Impact of the Monte Carlo code GATE for Targeted Radionuclide Therapy (TRT): imaging and dosimetric calculations.

Objectives Targeted Radionuclide Therapy (TRT) combines the specificity of a biological vector with the action of short-range radiations. Assessing the absorbed dose delivered to tumors and healthy tissues participates to the evaluation and optimization of the therapy. It implies combining quantitative imaging and absorbed dose calculations. Monte Carlo modeling can be involved in both steps: firstly as a process control aid for the improvement of image quantification, and secondly for the personalization of absorbed dose calculations by taking into account organ shape and density. The Monte Carlo code GATE [1] has been widely used for scintigraphic image modeling, both for preclinical and clinical PET/SPECT studies, and is available for absorbed dose calculations in nuclear medicine since its version 6.0. Therefore the possibility to use the same code to model both aspects of radiopharmaceutical dosimetry was considered as an asset. This works investigates the input of GATE as a toolkit for applications in internal dosimetry in a context of improvement of dosimetric methods. A wide heterogeneity is observed in dosimetric approaches and no standardized dosimetric protocol has been proposed to date. The DosiTest project (www.dositest.com) aims at evaluating the impact of the various steps contributing to the realization of a dosimetric study, by means of a virtual multi-centric inter-comparison based on Monte-Carlo modeling, and eventually proposing a reference methodology.

Methods To this end, pharmacokinetics of two radiopharmaceuticals (OCTREOSCAN[TM] and LUTATHERA[TM]) was created following a compartmental modeling. Two virtual patients were defined from these radiopharmaceuticals and from two anthropomorphic models (XCAT [2] and ICRP 110 [3] female reference computational model) split in functional compartments. The generation of scintigraphic images was performed with GATE v6.2 from these virtual patients following imaging protocols dedicated to each radiopharmaceutical.

Results Computation times (on a 480 virtual cores computing cluster), for “step and shoot” whole body simulations, with acceptable statistics, of the XCAT/OCTREOSCAN[TM] patient were: 10 days for extra-vascular fluid, 28h for blood, 12h for liver, 7h for kidneys, and 1-2h for bladder, spleen and liver tumors. With 15 projections per head (on a virtual 4-headed gamma-camera), tomographic simulations were 3 times longer. Reference dosimetric calculations were performed with GATE v7.1, for a further comparison with results obtained by participating centers: for the ICRP 110 virtual patient, after an injection of 5.5 GBq of LUTATHERA[TM], absorbed dose results were 1.00 Gy in liver, 7.77 Gy in spleen, 5.59 Gy in kidneys, 7.45 Gy in urinary bladder wall and respectively 60.23 and 3.78 Gy in the liver tumors.

Conclusions The ability to generate both scintigraphic images and absorbed dose maps with GATE was highlighted in this study. The next steps of DosiTest will be to decrease computation times, then to generate scintigraphic images for several clinical centers, in order to compare dosimetric results with our reference.