Title :
Artificial Neural Network Model for Spectral Construction of a Linear Accelerator Megavoltage Photon Beam
Author_Institution :
Univ. of Birmingham, Birmingham, UK
Abstract :
In photon radiotherapy, one of the most critical parameters for the delivery of highly target-conformal dose to the target volume is the determination of the energy spectra incident on the surface of the target volume. This paper proposes an artificial neural network (ANN) model for the determination of megavoltage photon beams spectra. We have computed bremsstrahlung spectra of 4 to 25 MV beams for different medical linear accelerators. The ANN model is trained with thousands of data sets obtained from the Monte Carlo (MC) constructed spectra data of different energies. The computed spectra from the ANN model are verified by its comparison with published Monte Carlo spectra data and we have found an excellent agreement in the results. A major advantage of the ANN model is its ability of instantaneous data analysis after learning phase, enabling automated data analysis within short period for dose optimization with greater accuracy.
Keywords :
Monte Carlo methods; biological effects of radiation; bremsstrahlung; dosimetry; linear accelerators; medical computing; neural nets; photons; radiation therapy; ANN model; Monte Carlo spectra data; artificial neural network; bremsstrahlung spectra; energy spectra; instantaneous data analysis; learning phase; linear accelerator megavoltage photon beam; medical linear accelerator; megavoltage photon beams spectra; photon radiotherapy; spectral construction; Artificial neural networks; Computational modeling; Computer networks; Data analysis; Linear accelerators; Monte Carlo methods; Optimization methods; Particle beams; Robustness; Structural beams; Monte Carlo methods; artificial neural network; dose optimization data analysis; dosimetry; high-energy photon spectrum; linear accellerator;
Conference_Titel :
Intelligent Systems, Modelling and Simulation (ISMS), 2010 International Conference on
Conference_Location :
Liverpool
Print_ISBN :
978-1-4244-5984-1
DOI :
10.1109/ISMS.2010.27