Relationships Between Noise-Equivalent Count Rates For Extended Nema Nu 2-Like Scatter Phantoms And A Human Subject Scanned Using The Explorer Total-Body Pet Scanner
Edwin Leung,Xuezhu Zhang,Eric Berg,Hongcheng Shi,Shuguang Chen,Pengcheng Hu,Yun Dong,Yong Zhao,Songsong Tang,Xinyu Lv,Shaoping Chu,Weiping Liu,Hongdi Li,Terry Jones,Simon Cherry,Ramsey Badawi
2019-01-01
Journal of Nuclear Medicine
Abstract:1385 Introduction: Noise-equivalent count rate (NECR) is a performance metric that is used to evaluate the performance of PET scanners (Strother 1990). Conventionally, NECR is measured by scanning the NEMA NU 2 scatter phantom (20 cm diameter x 70 cm long). For the EXPLORER total-body PET scanner, however, the NU 2 phantom is considerably shorter than the axial FOV. We measured NECR using extended (ext.) NU 2-like scatter phantoms, in addition to the NU 2 phantom. We then compared the phantom NECR measurements to NECR measurements derived from a human subject to determine the phantom length that best approx. the count rate and noise properties in the scanner. Methods: Phantoms: a NEMA NU 2 phantom, and 2 ext. NU 2-like phantoms (175 cm and 210 cm long) were used to measure the NECR of the scanner. Data acq. and analysis were performed according to the NEMA NU 2-2012 protocol for all phantoms. Human subject: a female subject (height: 156 cm, weight: 56 kg) was injected with 256 MBq of 18F-FDG. 5 min of list-mode data were used (from 55 min post-injection) to calculate human-derived NECR. The volume of the subject was determined based on a CT volume excluding voxels below the attenuation coefficient of the lungs in the volume. A block-based scatter sinogram, generated from a Monte Carlo simulation and fully compressed axially was thresholded to determine an approx. outline of the subject. Randoms, prompt and scattered coincidence data falling outside of the thresholded region were excluded. Maximum unit difference: the EXPLORER scanner consists of 8 PET detector module ring units, with each unit being ~ 24 cm in length. The NECR was measured for varying maximum unit differences (MUDs) ranging from 0 - 4 to assess the effect of acceptance angle on NECR. The MUD range corresponds to an axial extent from ~ 24 cm - 120 cm. Results: The results showed that at 3.2 kBq/cc, the NEC1R of the subject was the highest with a MUD = 2 (1294 kcps), following by MUDs = 3 (1255 kcps), 4 (1177 kcps), 1 (993 kcps) and 0 (434 kcps). Also, only the NU 2 phantom had the highest NEC1R peak with MUD = 4. The two ext. NU 2-like scatter phantoms had the highest NEC1R peaks at either MUD = 2 or 3. Discussion At an activity conc. of 3.2 kBq/cc, the NECR of the subject was much closer to the NECR of the ext. phantoms than the NU 2 phantom. There were large differences in NECR between the human/ext. phantoms and the NU 2 phantom for all MUDs. At first glance, it may seem natural to think that the larger the MUD, the higher the peak NECR since more true coincidences are collected when using a wider acceptance angle. However, a larger MUD also leads to increased scattered and random coincidences. Our results suggest that careful choice of the MUD is needed depending on the imaging conditions. Our results also seem to agree with the data presented in Poon et al. (2012) where, for a phantom covering the full extent of the axial FOV, there is “no advantage in increasing the ring difference corresponding to an axial extent greater than ~ 700 mm” (which corresponds to a MUD = ~ 2 as the axial length of an EXPLORER PET detector unit is ~ 24 cm). Although TOF was not incl. in the NECR assessment, it is expected that the NECR (w/ TOF) will increase since TOF increases the value of true coincidences and decreases for random coincidences. The optimal MUD may also change after incl. TOF in the NECR assessment. The excellent agreement between the human NECR and the ext. phantom NECR should not be over-interpreted. Human body habitus varies widely and a number of approx. were made in estimating the NECR for the human subject. Conclusions: We conclude that ext. NEMA NU 2-like phantoms may be more appropriate for assessing the NECR of long axial FOV PET scanners, such as the EXPLORER scanner. We showed that ext. NU 2-like phantoms may more appropriately reflect the count rate conditions when imaging a human subject compared to the NU 2 phantom, and suggests their future use for estimating the performance gains of long axial FOV PET scanners.