Yttrium-90 post-radioembolization patient dosimetry with the Biograph Vision Quadra

Introduction: The purpose of the study is to evaluate the image quality and post-treatment dosimetry of patient data after Y-90 liver radioembolization using a whole-body Biograph Vision Quadra long axial field-of-view (LAFOV) PET/CT system, in terms of number of iterations and reconstruction times.

Methods: Data from one patient (male, 62 years old, injected activity 1.83 GBq) were acquired for 20 minutes with a Biograph Vision Quadra (Siemens Healthineers, Knoxville, TN, USA) which provides an AFOV of 106 cm. After list-mode data rebinning, images were reconstructed also with 15, 10, 5 and 1-minute frame durations. Each of the data were reconstructed with 2, 4, 6, and 8 iterations with 5 subsets and a Gaussian filter (2mm FWHM). The data were reconstructed with the ultra high-sensitivity (UHS) mode using the scanner's full acceptance angle with 3D OSEM, PSF + TOF, and a 220 x 220 matrix. 3D scatter correction was performed. Simplicit⁹⁰Y (Boston Scientific, MA, USA) was used for the pre- and post-treatment dosimetry calculations and syngo.via (Siemens Healthineers), for image processing and SUV evaluation. Spherical VOIs were placed on the most conspicuous lesion and on the liver background. The mean and standard deviation (SD) values were extracted. The SUV_bw-mean, of the lesion with the highest uptake was extracted. Threshold used was 40% isocontour of the SUV_bw-max. SNR was defined as (mean lesion VOI signal)/(background VOI SD). For the post-treatment dosimetry, the D₂₀, D₅₀ denote the minimum dose in Gray received by the 20, and 50% of the perfused VOI. V₂₀₅ is the percentage volume, which reached a minimum dose of 205 Gy (generally accepted target dose). The planned perfused viable tumor and perfused tumor absorbed doses were 400 and 386.6 Gy, respectively with a required activity of 1.81 GBq, based on the pre-treatment dosimetry. No statistical test was performed due to lack of enough data.

Results: The SNR exhibited similar values (<5% difference) for the 5, 10, 15 and 20 minutes reconstructions for all iterations. For the 1-minute reconstruction, the SNR was 16%, 28%, 37% and 41% lower than the rest of the reconstructions for the 2, 4, 6 and 8 iterations, respectively. SUV_bw-mean exhibited similar values for the 4, 6, and 8 iterations (range 2132-2401) while for the 2 iterations was lower (range 1903-2024) following a more stable plateau for all reconstructed times.

The whole liver normal tissue absorbed dose did not exhibit differences between different number of iterations and reconstructed times and was measured 35.3 ± 0.4 Gy. The perfused tumor absorbed dose was stable between iterations (396.5 ± 0.9), (379.1 ± 0.6), (359.3 ± 0.8), (351.1 ± 1.2) and (355.8 ± 1.1) Gy for the 1, 5, 10, 15 and 20-minutes reconstructed times, respectively. The perfused viable tumor absorbed dose was measured similarly (409.0 ± 1.4), (388.3 ± 0.7), (367.4 ± 0.9), (359.3 ± 1.3) and (363.8 ± 1.3) Gy.

For the D₂₀, the 1-minute reconstructions showed 298, 268, 231 and 201 Gy for 2, 4, 6 and 8 iterations, respectively while for the rest of the reconstructed times the values were similar for the different number of iterations (277 ± 2.5) Gy. Similarly for the D₅₀: 32, 13, 5 and 2 Gy for the 1-minute reconstructions and 53, 44, 35 and 29 Gy for the 5-minutes reconstructions while for the rest they showed more homogeneity (46.5 ± 5.6) Gy. The V₂₀₅ showed consistency for all reconstructed datasets (26 ± 1.3) %.

Conclusions: This preliminary study showed that the image SNR at 5-minutes reconstructed time is comparable to the original acquisition time for all datasets, while the SUV_bw-mean shows consistency for the 4, 6 and 8 iterations for the 5, 10, 15 and 20-minutes reconstructions while it is lower for the 2 iterations. For the post-treatment dosimetry, D₂₀ and D₅₀ quantities showed homogeneity from the 5-minutes reconstructions and above while the V₂₀₅ was stable for all reconstructed datasets. The study is ongoing and more patients will be evaluated, to reach a reliable conclusion.