SPECT Simulation of a Digital Anthropomorphic Brain Phantom

A digital brain phantom was created from 3D CAD images and imported to a simulation of a Siemens Healthineers Symbia Evo SPECT system. The simulation was performed using GAPE, a Geant-4 based Monte Carlo toolkit, which included features to import 3D CAD structures. The brain object was generated using an MRI study of a human brain and imported to GATE as a tessellated mesh filled with radiation infused water that had a radiation spectrum matching Tc-99m and spatial distribution that filled the brain volume. The simulation was configured for a dual head SPECT with a LEHR parallel collimator and NaI scintillation detectors on a brain phantom filled with water. The resulting projection matrix was reconstructed using proprietary software to create an image that was validated against the simulation ground truth. The known activity distribution. The processes were parallelized across a local cluster that generated data with the SPECT heads at predefined angles for each slice, which were then recombined to create the full 180 degrees step-and-shoot acquisition. The simulation methodology was tested using a simulated point source and a hot-cold spheres phantom. Possible future applications include utilizing the method to generate training data for deep learning and the extension to other organs.