Room: Track 1
We aimed to calculate bone lesion absorbed doses for Lu-177-PSMA-therapy of metastatic castration-resistant prostate cancer (mCRPC) patients with respect to the heterogeneous bone structure. Full Monte Carlo (MC) simulations (considered gold standard) accounts fully for different tissue structures using information obtained from a patient CT scan and time-integrated activity distributions from SPECT scans, while other dosimetric methods do not account for differences in tissue densities. Since MC remains complex and time-consuming, we compare the MC obtained absorbed doses to those estimated with a simpler and faster voxel S-value method. The latter, has become clinically available.
7 patients (1st cycle Lu-177-PSMA, therapy activities 7.4-9GBq) were included in this study. MC dose simulations were performed using GATE v8.2. The simulations used patient CT and 3D time-integrated activity distribution map, obtained from 24h, 48h, 72h QSPECT/CT with three bed positions. Voxel S-value kernels were also simulated for the ICRP soft tissue. Dose estimates with the S-value kernels were calculated in 2 ways: a) using the soft tissue S-value kernels and b) using CT-based voxel density weighting for the doses in a. Lesion-wise dose estimates, with and without density weighting, were compared to MC.
A total of 102 bone lesions were investigated. Absorbed dose estimates from voxelized S-values without density weighting overestimated the doses by +19% (range: -12% to +57%) when compared to those obtained with MC. Applying CT-based density weighting reduced the deviation to -4% (range: -9% to -3%).
Our results suggest that bone lesion doses obtained with voxelized S-values in combination with CT-based density weighting agree well with those obtained with MC. This is of great interest for clinical routine as it can enable a fast and accurate dose assessment for tumor metastases in heterogeneous bone structures for mCRPC patients undergoing Lu-177-PSMA-therapy and other radionuclide therapies.