Optimization of Y90 bremsstrahlung image reconstruction using multiple energy window subsets

Abstract

1680

Objectives: Yttrium-90 (Y90) is a promising isotope for radionuclide therapy. Its beta- emission allows for an optimized dose deposition but its lack of gamma emission is complicating the task of imaging for a personalized dosimetry. Bremsstrahlung of the emitted electrons generates however a continuous spectrum of gamma rays that can be used for imaging. A previous study, focused on the history of the photons arriving on the detector, showed the importance of using a low and large energy window (50-150keV) for maximizing the amount of geometric photons while maintaining a reasonable sensitivity. The work presented here investigates the use of multiple energy windows for the reconstruction of Y90 images.

Methods: The MLEM algorithm was used to reconstruct images of a Jaszczak phantom. The 50-150 keV energy window was divided into 3 subsets (50-65 keV, 65-95 keV, 95-150 keV) used alternatively for the reconstruction of the images. For each subset, a Monte Carlo modeling of the detector response was incorporated into the MLEM algorithm in both forward and back projections. Attenuation maps, based on the mean energy of the subset, were used for attenuation correction. Results for contrast recovery and noise were compared to previous results obtained with a single energy window.

Results: When correcting for the attenuation and for an equal level of noise, results show an improvement in contrast recovery for the images reconstructed using multiple energy window subsets. This can be explained by the better approximation of the detector response and a better estimation of the attenuation of the photons when multiple energy windows are considered.

Conclusions: It is shown in this study that using multiple energy subsets in the reconstruction process can enhance image quality for bremsstrahlung imaging.