Modeling of 3D Positron Emission Activity Distributions Induced by Proton Irradiation: A Semi-Empirical Method
O Lopatiuk-Tirpak1*, Z Su2, W Hsi3, O Zeidan3, S Meeks1, (1) MD Anderson Cancer Center Orlando, Orlando, FL, (2) University of Florida, Jacksonville, FL, (3) ProCure Treatment Centers, OKLAHOMA CITY, OKSU-E-T-14 Sunday 3:00:00 PM - 6:00:00 PM Room: Exhibit Hall
Purpose: to present and validate a method for modeling three-dimensional positron emission (PE) activity distributions induced by proton beam irradiation for PET/CT delivery verification studies in homogeneous media.
Methods: the method relies on modeling the 3D proton flux distribution by combining the analytical expression for the depth reduction of proton flux with the empirically obtained lateral distribution. The latter is extracted from the corresponding dose distribution under the assumption that the projectile energy is nearly constant in the lateral plane. The same assumption allows calculating the 3D induced activity distributions from proton flux distributions by parameterizing the energy-dependent activation cross-sections in terms of depth via the energy-range relation. Results of this modeling approach were validated against experimental PET/CT data from three phantom deliveries: unmodulated (pristine) beam, spread-out Bragg peak (SOBP) delivery without a range compensator, and SOBP with a range compensator. BANG3-Pro2 polymer gel was used as a phantom material because of its elemental soft-tissue equivalence.
Results: the agreement between modeled and measured activity distributions was evaluated using 3D gamma index analysis method, which, despite being traditionally reserved for dose distribution comparisons, is sufficiently general to be applied to other quantities. The evaluation criteria were dictated by limitations of PET imaging and were chosen to correspond to count rate uncertainty (6% value difference) and spatial resolution (4 mm distance to agreement). With these criteria and the threshold of 6%, the fraction of evaluated voxels passing the gamma evaluation was 97.9% for the pristine beam, 98.9% for the SOBP without compensator, and 98.5% for SOBP with compensator.
Conclusions: results of gamma evaluation indicate that the activity distributions produced by the model are consistent with experimental data within the uncertainties of PET imaging for clinical proton beams deliveries.