Development and implementation of a non Gaussian model for the lateral dose prediction in a proton therapy treatment planning system

Challenging issues in treatment planning system for hadrontherapy are the accurate and fast calculation of dose distribution, the reduction in memory space required to store the dose kernel of individual pencil beams and the shortening of computation time for dose optimization and calculation. In this framework, the prediction of lateral dose distributions is a topic of great interest because currently the double gaussian parametrization is typically used as approximation although other parameterizations are also available. The best accuracy for this kind of calculations can be obtained by Monte Carlo methods, at the expense of a long computing time. This work aims to present a flexible computational model for the calculation of the lateral profile of a pencil proton beam and the results of its implementation in a treatment planning system. The model calculation are compared with the currently used double gaussian approximation and the Monte Carlo calculations, and the tests are performed in water and in presence of inhomogeneities.