Room: Exhibit Hall
Purpose: Dosimetry in multi-seed brachytherapy implants are done by using superposition of the dose around single sources, ignoring the inter-seed effect. The purpose of this study is to use feed-forward neural network for correction of the inter-seed effects in prostate brachytherapy using I-125, and Pd-103 sources.
Methods: The dose distribution around different models of I-125, and Pd-103 brachytherapy sources were obtained using the MCNP Monte carlo simulations. Dummy pellets were then simulated at different distances, and angles from the active source, and its effect was obtained by MCNP5 simulation. Then the simulation results were used for training a feed-forward three-layer artificial neural network. Then the network was used for correcting the dose distribution obtained by superposition.
Results: According to the results of this study, the feed-forward network was able to predict the inter-seed effect with less than 1% for I-125, 2% for Pd-103. For a combination of sources used for brachytherapy treatment, the dose distribution may be obtained with high accuracy.
Conclusion: The method proposed in this study, the dose distribution in prostate obtained by adding the dose of each single source, can be effectively corrected for inter-seed effect.