Early-phase pelvic bone SPECT: Simulation and comparison of several acquisition protocols to reduce bladder artifact and improve image quality

Tomoscintigraphic reconstruction in nuclear medicine assumes that the distribution of the tracer is unchanged in the volume of interest throughout the duration of the acquisition. This condition is however not met in early-phase bone scintigraphy and early-phase pelvic SPECT may display helical artifacts due to the filling of the bladder. Those artifacts may hamper proper interpretation of surrounding bone areas. The aim of this study was to construct a 4D digital pelvic phantom to simulate different acquisition protocols and optimize the acquisition. A 4D digital pelvic phantom was generated with a dynamic component consisting in an expanding bladder with 2 ureters and a static part consisting in the 2 kidneys, bone structures, and soft tissues. Projection data were obtained using an attenuated Radon transform function. Four acquisitions protocols were tested: 32 projections of 16 seconds (32–16–1), 32 projections of 8 seconds (32–8–1), 2 consecutive SPECT of 32 projections of 4 seconds (32–4–2) and 2 consecutive SPECT of 16 projections of 8 seconds (16–8–2). The optimal protocol was then tested on one patient. The amplitude of the artifacts was reduced with the 32–8–1, 32–4–2, and 16–8–2 protocols. The 16–8–2 protocol had the highest signal to noise ratio among those 3 protocols. The bladder artifact was visually markedly reduced on the patient acquisition with a 16–8–2 protocol. Two successive early-phase bone SPECT, with a lower number of projection than the usual protocol reduce the impact of the helical artifacts around the bladder.