Therapeutic gain prediction for evaluation and optimization of neutron spectra in BNCT

In this work, we propose a method for fast evaluation of neutron beam spectra for the treatment of the brain tumors using boron neutron capture therapy. This method applies an artificial neural network to predict the therapeutic gain which is a very important parameter in the evaluation of the quality of neutron beam spectra. In this way, we calculated the dose delivered to the tumor region and the maximum dose delivered to healthy tissues, for various neutron beam energy levels through the Snyder head phantom. The calculations were carried out using the MCNP Monte Carlo code and the results were used to train the artificial neural network in the learning process. The trained network can be assumed as a function that predicts the therapeutic gain of any neutron spectrum. The results of this study indicated that the trained artificial neural network was able to produce an accurate prediction of the therapeutic gain for any neutron beam spectrum.

► A new method was proposed for fast evaluation of neutron beam spectra in BNCT.
► An ANN was used for the prediction of TG for the evaluation of the neutron beam spectra.
► Dose calculations were performed using the MCNP code and were used to train the ANN.
► Trained ANN was able to produce accurate prediction of the TG for any neutron beam energy.