Abstract
89
Objectives: This article presents studies performed with the Discovery MI PET/CT system, a new time-of-flight (TOF) system based on silicon photomultipliers. System performance was characterized according to the NEMA NU-2 2012 standards. Comparisons of performance and clinical images were also made between this next-generation system and other commercially available PET/CT and PET/MR systems, as well as between different image reconstruction algorithms.
Methods: Spatial resolution, sensitivity, NECR, scatter fraction, count rate accuracy, and image quality were characterized according to the NEMA NU-2 2012 standards. In addition, energy and coincidence time resolution were measured using a line source at the center of the field-of-view (CFOV). Tests were conducted independently on two Discovery MI scanners installed at Stanford University Hospital and Uppsala University Hospital, and results were averaged between the two systems. In addition, back-to-back patient scans were performed between the Discovery MI PET/CT, Discovery 690 PET/CT, and SIGNA PET/MR systems. Clinical images were reconstructed with both ordered-subset expectation maximization (OSEM) reconstruction algorithms and the block-sequential regularized expectation maximization (BSREM) "Q.Clear" reconstruction algorithm, and examined qualitatively.
Results: The averaged FWHM of the radial, tangential, and axial spatial resolution reconstructed with filtered backprojection (FBP) at 1/10/20 cm from the system center are, respectively, 4.10/4.19/4.48 mm, 5.47/4.49/6.01 mm, and 7.53/4.90/6.10 mm. The averaged sensitivity is 13.7 cps/kBq at the center and 13.4 cps/kBq at 10 cm radial offset from the center. Averaged peak noise equivalent count rate (NECR) is 193.4 kcps at 21.9 kBq/mL with a scatter fraction (SF) of 40.6%. The averaged contrast recovery (CR) coefficients for the image quality (IQ) phantom are 53.7/64.0/73.1/82.7/86.8/90.7 for the 10/13/17/22/28/37 mm diameter spheres over 3 separate acquisitions. The average photopeak energy resolution is 9.40% FWHM and the average coincidence time resolution is 375.4 ps FWHM. Clinical image comparisons between the PET/CT systems demonstrate the very high quality of the Discovery MI system. Comparisons between the Discovery MI PET/CT and SIGNA PET/MR systems, which contain identical detector architectures but with different detector diameters, show similar spatial resolution and overall imaging performance. Lastly, results indicate significant image quality and contrast-to-noise performance enhancement for the "Q.Clear" reconstruction algorithm when compared to OSEM.
Conclusion: Excellent performance was achieved with the new Discovery MI system, including 375 ps FWHM coincidence time resolution and sensitivity of 14 cps/kBq. Comparisons between different image reconstruction algorithms and other multimodal SiPM and non-SiPM-based PET detector system designs indicate substantial performance enhancements are possible with this next-generation system. Research Support: None