High-resolution EM-CCD scintillation gamma cameras

Abstract

The development of medical imaging techniques has dramatically changed clinical practice and biomedical science in the 20th century. Nuclear Medicine imaging techniques reveal the function of organs and tissues in vivo with the aid of radioactively labeled tracer molecules. These techniques, such as Single Photon Emission Computed Tomography (SPECT) and Positron Emission Tomography (PET), are clinically applied in the fields of neurology, oncology, cardiology, skeletal and thyroid imaging. Applied to small animals, these techniques are important tools in development of better diagnostics and new drugs and fundamental biological research. For small animal SPECT systems, pinhole magnification enables image resolutions below 0.3 mm using conventional Anger cameras with an intrinsic spatial resolution of only 3 to 4 mm. There is a large desire to improve the spatial resolution of the gamma detectors to further improve the preclinical and clinical SPECT performance as well as to make systems more compact. This thesis focuses on EM-CCD based scintillation gamma cameras with a very high spatial resolution, more than 10 times better than conventional Anger cameras. Therefore these EM-CCD gamma cameras are a candidate for the next generation of pre-clinical-SPECT scanners. In pinhole SPECT cameras the spatial resolution is severely degraded for gamma photons incident at an angle, called the depth-of-interaction effect. We showed that this effect can be overcome using a novel curved scintillator geometry which results in a scintillation light centroid independent of the depth-of-interaction. Alternatively we strongly reduced the depth-of-interaction effect by using a multi scale algorithm that detects the scintillation depth. This depth is determined by modelling the variation of the scintillation light spread with depth. If we also take the statistics of the detector into account using a Maximum Likelihood scintillation detection algorithm the energy resolution of the detector improves significantly. Furthermore, using the Cramer Rao lower bound we showed that the EM-CCD based gamma camera performance is very sensitive to the scintillator light yield and the Clock Induced Charge (CIC) and dark current noise. A feasible reduction of these noise sources would improve the gamma camera performance significantly.