Abstract
585
Objectives: Advances in SPECT brain imaging agents, that could support beta-amyloid and tau research, should be complemented by development of high-performance brain-dedicated imaging systems. We are designing a multi-pinhole brain SPECT system (AdaptiSPECT-C) capable of both static and dynamic imaging. One major concern in designing the system is sampling sufficiency of the acquired data to reconstruct 3D activity distribution in the object. We previously developed a shutter mechanism for multi-pinhole collimators for pre-clinical applications. In this work, we investigate the effect on image quality of temporal shuttering of the pinholes of the current design of the system as a mechanism for enhancing axial and angular sampling.Methods: AdaptiSPECT-C currently consists of 3 rings of detectors in a partially spherical configuration. Caudal and middle rings have 9 detectors whereas quasi-vertex ring has 5 detectors. The volume of interest for imaging is a 21-cm diameter sphere. In the original design, each detector corresponds to a single pinhole. Herein, we investigate adding 4 more apertures, 2 axially and 2 laterally, besides each main pinhole in the caudal and the middle rings for enhancing axial and angular sampling. Using temporal shuttering (opening/closing) of the main, lateral, and axial apertures, we can have multi-frame acquisition. We simulated 4 acquisition modes: mode M, single frame using only the main apertures (the original design); mode M-L, two frames using the main apertures and the lateral apertures in frames 1 and 2, respectively; mode M-A, two frames using the main apertures and the axial apertures in frames 1 and 2, respectively; and mode M-L-A, three frames using the main apertures, lateral apertures, and axial apertures in frames 1, 2, and 3, respectively. We avoid more than slight multiplexing in projections by aperture shuttering in multi-frame acquisitions. However, projections of the lateral and axial apertures are truncated due to smaller acceptance angle and finite detector size. We used a customized Defrise phantom and an XCAT brain perfusion phantom as activity sources for simulations. We reconstructed the images using an MLEM algorithm.Results: While the original design of the system can’t recover all the hot disks in the Defrise phantom, most of them can be recovered using the additional apertures. Also, the visual quality of the images of the brain perfusion phantom was improved using the proposed technique. Table 1 shows the quantitative assessment of relative image quality in terms of normalized mean square error (NMSE).
Conclusions: Our simulations indicated that for the AdaptiSPECT-C system, additional apertures in a multi-frame data acquisition can improve image quality by increasing the number of axial and angular samplings while avoiding degradation from multiplexing by using temporal shuttering of apertures.Research Support: National Institute of Biomedical Imaging and Bioengineering (NIBIB), National Institutes of Health, Grant No R01 EB022521.