Sub-millimeter nuclear medical imaging with high sensitivity in positron emission tomography using beta-gamma coincidences

TL;DR

This paper introduces a novel gamma-PET imaging technique that achieves sub-millimeter spatial resolution using triple-gamma coincidences, significantly reducing the number of events needed and enabling faster, potentially real-time, nuclear medical imaging.

Contribution

The paper presents a new gamma-PET method utilizing triple-gamma coincidences for high-resolution imaging with fewer events, improving speed and sensitivity over conventional PET.

Findings

Achieves ca. 0.4 mm spatial resolution in simulations.

Requires only about 40 reconstructed intersections for high-resolution imaging.

Reduces acquisition time to approximately 140 seconds for effective imaging.

Abstract

We present a nuclear medical imaging technique, employing triple-gamma trajectory intersections from beta^+ - gamma coincidences, able to reach sub-millimeter spatial resolution in 3 dimensions with a reduced requirement of reconstructed intersections per voxel compared to a conventional PET reconstruction analysis. This '$\gamma$-PET' technique draws on specific beta^+ - decaying isotopes, simultaneously emitting an additional photon. Exploiting the triple coincidence between the positron annihilation and the third photon, it is possible to separate the reconstructed 'true' events from background. In order to characterize this technique, Monte-Carlo simulations and image reconstructions have been performed. The achievable spatial resolution has been found to reach ca. 0.4 mm (FWHM) in each direction for the visualization of a 22Na point source. Only 40 intersections are sufficient for a reliable sub-millimeter image reconstruction of a point source embedded in a scattering volume of water inside a voxel volume of about 1 mm^3 ('high-resolution mode'). Moreover, starting with an injected activity of 400 MBq for ^76Br, the same number of only about 40 reconstructed intersections are needed in case of a larger voxel volume of 2 x 2 x 3~mm^3 ('high-sensitivity mode'). Requiring such a low number of reconstructed events significantly reduces the required acquisition time for image reconstruction (in the above case to about 140 s) and thus may open up the perspective for a quasi real-time imaging.