Robust Detector Calibration for a Novel PET System Based on Cross-Strip CZT Detectors

Abstract

In positron emission tomography (PET) imaging, high energy and spatial resolution are necessary to obtain high quality images. This paper presents a photon interaction position dependent energy calibration for a small animal PET system built with cadmium zinc telluride (CZT) crystals. 3-D-resolving CZT detection systems employ up to 103–105 read-out channels which require individual calibration due to position-varying crystal quality. This paper aims to establish an automated data processing algorithm to minimize manual intervention and guarantee consistent calibration. The algorithm analyzes the data using a two-step process with a calibration and validation step. The validation step is a spectrum analysis which utilizes a non-Gaussian shape function. The calibration step is a position dependent photopeak recovery where three different methods were studied. The analysis of the uncorrected data reveals an energy resolution of (8.4%±0.2%) FWHM at 511 keV. The best performance is achieved with the proposed processing, which yields an energy resolution of (7.0%±0.1%) FWHM at 511 keV. The robustness of the shape function used to determine the energy resolution was evaluated and was found to be superior compared to a pure Gaussian shape function. In summary, the proposed data algorithm can generate calibration parameters in a robust manner and provides a consistent calibration.