Distant Ray Therapy Planning
Algorithms and Codes for Distant Ray Therapy Optimal Planning Based on Transport Theory and Parallel Calculations (Prototype of New Generation of Ray Therapy Computation Complex)
Tech Area / Field
- BIO-CGM/Cytology, Genetics and Molecular Biology/Biotechnology
3 Approved without Funding
MIFI, Russia, Moscow
- Cancer Research Center, Russia, Moscow\nKeldysh Institute of Applied Mathematics, Russia, Moscow
- University of Maryland / School of Medicine / Department of Radiation Oncology, USA, MD, Baltimore\nUniversity of Thessaloniki / Chaos @ Innovation Unit, Greece, Thessaloniki\nBaylor College of Medicine, USA, TX, Houston\nUniversity of Michigan / Colledge of Engeneering, USA, MI, Ann Arbor\nKrolinska Institute / Department of Medical Radiation Physics, Sweden, Stockholm
Project summaryFor recent years medicine accelerators, modern systems of radiation collimation, various high precision tomographs and another types of high precision devices become to be widely exploited at the radiation therapy. Such wide introduction of high precision devices inspired the development of new modern algorithms for dose distribution calculation, that guarantee high calculation accuracy, consistent with modern radiology requirements. As a result 3D systems of dosimetry planning optimization and corresponding algorithms of 3D dose calculation and optimization come to be exploited in leading hospitals.
Currently two types of numerical methods are mainly used for 3D dose calculation:
1) convolution/superposition method based on dose kernel calculation for differential pencil beam (dose spread function); 2) various modifications of pencil beam method. These methods replaced previously used empirical and semi-empirical methods. The new methods are sufficiently fast and provide necessary accuracy in the case of homogeneous media. However, the error drastically increases in presence of heterogeneity inside calculation domain and also in the vicinity of patient body surface.
The statistical Monte Carlo method (MCM), developed for ionizing radiation transport problems, is usually regarded as the unique one to guarantee the accuracy, recommended by international committees for calculations in domains of complicated geometry.
Nevertheless, a number of different methods exists in transport theory, providing the capability to calculate ionizing radiation field characteristics in domains of complicated geometry. Among them the Discrete Ordinates Methods (DOM) can be viewed as the most advanced ones. However, despite their advantages in comparison with another methods, both DOM and MCM require for 3D geometry very massive calculations. These calculations can take many running hours while being realized on single-processor personal computers (even of Pentium-4 type).
The situation is certainly unacceptable for multiple calculations in tasks of radiation dose planning. At the same time the perspective of precise calculation fulfillment under acceptable running time in the case of 3D geometry problems, including any nonhomogeneity types, represents actual problem needing to be solved. The solution can be achieved via development of parallel calculation algorithms implemented on multi-processor computers. The Project authors have a significant experience on parallel computation code creating for 3D physical nuclear reactor calculations and also for calculations of nuclear power plant radiation shielding.
The first activity direction within the Project framework is related to development of code modules for parallel architecture computers, based on DOM method, for calculations of radiation, generated by machine head, and dose distributions from г/е-radiation beams.
The second direction is related to creation of superfast code module based on MCM for 3D dose distribution calculations in dose planning problems.
Today it takes several hours for solving of irradiation planning problems in regime of photon beam intensity modulation (IMRT) with necessary accuracy, using the most fast modern MCM algorithms on modern computer work stations.
The Project authors previously developed significantly new fast version of MCM for radiation dose calculations inside patient body. The algorithm main feature is that high computation speed is achieved not only via reduction of running time, necessary for trajectory modeling (like it is usually done in similar codes), but mainly via application of new types of dose estimating (namely, PL-estimations). The PL-estimation introduction allowed to gain almost order in running time reduction. Besides, the algorithm admits almost 100% computation parallelizing. In a whole these algorithm features will permit to reduce running time, necessary for dose distribution calculation in irradiated volume by MCM up to one-two minutes. However, to achieve high accuracy of dose distribution calculation in the whole volume the detailed precise information about the radiation, incident on patient, is necessary. The exploitation of code module based on DOM seems to be more effective for getting these data.
The most important stage of modern dosimetry planning consists in searching of optimal irradiation planning, including definition of irradiating beam number, beam directions and optimal beam intensity transversal modulation. Correct solving of all problem complex under all the factors taking into account requires massive computations and corresponding great running times in the case of single-processor regime of code implementation. Within the Project framework we suggest to develop new regularization algorithms and new parallel computation codes based on previously developed ones by the authors within the framework of ISTC Project №1079/99. The new codes will guarantee stability and correctness of numerical solution for the whole optimization problem complex and besides will provide acceptable for clinical practice running time of problem solving.
Thus, the main Project objectives consist in:
1. Development of new modern computation complex of external radiation therapy for fast and accurate solving of 3D problems of г/е-irradiation planning and optimization, including code implementation on parallel architecture computers. In particular it includes:
· computation module for calculation of spatial, energetic, angular characteristics of radiation field, generated by accelerator head, and radiation dose distributions inside patient body (numerical algorithm is based on DOM);
· computation module for calculation of dose distributions inside irradiated volume via Monte Carlo method with significantly new high effective method of dose estimation;
· computation module for irradiation optimization.
2. Testing of the new created complex at clinical base of Cancer Research Center.
The International Science and Technology Center (ISTC) is an intergovernmental organization connecting scientists from Kazakhstan, Armenia, Tajikistan, Kyrgyzstan, and Georgia with their peers and research organizations in the EU, Japan, Republic of Korea, Norway and the United States.
ISTC facilitates international science projects and assists the global scientific and business community to source and engage with CIS and Georgian institutes that develop or possess an excellence of scientific know-how.