Abstract
351
Objectives We have used a phantom that represents an average patient's dimensions to measure time-of-flight (TOF) image quality improvements.
Methods The study was performed on a Discovery 690 PET/CT (GE Healthcare). This system has LYSO crystals (4.25 x 6.3 x 25 mm3) in four rings of 6x9 blocks. The phantom has an oval cross section (34x20 cm2 internal) modeled after average patient dimensions. It doesn't include heterogeneous tissue effects. The phantom had a uniform background and 1 cm hot spheres (six at 8:1 and six at 4:1). In addition, two cold bottles (filled with non-radioactive water) were placed in the phantom. Each bottle contained one 4:1 sphere. Images were reconstructed with and w/o TOF information, using 24 subsets and varying iterations: 1-5 for TOF and 1-9 w/o TOF. All corrections were applied. ROI's were placed for each hot sphere based on CT images. Additional 1 cm regions were placed to measure background variability, and regions were placed on the cold bottles. Contrast was measured and averaged for all 8:1 spheres and for all 4:1 spheres.
Results Each iteration of TOF image yielded contrast requiring at least 2x as many non-TOF iterations. Consistent with image appearance, TOF demonstrated 20%-50% reduction in noise at a particular contrast, compared to non-TOF. Contrast of the hot spheres in the cold bottles increased more rapidly with iteration and acheived higher levels with TOF.
Conclusions Given the constraints of this phantom in not representing the true variability of radiotracer uptake in patients, TOF capability added substantially to the quality of images of this medium-size phantom.
- © 2009 by Society of Nuclear Medicine