Submillimeter nuclear medical imaging with a Compton Camera using triple coincidences of collinear \beta+ annihilation photons and \gamma-rays

TL;DR

This paper introduces a gamma-PET imaging method using triple coincidences of collinear eta+ annihilation photons and gamma-rays, achieving submillimeter resolution with significantly lower activity than traditional PET.

Contribution

It presents a novel gamma-PET technique that leverages triple coincidence detection to improve spatial resolution and reduce required activity in nuclear medical imaging.

Findings

Achieves 0.2 mm spatial resolution (FWHM), surpassing conventional PET.

Demonstrates feasibility with low activity sources (~0.7 MBq).

Provides a coincidence detection efficiency of 1.92e-7 per decay.

Abstract

Modern PET systems reach a spatial resolution of 3-10 mm. A disadvantage of this technique is the diffusion of the positron before its decay with a typical range of ca. 1 mm (depending on its energy). This motion and Compton scattering of the 511 keV photons within the patient limit the performance of PET. We present a nuclear medical imaging technique, able to reach submillimeter spatial resolution in 3 dimensions with a reduced activity application compared to conventional PET. This 'gamma-PET' technique draws on specific positron sources simultaneously emitting an additional photon with the \beta+ decay. 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 test the feasibility of this technique, Monte-Carlo simulations and image reconstruction has been performed. The spatial resolution amounts to 0.2 mm (FWHM) in each direction, surpassing the performance of conventional PET by about an order of magnitude. The simulated detector geometry exhibits a coincidence detection efficiency of 1.92e-7 per decay. Starting with only 0.7 MBq of source activity (ca. 200-500 times less compared to conventional PET) an exposure time of 450 s is sufficient for source reconstruction.