Development of tools for quality control on therapeutic carbon beams with a fast-MC code (FRED)

In the fight against tumors, different types of cancer require different ways of treatment: surgery, radiotherapy, chemotherapy, hormone therapy and immunotherapy often used in combination with each other. About 50% of cancer patients undergo radiotherapy treatment which exploits the ability of ionizing radiation to damage the genetic heritage of cancer cells, causing apoptosis and preventing their reproduction. The non-invasive nature of radiation represents a viable alternative for those tumors that are not surgically operable because they are localized in hardly reachable anatomical sites or on organs which removal would be too disabling for the patient. A new frontier of radiotherapy is represented by Particle Therapy (PT). It consists of the use of accelerated charged particle beams (in particular protons and carbon ions) to irradiate solid tumors. The main advantage of such a technique with respect to the standard radiotherapy using x-rays/electron beams is in the different longitudinal energy release profiles. While photons’ longitudinal dose release is characterized by a slow exponential decrease, for charged particles a sharp peak at the end of the path provides a more selective energy release. By conveniently controlling the peak position it is possible to concentrate the dose (expressed as the energy release per unit mass) to tumors and, at the same time, preserve surrounding healthy tissues. In particle therapy treatments, the achieved steep dose gradients demand highly accurate modelling of the interaction of beam particles with tissues. The high ballistic precision of hadrons may result in a superior delivered dose distribution compared to conventional radiotherapy only if accompanied by a precise patient positioning and highly accurate treatment planning. This second operation is performed by the Treatment Planning System (TPS), sophisticated software that provides position, intensity and direction of the beams to the accelerator control system. Nowadays one of the major issues related to the TPS based on Monte Carlo (MC) is the high computational time required to meet the demand for high accuracy. The code FRED (Fast paRticle thErapy Dose evaluator) has been developed to allow a fast optimization of treatment plans in proton therapy while profiting from the dose release accuracy of a MC tool. Within FRED, the proton interactions are described with the precision level available in leading-edge MC tools used for medical physics applications, with the advantage of reducing the simulation time up to a factor of 1000. In this way, it allows a MC plan recalculation in a few minutes on GPU (Graphics Processing Unit) cards, instead of several hours on CPU (Central Processing Unit) hardware. For the exceptional speed of the proton tracking algorithms implemented in FRED and for the excellent results achieved, the door to several applications within the particle therapy field has been opened. In particular, the success of FRED with protons determined the interest of CNAO (Centro Nazionale di Adroterapia Oncologica) center in Pavia to develop FRED also for carbon therapy applications, to recalculate treatment plans with carbon ions. Among the several differences between proton and carbon beams, the nuclear fragmentation of the projectile in a 12C treatment, which does not occur with protons, is certainly the most important. The simulation of the ion beam fragmentation gives an important contribution to the dose deposition. The total dose released is due not only to the primary beam but also to secondary and tertiary particles. Also for proton beams, there are secondary particles, mostly secondary protons from target fragmentation, which contribute on the level of some percent to the dose deposition for higher proton beam energies. However, fragments of the projectile, produced only by carbon beams, having on average the same energy per nucleon of the primary beam and a lower mass, can release dose after the peak causing the well-known fragmentation tail. This thesis is focused on the development of a fast-MC simulating the carbon treatment in particle therapy, with an entirely new nuclear interaction model of carbon on light target nuclei. The model has been developed to be implemented in the GPU based MC code, FRED. For this reason, in developing the algorithms the goal has been to balance accuracy, calculation time and GPU execution guidelines. In particular, maximum attention has been given to physical processes relevant for dose and RBE-weighted dose computation. Moreover, where possible, look-up tables have been implemented instead of performing an explicit calculation in view of the GPU implementation. Some aspects of the interaction of carbon ions with matter are analogous to the ones already used in FRED for proton beams. In particular, for ionization energy loss and multiple scattering, only a few adjustments were necessary. On the contrary, the nuclear model was built from scratch. The approach has been to develop the nuclear model parameterizing existent data and applying physical scaling in the energy range where the data are missing. The elastic cross-section has been obtained from ENDF/B-VII data while the calculation of the non-elastic cross-section was based on results reported on Tacheki, Zhang and Kox papers. Data used for the sampling of the combination of emitted fragments, energy and angle distributions, are relatives to the Dudouet and Divay experiments. To fill the gaps in the experimental data, an intercomparison between FRED and the full-MC FLUKA has been of help to check the adopted scaling. The model has been tested against the full-MC code FLUKA, commonly used in particle therapy, and then with two of the few experiments that it is possible to find in literature. The agreement with FLUKA is excellent, especially for lower energies