Objective: An anthropomorphic pelvis phantom was made by the modification of a National Electrical Manufacturers Association (NEMA) phantom, currently the most popular one, and its clinical usefulness was evaluated.
Methods: The NEMA plus bone phantom was made by placing the pelvic bone model in the NEMA phantom. The NEMA plus bladder phantom was made by placing an imitation of the urinary bladder in the NEMA phantom. The pelvis phantom was also made by placing both the pelvic bone model and the bladder in the NEMA phantom. Four kinds of phantoms were imaged by both 2D and 3D dynamic modes, and for each phantom, prompt coincidence count rates, random ones, true plus scatter ones, and single photon rates were measured and these coincidence count rates were compared with those from the actual clinical data. After image reconstruction, the contrast ratio and image noise were also investigated.
Results: For the random coincidence count rate, the data obtained from each phantom showed good correspondence to the clinical data. The prompt coincidence count rates and true plus scatter ones of the clinical data were different from those obtained from NEMA phantom, NEMA plus bone one and NEMA plus bladder one, whereas there was a good correspondence between the data of the pelvis phantom and the clinical data. For the contrast ratio and image noise, there were discrepancies between the data of NEMA phantom and pelvis phantoms.
Conclusions: We made an anthropomorphic pelvis phantom by the simple modification of a NEMA phantom. This phantom showed performance similar to that of the actual human pelvis, suggesting clinical usefulness in the evaluation of new acquisition protocols and reconstruction algorithms.