Room: Stars at Night Ballroom 1
Purpose: To demonstrate the efficacy of using Tc-99m MAA SPECT/CT to predict Y-90 microsphere voxel-based tumor dose.
Methods: A commercial software imaging and dosimetry platform (Velocity, Varian Medical Systems) was used to evaluate ten gross tumor volumes (GTVs) for nine liver patients that were imaged using Tc-99m MAA SPECT/CT prior to Y-90 therapy and quantitative bremsstrahlung SPECT/CT after Y-90 injection. For each case, SPECT/CTs were registered to a diagnostic MR or CT using deformable or rigid registration algorithms. Voxel-based dose distributions were computed using the local deposition method (LDM) and the known injected Y-90 activity. Dose-volume histograms were calculated for the GTVs.
Results: Injected Y-90 activities and tumor volumes ranged from 1.330-3.583 GBq, and 1.8-73.8 cc, respectively. GTV mean doses ranged from 150.3–361.8 Gy (244.4 +/- 57.6 Gy) and 140.2-345.9 Gy (244.9 +/- 66.4 Gy), with a Pearson correlation coefficient of R = 0.925 (strong correlation) between the Tc-99m MAA and Y-90 studies, respectively. On average, the Tc-99m MAA imaging study could be used to predict the GTV mean dose to within +/- 10% (range -13.2% to 16.5%) of the value calculated in the Y-90 imaging study.
Conclusion: The use of Tc-99m MAA SPECT/CT for predictive dosimetry of Y-90 tumor absorbed dose displayed reasonable accuracy. However, due to the inherent physical difference between the MAA particles and the Y-90 microspheres, a strong correlation between the two modalities is dependent on mitigating the differences in catheter position, the time between imaging studies, the imaging voxel size, and image registration uncertainties. With further development, this method could potentially be used to facilitate patient-specific activity prescriptions for Y-90 radionuclide therapy.
Funding Support, Disclosures, and Conflict of Interest: Potrebko, Andl, Bose, and Fox are employees of Varian Medical Systems.