Click here to


Are you sure ?

Yes, do it No, cancel

High-Precision Dosimetry in Yttrium-90 Radioembolization Through Post-Procedural CT Imaging of Radiopaque Microspheres in a Porcine Model

C Henry1*, M Strugari1,2, G Mawko1,3,4,6, K Brewer1,2,5,R Abraham4,7, A Syme1,3,6, (1) Department of Physics, Dalhousie University, Halifax, NS, CA, (2) Biomedical Translational Imaging Centre (BIOTIC), Halifax, NS, CA, (3) Department of Medical Physics, Nova Scotia Health Authority, Halifax, NS, CA, (4) Department of Diagnostic Imaging and Interventional Radiology, Dalhousie University, Halifax, NS, CA, (5) Department of Biomedical Engineering, Dalhousie University, Halifax, NS, CA, (6) Department of Radiation Oncology, Dalhousie University, Halifax, NS, CA, (7) ABK Biomedical Inc., Halifax, NS, CA


(Sunday, 7/12/2020)   [Eastern Time (GMT-4)]

Room: AAPM ePoster Library

Purpose: To perform 3D convolution-based dosimetry in radioembolization through CT imaging of radiopaque microspheres administered to a porcine model.

Methods: The kidney of a hybrid farm pig was embolized with radiopaque 8?Y microspheres and imaged with CT [SOMATOM Definition AS] at 80 kVp. A total of 2 GBq of ?°Y activity was distributed throughout the delineated source voxels based on Hounsfield Units [HUs] to generate a CT-based activity distribution as previous work demonstrated linearity between HUs and radiopaque microsphere concentration. This CT-based activity distribution was then downsampled to simulate a SPECT-based activity distribution. Using the GATE v8.2 Monte Carlo toolkit, ?°Y dose-voxel kernels [DVKs] were generated in 3D grids with isotropic voxels of 3.375 mm³ and 125 mm³. CT- and simulated SPECT-based dosimetry was performed through 3D convolution of activity distributions with ?°Y DVKs. Dose-volume histograms were calculated to extract relevant dose metrics: D(1%) and D(avg).

Results: DVKs had relative uncertainties = 0.013% and = 0.367% at the central voxel and X90, respectively. CT-based dosimetry revealed metrics of D(avg) = 10.32 Gy and D(1%) = 10,480 Gy. Simulated SPECT-based dosimetry revealed metrics of D(avg) = 10.19 Gy and D(1%) = 5,140 Gy. While D(avg) values are comparable, the simulated SPECT-derived D(1%) is reduced by a factor of ~2 relative to the CT-derived D(1%). These results confirm the extreme dose gradients observed in histological studies of ex vivo liver samples receiving similar quantities of microspheres.

Conclusion: CT imaging of in vivo radiopaque microsphere distributions depicts dose distribution heterogeneity with increased accuracy relative to SPECT-based imaging due to the increased spatial resolution of CT. Precise knowledge of the absorbed dose distribution in radioembolization is essential for identifying undertreated tumour volumes, identifying radiation toxicity in adjacent healthy tissue, and developing NTCP and TCP models to establish tumour dose and healthy tissue toxicity thresholds.

Download ePoster [PDF]

Funding Support, Disclosures, and Conflict of Interest: C Henry is a trainee in the Cancer Research Training Program of the Beatrice Hunter Cancer Research Institute, with funds provided by the DMRF C. MacDougall Studentship.


Dosimetry, CT, Nuclear Medicine


IM- CT: Quantitative imaging/analysis

Contact Email