaneously via a single SPECT acquisition. Thereby the overall examination time can be reduced from 4-5 hours to about 1 hour. In addition the currently performed and straining second SPECT acquisition can be omitted. An accurate scatter and attenuation correction during reconstruction is essential for the analysis of simultaneously acquired distributions of both tracers, since scattered high-energy ^99mTc photons contaminate the lower energy ²⁰¹Tl emissions. Besides the development of the simultaneous dual isotope protocol, the technical feasibility is proven by phantom studies and the clinical applicability is confirmed by a clinical study. In 31 out of 51 cases SDI and standard imaging used as reference led to consistent results. In 15 patients clinical follow examinations showed that the SDI images depict the clinical situation more accurately. For the remaining five cases the differing results of SDI imaging could not be confirmed (3) or no further, valuable clinical follow-up was available (2). Absolute quantification of tracer uptake promises an improved diagnosis compared to current qualitative procedures in particular in case of multi-vessel-disease. In this thesis a quantitative image analysis is developed and evaluated by phantom and patient data. Hereby an accurate method for scatter and attenuation correction is also crucial for determining the spatial tracer distribution. The special properties of the used reconstruction algorithm optimized for this task are examined in detail in the third section of the thesis. Besides the aforementioned requirements regarding the reconstruction, the influence of residual activity after injection is examined and a model-based segmentation of the heart is adapted to SPECT imaging. Furthermore, a newly developed method for calibration from reconstructed counts to absolute activity concentrations including a suitable phantom is presented. Hereby, it was possible to prove that even mis-registrations on the sub-voxel scale can have a significant influence on the calibration which led to a significant advancement in form of an automated correction method. The subsequently derived results show that an accuracy of 7.5% can be achieved in phantom studies. Likewise the results of the clinical study are in good accordance to the expected values. For patients with a known coronary artery disease a significant reduction of the ratio of uptake at stress and at rest can be confirmed. However, a concrete diagnostic added value by quantification cannot be derived for the examined patient collective. In the third section of the thesis an image quality assessment procedure based on signal and noise power spectra for SPECT and PET (Positron Emission Tomography) images is developed to evaluate the influence of different imaging and reconstruction parameters (phantom size, number of iterations, resolution recovery and scatter correction). The method is applied to simulated and measured data and compared to the established method of covariance matrix analysis. As a result, a good accordance between both methods is found and also between simulated and measured data. By this method the great impact of scatter correction on the reconstructed contrast could be proven and an analytical explanation for the image quality improvement by Time of Flight (ToF) could be provided.