Advances in Monte Carlo patient-specific internal dosimetry for 90 Y-TARE treatments.

Three-dimensional internal dosimetry (ID) is increasingly used in planning Trans-Arterial Radio-Embolization (TARE) treatments of liver tumors, with growing interest in the patientspecific Monte Carlo (MC) approach. Direct MC simulation is the gold standard for ID, but requires relatively long computational times. We present our recent studies aiming at speeding up MC 90Y-TARE ID through the optimization of parameters such as CT resolution and production cuts on secondary particles. Using the GEANT4-based toolkit GATE, we carried out multiple simulations varying CT resampling and production cuts; we found combinations of parameters reducing simulation time up to 30–40% while maintaining high dosimetric accuracy. We moreover discuss ongoing studies focusing on further optimization of MC 90Y-TARE ID by reducing background noise and dose artefacts in air-rich regions of patient’s body. For this purpose, we developed filtering techniques of the input SPECT data. Comparing the results obtained with filtered and unfiltered data, non-negligible discrepancies in lungs doses emerge, with relative differences exceeding even 40% depending on the adopted filtering procedure.