Focused Multi-pinhole SPECT

This thesis describes the development and validation of image acquisition methods and reconstruction techniques of focused small animal multi-pinhole Single Photon Emission Computed Tomography (SPECT). Multi-pinhole SPECT can achieve sub-half-millimeter resolution with very high sensitivity by employing collimators with multiple pinholes that are focused at the same centrally located region. The U-SPECT-II, a small-animal SPECT scanner developed at UMC Utrecht, TU Delft and MILabs, employs such multi-pinhole geometries, and is used in all studies described in this thesis. The first study demonstrates that focused multi-pinhole SPECT images can be improved when scans are better targeted to the organ or tissue of interest. New tools are presented that facilitate targeted imaging of specific organs and tumors and a study was conducted to validate the effects of improved targeting of the pinhole focus. By restricting focused SPECT scans to a small volume, count yield was increased, and visibility of small structures was significantly enhanced. At equal noise levels, small-lesion contrast measured in a mouse myocardial phantom was increased substantially, and noise in in vivo images of a tumor and the mouse heart was reduced. Ordered Subset Expectation Maximization (OSEM) can be used to accelerate iterative image reconstruction in SPECT. When subsets consist of complete projection views, as is common in clinical SPECT with parallel hole collimators, even moderate acceleration factors lead to inaccurate reconstructions. In this thesis, Pixel-based Ordered Subsets Expectation Maximization (POSEM) is introduced, which is based on an alternative subset choice. Performance was compared with traditional OSEM and Maximum Likelihood Expectation Maximization (MLEM) for a rat total body bone scan, a gated mouse myocardial perfusion scan and a Defrise phantom scan. The results show that in many cases POSEM can be operated at acceleration factors that are an order of magnitude higher than in traditional OSEM. Another study investigates the influence of unwanted high tracer uptake outside the scan volume on reconstructed tracer distributions inside the scan volume, for Tc-99m-tetrofosmin myocardial perfusion scanning in mice. Simulated projections of a digital mouse phantom in a focusing multi-pinhole SPECT system were generated. Various scan volumes were tested with increasing levels of truncation. Despite an overall negative bias, all selected volumes resulted in visually similar images. After normalization to the mean, quantitative differences were small. Absolute deviations could also be reduced by performing a short survey scan of the exterior activity and focusing the remaining scan time at the organ of interest. The last study validates the accuracy of images of intratumoral antibody distributions obtained with focused multi-pinhole SPECT. Zalutumumab, a human monoclonal EGFr-targeting antibody was radiolabeled with In-111 and administered to mice with xenografted A431 tumors. Total-body and focused tumor 3D SPECT images were acquired at 0h and 48h after injection. Global SPECT imaging showed that at 48h after injection, activity was predominantly concentrated in the tumor. Full 3D histological analysis was performed, indicating that the SPECT distribution was morphologically very similar to the ex vivo EGFr distribution and that regions showing little SPECT activity were necrotic or exhibited low EGFr expression.