Positioning of Multiple Photon Interaction Events in Cross Strip Cadmium Zinc Telluride Detectors for PET

TL;DR

This paper presents an algorithm for accurately locating photon interactions in CZT detectors for PET, improving position resolution by resolving event degeneracy using trigger timing and pulse height ratios.

Contribution

The study introduces a novel algorithm that resolves photon event positioning degeneracy in CZT PET detectors, enhancing spatial accuracy beyond previous methods.

Findings

Algorithm achieves 59.7% to 66.7% accuracy in identifying true interaction locations.

Improves position resolution by 19.4% to 33.4% over unbiased random guessing.

Potential to increase system sensitivity by better utilizing ambiguous events.

Abstract

This report summarizes work done during PhD rotation that proposes an algorithm that uses anode- cathode trigger time difference and cathode to anode pulse height ratio to resolve photon event positioning degeneracy in a novel Cadmium Zinc Telluride (CZT) photon detector for positron emission tomography (PET). The detectors used comprise 40 mm x 40 mm x 5 mm monolithic CZT crystals that employ a cross-strip electrode pattern with interspersed steering electrodes to obtain high spatial and energy resolution. The specific positioning degeneracy studied results when a photon event interacts in two detector pixels, triggering two anode and two cathode strips, hence presenting four pixels as potential interaction locations. An experiment was performed where a 1 mm diameter collimated pencil beam of annihilation photons was used to target a detector pixel of known location. The data acquired provided an empirical basis to train and test the algorithm. The study shows that, depending on the restrictions applied on the data set, the proposed algorithm accurately identifies the true photon interaction locations 59.7% to 66.7% of the time, which is 19.4% to 33.4% better than an unbiased Bernoulli trial. Further improvements of the algorithm will enhance the effective system photon sensitivity by extracting useful information from events which would otherwise be discarded.