90Y Imaging using both SPECT and PET

Objectives 90Y-DOTATOC has recently shown great promise in treating patients with neuroendocrine tumors in clinical practice. The aim of this study is to perform SPECT and PET imaging studies in realistic distribution of 90Y-DOTATOC and identify the best approach for 90Y imaging.

Methods An anthropomorphic torso phantom was filled with 21.1 mCi of 90Y chloride with realistic contrast ratios for the background to liver, spine, and a number of spherical “lesions” ranging from 2.2 to 3.8 cm in diameter. The spine was filled with liquid of bone-equivalent density to mimic bony uptake. The spheres were attached to the liver (SP-LIVER), left lung (SP-LUNG), and spine (SP-SPINE and SP-SPINE-BONE). All of the spheres were filled with radioactive water except SP-SPINE-BONE, which was filled with liquid bone. Noise-free 90Y bremsstrahlung SPECT projections were acquired with both medium-energy (ME) and high-energy (HE) collimators. The ME data were scaled to the total counts of a typical clinical study using ME collimators. The HE data were scaled to match the ME imaging time of the same activity. Poisson deviates were generated for each study separately. 90Y bremsstrahlung projections were reconstructed using OSEM (8 subsets, 10 iterations) while modeling the point spread function (PSF). Next, 90Y PET list-mode data were acquired on a PET-CT (Siemens Biograph) with equivalent imaging time and reconstructed using OSEM-PSF (8 subsets, 10 iterations). No post-reconstruction filtering was applied.

Results Noise levels were 31.1%, 47.1% and 141.3% for SPECT-ME, SPECT-HE and PET, respectively. At the 10th iteration, the lesion-to-background contrasts for ME/HE were 7.0/9.3, 4.8/10.7, 2.2/3.1 and 20.2/43.7 for SP-LIVER, SP-LUNG, SP-SPINE and SP-SPINE-BONE, respectively.

Conclusions HE collimator yielded better lesion contrast than ME collimator for 90Y bremsstrahlung imaging. For 90Y-DOTATOC imaging, SPECT yielded superior image quality to PET.

Research Support NIH R21CA140771