Ferramentas Computacionais para Otimização de Mapa de Fluência Radioterápico

Abstract

Cancer is one of the main causes of death in the world, for which the main form of treatment is radiotherapy, with Intensity Modulated Radiotherapy (IMRT) being the most advanced technique in terms of dose delivery. Different mathematical optimization models can be built as well as different solvers used to propose the dose to be emitted so that the dose absorbed in the tumor is the prescribed one and in other tissues it is the minimum. Thus, this work presents three instances, including two linear models and a quadratic model, which were used to evaluate their performance and implications regarding the construction of the optimization function in models that deal with the beam creep problem ( FMO) for the treatment of cancer in the prostate region. Open source and commercial solvers were used in the computational tests. Open source solvers included Clarabel, COSMO, HiGHS, and OSQP, while commercial solvers comprised Gurobi and CPLEX. In all instances, the results were evaluated in relation to the dose coverage in the tumor, and the percentage dose limits in the organs at risk, in addition to evaluating the performance in the different solvers. The results obtained showed that both commercial and open source solvers can find the solution to the problem in a short space of time. The histograms demonstrate that by minimizing the largest dose deviation with the N3 instance it is possible to reach 100% of the tumor tissue. By minimizing the average dose deviation, it is possible to reach 60% of the tissues of the tumor structure through instance N1 with the prescribed dose of 60 Gy, while instance N2 reaches only 33% of tumor tissue with 60 G