Suppressing bladder artifacts in bone SPECT of the pelvis.

ABSTRACT

OBJECTIVE: Bladder-filling reconstruction artifacts have a detrimental effect on the image quality of pelvic bone single photon emission computed tomography (SPECT). Using a simple protocol consisting of forced diuresis coupled with intravenous (IV) hydration, this study was undertaken to obtain an artifact-free pelvic SPECT after discarding the residual urinary activity. METHODS: Thirty patients were enrolled. In group I, pelvic SPECT was performed directly after normal void, whereas in group II, SPECT was preceded by IV injection of 0.5 mg/kg furosemide (maximum 40 mg) coupled with IV infusion of 500 cc of physiologic saline. Bladder-filling reconstruction artifacts were analyzed in group I patients, who had their images reconstructed using both filtered backprojection and iterative algorithms, both qualitatively and quantitatively by means of regions of interest (ROIs) drawn around the artifact-bearing bone areas as well as the corresponding contralateral sites. For group II patients, besides visual analysis, ROIs were placed over the sites corresponding to those of the group I patients. In every patient, total counts of each ROI were normalized to a reference ROI placed over the sacrum, and a ratio was created. RESULTS: Using filtered backprojection, two forms of artifacts were identified in group I patients first, a streak pattern that extended to the sacro-iliac joint in nine (60%) patients, the hip joint in five (33%), the superior pubic rami in four (27%), the sacrum in three (20%), and the ischium in one (6%); second, a count loss subtype which extended to the hip joints in nine (60%) patients. Corresponding values after iterative reconstruction were two (13%) for the sacro-iliac joint, three (20%) for the hip joint, one (6%) for the superior pubic ramus, and one (6%) for the sacrum. In five (33%) patients, residual count loss artifacts were still identifiable after iterative reconstruction. However in group II, no such effects were observed because the bladder activity reached near background level in 14 (93%) of 15 patients after three successive voids with a 3.5-fold decrease in the mean value of total bladder count in comparison with group I patients. A statistically significant difference was found between artifact- and non-artifact-harboring ROIs in group I whichever the method used for reconstruction, whereas the values of right and left hemi-pelvis ROIs/sacrum in group II were almost identical. CONCLUSIONS: Forced diuresis coupled with parenteral hydration facilitates the acquisition of an artifact-free pelvic SPECT. Especially for clinical questions that focus on femoral heads and pubic bones, applying the aforementioned protocol may improve the diagnostic accuracy of pelvic bone SPECT.