An optimized configuration of multi-pinhole aperture can improve the spatial resolution and the sensitivity of pinhole SPECT simultaneously. In this study, an optimization strategy of the multi-pinhole configuration with a small detector is proposed for mouse cardiac imaging. A 14 mm-diameter spherical field-of-view (FOV) is used to accommodate the mouse heart. To accelerate the optimization process, the analytic models are applied to rapidly obtain the projection areas of the FOV, the sensitivities and the spatial resolutions of numerous system designs. The candidates of optimal multi-pinhole configuration are then decided by the preliminary evaluations with the analytic models. Subsequently, the pinhole SPECT systems equipped with the designed multi-pinhole apertures are modeled in GATE to generate the imaging system matrices (H matrices) for the system performance assessments. The area under the ROC curves (AUC) of the designed systems is evaluated by signal-known-exactly/background-known-statistically detection tasks with their corresponding H matrices. In addition, the spatial resolutions are estimated by the Fourier crosstalk approach, and the sensitivities are calculated with the H matrices of designed systems, respectively. Furthermore, a series of OSEM reconstruction images of synthetic phantoms, including the hot-rod phantom, mouse heart phantom and Defrise phantom, are reconstructed with the H matrices of designed systems. To quantify the sensitivity and resolution competition in the optimization process, the AUC from the detection tasks and the resolution estimated by the Fourier crosstalk are used as the figure of merits. A trade-off function of AUC and resolution is introduced to find the optimal multi-pinhole configuration. According to the examining results, a 22.5° rotated detector plus a 4-pinhole aperture with 22.5° rotation, 20% multiplexing and 1.52X magnification is the optimized multi-pinhole configuration for the micro pinhole-SPECT applied to mouse cardiac imaging with a camera of 49 × 49 mm2 active area.
- image quality
- system performance