Development and Full Validation of a Monte Carlo Single-photon Emission Computed Tomography Camera Proposed for Coded Aperture Breast Tumor Imaging

Abstract

The Modified Uniformly Redundant Array (MURA) Coded Aperture (CA) imaging technique,
initially developed for astronomical X-rays in far-field geometry, offers high Signal-to-Noise Ratio (SNR) and
artifact-free imaging. Interest has grown in adapting MURA-CA to near-field geometry due to its high sensitivity
and superior resolution, despite challenges related to artifacts. Materials & Methods: This study utilized MURA
masks, anti-masks, and subtraction techniques to exploit their delta correlation function. Simulations were
conducted with varying planar source sizes using 140 keV photons and Gaussian blurring to replicate real-world
conditions. Separate decoding functions were applied to reconstruct images from mask- and anti-mask-projected
data. Metrics including resolution Full Width at Half Maximum (FWHM), contrast, SNR, root-mean-square error
(RMSE), and correlation coefficients were evaluated. Results: MURA-CA exhibited significant SNR advantages
compared to pinholes and collimators. Near-field imaging, however, introduced artifacts that could be mitigated
by transitioning to far-field geometry, though at the expense of counting efficiency and SNR. Subtraction of mask
and anti-mask images effectively minimized side lobe effects and reduced RMSE, achieving results comparable to
individual mask and anti-mask projections. Further enhancements were made by optimizing the trade-off between
SNR and resolution under fixed time and dose conditions. Conclusion: MURA-CA is particularly well-suited for
far-field applications like breast tumor imaging, where artifacts are minimal. This study highlights the potential of
a MURA-CA system specifically designed for high-resolution SPECT imaging, integrated with a clinical gamma
camera, to enable early tumor detection with improved imaging performance.