Reconstruction of the electron source intensity distribution of a clinical linear accelerator using in-air measurements and a genetic algorithm

BACKGROUND AND PURPOSE: The electron source intensity distribution of a clinical linear accelerator has a great influence on the calculation of output factors for small radiation fields where source occlusion by the collimating devices takes place. The purpose of this study was to present a new method for the electron source reconstruction problem. MATERIALS AND METHODS: The measurements were performed in-air using diode and 6 MV 1 × 1 cm(2) photon field in flattening filter-free mode. In Monte Carlo simulation, an electron target area was divided into a number of square subsources. Then, the in-air doses in 2D silicon chip array were calculated individually from each subsource. A genetic algorithm search was applied in order to determine the optimal weight factors for all subsources that provide the best agreement between simulated and measured doses. RESULTS: It was found that the reconstructed electron source intensity from a clinical linear accelerator has the two-dimensional elliptical double Gaussian distribution. The source intensity distribution consisted of two intensity components along the in-plane (x) and cross-plane (y) directions characterized by full width half-maximum (FWHM): FWHM(x1) = 0.27 cm, FWHM(x2) = 0.08 cm, FWHM(y1) = 0.24 cm, FWHM(y2) = 0.06 cm, where broader components are 81% and 53% of the total intensity along × and y axis respectively. CONCLUSIONS: The obtained results demonstrated an elliptical double Gaussian intensity distribution of the incident electron source. We anticipate that the proposed method has universal applications independent of the type of linear accelerator, modality or energy.