PT  - JOURNAL ARTICLE
AU  - Kimberly Widrick
AU  - Xin Li
AU  - Michael King
AU  - Phillip Kuo
AU  - Lars Furenlid
TI  - Photon-transport Simulations of a Curved Gamma-Ray Detector for use in a dedicated Brain SPECT System
DP  - 2019 May 01
TA  - Journal of Nuclear Medicine
PG  - 1393--1393
VI  - 60
IP  - supplement 1
4099  - http://jnm.snmjournals.org/content/60/supplement_1/1393.short
4100  - http://jnm.snmjournals.org/content/60/supplement_1/1393.full
SO  - J Nucl Med2019 May 01; 60
AB  - 1393Objectives: Our aim is to develop a detector with high intrinsic spatial resolution for use in a clinical brain SPECT system. Traditional brain SPECT is performed with a 2-or 3-headed rotating gantry system, whereas our system proposes a stationary hemisphere of modular detectors, each with their own pinhole collimator. To obtain increased resolution and sensitivity we propose a detector with a cylindrically-curved scintillation crystal read out with a matching fiber-optics transfer plate. Methods: The prototype was simulated as a cylindrically-curved (R=305mm) 157mm x 157mm detector with a NaI(Tl) scintillation crystal that was 6-8 mm thick with a curvature-matched light guide placed behind it with a thickness from 1-3mm. After the light guide, a curved 1-inch-thick fiber-optic plate with a planar exit face was used to transfer light to a planar light guide. This light guide transfers the scintillation light to an array of SiPMs. To determine the best spatial resolution attainable from this design we used a scintillation-photon transport code written in C++ to generate the mean detector response function (MDRF), and then used the MDRF to calculate the Fisher Information Matrix and resolution as a function of light guide and scintillation crystal thicknesses. The code generates optical photons at a gamma-ray interaction location, each with a randomized direction vector, and then propagates the photon through the materials of the detector including reflective and refractive interactions at interfaces. The 3D MDRF was created by initiating the photons on a grid of 1mm x 1mm interaction locations at a uniform depth within the crystal, and then repeating the simulations at a grid of depths. The SiPMs were simulated as a 6x6 array of 1-inch x 1-inch active surfaces that model commercially available units. This data was then used to determine the Fisher Information and intrinsic resolution via the Cramér-Rao Lower Bound using an analysis code written in MATLAB. Results: We found that the curvature of the detector improved resolution along the direction of curvature. Among the initial set of candidate designs, the best average spatial resolution was found using a 3mm light guide between the crystal and the fiber optic plate. Despite a slight decrease in resolution, we found that an 8mm thick crystal was best for our application due to the increased stopping power. Conclusion: Using a curved detector with fiber optics in a SPECT system is an effective way to increase resolution over a planar detector. In addition, this simulation methodology could be used for more camera geometries to provide a more economic way of determining the best design for cameras in the future. Acknowledgement: This work was partially supported by NIH/NIBIB grant R01-EB022521.