Investigation Of Sub-500μm PET Image Based On350μm Pitch Pixelated CdZnTe Detectors

We are currently investigating the feasibility of using highly pixelated Cadmium Zinc Telluride (CdZnTe) detector for sub-500μm resolution PET imaging applications.20mm×20mm×5mm CdZnTe substrate was fabricated with350μm pitch pixels (250μm anode pixels with100μm gap) and coplanar cathode. The same CdZnTe detector was polished and deposited again with600μm pitch pixels (500μm anode pixels with100μm gap) in order to compare the results of these two different pitch detectors.In addition to non-collimated y-ray source measurements (59.5keV,122keV and511keV), we scan across pixels with a collimated Co-57source (collimated beam size<pixel size) with a step size50μm. We sorted events based on the number of pixels that detect a charge signal (single, double) and analyzed the charge sharing at different energies for detectors of different pitches. Results show the number of charge sharing events increases when y-ray energy increases or when pixel size decreases. For both350μm pitch and600μm pitch CdZnTe detectors, the scatter plots of signal amplitudes of two neighboring pixels clearly show more charge sharing when the collimated beam hits the gap between adjacent pixels. And, the energy resolution of600μm pitch detector is better than the energy resolution of350μm pitch detector.Using collimated122keV and511keV beams, we measured the intrinsic spatial resolution of350μm pitch detector biased at-1000V. For122keV sources, the measured single-pixel event profile shows a resolution of340μm FWHM; while the double-pixel event profile shows360μm FWHM. The effective collimator size is270μm×270μm. For511keV source, the single-pixel event profile shows a resolution of410μm FWHM; and the double-pixel charge sharing event profile shows520μm FWHM. Depth of interaction was analyzed based on two types of methods, i.e. cathode/anode ratio and electron drift time. The method of cathode/anode ratio shows better depth information for350μm pitch detector in both122keV and511keV measurements. For single-pixel photopeak events, a linear correlation between cathode/anode ratio and electron drift time was shown, which was proposed to correct the timing resolution of CdZnTe detector in the coincidence measurement between CdZnTe and LSO detectors.A coincidence detection system between350μm pitch pixelated CdZnTe detector and a LSO-based Inveon PET detector was built up for PET imaging applications. Using different trigger modes, i.e. triggering time comes from LSO signal, triggering time comes from CdZnTe anode signal, and triggering time comes from CdZnTe cathode signal, we studied how the triggering strategy affects the PET imaging resolution. Results show the best image resolution comes from the LSO trigger mode. We recorded coincidence events between a central CdZnTe pixel and the LSO detector to image a NEMA NU4Na-22point source, which is250μm in diameter and embedded in an acrylic cube.In a VP-PET geometry, where the distance between Na-22source and CdZnTe detector is23mm, and the distance between the source and LSO detector is127mm, we obtained670μm at FWHM image resolution for coincidence events detected by CdZnTe central pixel only using2D FBP algorithm. The point source was moved1.2mm away and imaged again, the fused image show the two point sources can be clearly separated. If we include coincidence events from CdZnTe double-pixel charge sharing events (while the CdZnTe central pixel still got the biggest signal amplitude), the PET image resolution degrades to730μm FWHM, but the count rate increases2.5to3times. We measured the time difference between CdZnTe anode signal and LSO signal based on the coincidence setup. Combined with the electron drift time measurement from one flash ADC system and the Cathode/Anode ratio measured from the coincidence system, the timing resolution of CdZnTe detector was improved significantly using the proposed timing correction method.Then we double the distance between source and LSO detector and make the data acquisition time four times longer, we obtained590μm Full Width at Half Maximum (FWHM) PET image of the Na-22point source (without correcting the source dimension). If we included the coincidence events of double-pixel charge sharing events from CdZnTe, the PET image resolution degrades to640μm FWHM. To interpolate the position of the charge sharing events between two neighbor pixels, we divided the350μm pixel pitch into five regions based on the ratio of signal amplitude of central pixel and neighbor pixel and also the count rate of each region, then we reconstructed the PET image using events from different regions individually. By interpolating the location of charge sharing events, the reconstructed PET image of Na-22point source was improved from640μm to560μm. We concluded that the image resolution of an animal PET scanner could be improved to500μm FWHM if we use350μm pitch pixelated CdZnTe detector as a PET insert and integrate such device into the PET scanner.