Image reconstruction from tissue scattered events for ${\beta}^{+}\gamma$ coincidences in Compton-PET

TL;DR

This paper explores the feasibility of reconstructing images using only single scattered coincidences in a Compton-PET system, potentially improving imaging of non-pure beta emitters with fewer counts.

Contribution

It introduces a novel method for tissue scatter point localization and annihilation point estimation using Compton cones and time-of-flight data in Compton-PET imaging.

Findings

Feasibility demonstrated for imaging with single scattered events.

Reduction of tissue scatter points to two locations using Compton cones.

Enhanced confidence in image interpretation from single scattered data.

Abstract

For long time non-pure beta emitters are avoided from PET imaging due to extra dose and increase in background from Compton scattering. But advent of high-resolution Compton camera system opens up new domain of imaging. Various non-pure beta emitters are formed as beam irradiation byproduct in therapy which can be used in online beam range verification. In this case, the number of usable counts for imaging is generally 1-3 order lesser than normal PET scan. On the other hand, we know that in human PET scanner, 30-60\% can be tissue scattered coincidences in 3D case containing 80\% single scattered events. In this work, we have investigated feasibility of imaging using only single scattered coincidences for non-pure beta emitters in a Compton-PET system. The locus of tissue scatter point can be reduced to in generally two points after using Compton cone from both ends of 511 keV detections. Finally, annihilation point is estimated using Compton cone of 1157 keV gamma and time-of-flight information for the 511 keV. We believe independent assessment of underlying activity from single scattered data sets will increase confidence in image interpretation.