Abstract
543
Objectives 177Lu labelled somatostatin analog therapies require accurate patient specific post-therapy kidney dosimetry for patient safety. Physically accurate Monte Carlo (MC) simulations are unfortunately time consuming and often impractical to carry out in clinical environments. The aim of this study is to compare a new semi-MC dosimetry code with local electron absorption assumption combined with simplified MC-based photon simulation to exact MC simulations.
Methods A XCAT phantom was created with realistic 177Lu tissue activities and densities. The activities were based on an average of ten 177Lu-DOTATATE treated patients at 24 hours post-injection with spleen uptake. The phantom was used to generate SPECT projection data using an in-house MC-simulator with typical 177Lu SPECT acquisition parameters. Image reconstruction was carried out using OSEM with attenuation, collimator-detector-response and MC-based scatter compensations. Absorbed dose rate distributions were calculated based on the original phantom activity distributions (DPhantom) and activity distributions obtained from the reconstructed images (DImage) using a physically accurate PENELOPE based vxlPen code and semi-MC code. Absorbed doses to both kidneys were calculated for original phantom and reconstructed image based activity distributions.
Results In the phantom, total kidney absorbed dose difference between physically accurate dose DPhantom-vxlPen and semi-MC DPhantom-semiMC were 0.3% and 0.2% for left and right kidneys respectively. Whereas the differences in image based doses DImage-vxlPen and DImage-semiMC were 0.4% and 0.1%. Only 4% of the kidney dose was produced by photons in all cases. The semi-MC method was approximately 250 times faster than vxlPen.
Conclusions Kidney dosimetry for 177Lu labelled somatostatin analog therapies can be carried out very fast by assuming local electron absorption and using simplified photon MC-simulation.