Fifty-seventh annual meeting of the American association of physicists in medicine
TH-AB-204-07: Low Dose, Single-Step, X-Ray Phase Contrast Imaging with a Spectral Detection Scheme
Widely popular x-ray interferometric x-ray phase contrast methods are known to introduce high radiation dose and require multiple measurement steps per projection to extract absorption and phase signatures from intensity measurements. A newer method like coded-aperture technique mitigates many drawbacks of an interferometric method but still requires multiple measures involving movement of the aperture masks to obtain separate absorption and phase information. This abstract explores a novel method to enable a single-step x-ray phase contrast imaging using spectral detection scheme in a coded-aperture setting.
The proposed setup consists of a coded-aperture phase contrast imaging setup with a spectral detector. Spectral detection scheme allows detection of multiple energy-bin data simultaneously. We derived a new transport of intensity equation (TIE) that describes x-ray propagation and detection in such a setup. Using this new TIE equation in conjunction with a new retrieval algorithm we obtain a solution method for quantitatively accurate estimation of absorption, phase and differential phase contrast. A rigorous forward simulation was used to examine the retrieval accuracy. The phantom consisted of 5cm thick (50–50 density) breast phantom with thin tubes of PMMA, calcification spheres and cancerous tissue. An imaging dose of 1.5 mGy was used in the simulation.
Our results show excellent quantitative agreement for retrieved results in absorption, phase and differential phase contrast in comparison to the ground truth. Our method is the first single-step technique to achieve a low dose retrieval of multiple tissue properties.
Our proposed method based on a non-interferometric method for spectral solutions to absorption, phase and DP has potential to mitigate existing hurdles in phase contrast imaging. Further investigations are underway in terms of bench top system development and understanding optimal choice of energy windows, detector pixel size and sensor thickness for the described method.