Optimization of the mechanical collimation for minibeam generation in proton minibeam radiation therapy

Purpose: The dose tolerances of normal tissues continue to be the main barrier in radiation therapy. To lower it, a novel concept based on a combination of proton therapy and the use of arrays of parallel and thin beams has been recently proposed: proton minibeam radiation therapy (pMBRT). It allies the inherent advantages of protons with the remarkable normal tissue preservation observed when irradiated with submillimetric spatially fractionated beams. Due to multiple Coulomb scattering, the tumor receives a homogeneous dose distribution, while normal tissues in the beam path benefit from the spatial fractionation of the dose. This promising technique has already been implemented at a clinical center (Proton therapy Center of Orsay) by means of a first prototype of a multislit collimator. The main goal of this work was to optimize the minibeam generation by means of a mechanical collimation. Methods: Monte Carlo simulations (GATE V7.1) were used to evaluate the influence of the collimator material (brass, nickel, iron, tungsten), thickness, phantom-to-collimator distance (PCD), among other parameters, on the dose distributions. Maximization of the peak-to-valley dose ratios (PVDR) in normal tissues along with minimization of full width at half maximum, penumbras and neutron contamination were used as figures of merit. As a starting point for the optimization, the collimator employed in our previous works was used. It consisted in 400 m x 2 cm slits with a center-to-center distance (c-t-c) of 3200 m. As the main targets of pMBRT will be neurological cases, 100 MeV energy proton minibeams were considered. This energy range would allow treating tumors located at the center of the brain (the worst scenario). Results: Tungsten and brass are the most advantageous materials among those considered. A tungsten collimator provides the highest PVDR and lowest penumbra. Although the neutron yield generated in the tungsten collimator is 3 times higher than that of the other materials, the biologic neutron doses at the patient position amount to less than 0.05% and 0.7% of the peak and valley doses, respectively. In addition, shorter PCD than the one currently used (7 cm) leads to thinner beams (enhancing the dose-volume effects), accompanied, however, by an increase of neutron dose at the phantom surface. Finally, no gain in dose distributions is obtained by using nonparallel slits. Conclusions: The collimator design and irradiation configuration have been optimized to minimize the angular spread, deliver the highest PVDR and the lowest valley possible in the normal tissues in pMBRT. We have also confirmed that even though the neutron yield generated in the multislit collimator is higher with respect to the one produced by the collimators used in conventional proton therapy, the increase of biological neutron dose in the patient will remain low (less than 1%). (C) 2017 American Association of Physicists in Medicine