Determining the energy spectrum of clinical linear accelerator using an optimized photon beam transmission protocol

Purpose The complex beam delivery techniques for patient treatment using a clinical linear accelerator (linac) may result in variations in the photon spectra, which can lead to dosimetric differences in patients that cannot be accounted for by current treatment planning systems (TPSs). Therefore, precise knowledge of the fluence and energy spectrum (ES) of the therapeutic beam is very important. However, owing to the high energy and flux of the beam, the ES cannot be measured directly, and validation of the spectrum modeled in the TPS is difficult. The aim of this study is to develop an efficient beam transmission measurement procedure for accurately reconstructing the ES of a therapeutic x‐ray beam generated by a clinical linac. Methods The attenuation of a 6 MV photon beam from an Elekta Synergy Platform clinical linac through different thicknesses of graphite and lead was measured using an ion chamber. The response of the ion chamber as a function of photon energy was obtained using the Monte Carlo (MC) method in the Geant4 simulation code. Using the curves obtained in the photon beam transmission measurements and the ion chamber energy response, the ES was reconstructed using an iterative algorithm based on a mathematical model of the spectrum. To evaluate the accuracy of the spectrum reconstruction method, the reconstructed ES (ESrecon) was compared to that determined by the MC simulation (ESMC). Results The ion chamber model in the Geant4 simulation was well validated by comparing the ion chamber perturbation factors determined by the TRS‐398 calibration protocol and EGSnrc; the differences were within 0.57%. The number of transmission measurements was optimized to 10 for efficient spectrum reconstruction according to the rate of increase in the spectrum reconstruction accuracy. The distribution of ESrecon obtained using the measured transmission curves was clearly similar to the reference, ESMC, and the dose distributions in water calculated using ESrecon and ESMC were similar within a 2% local difference. However, in a heterogeneous medium, the dose discrepancy between them was >5% when a complex beam delivery technique composed of 171 control points was used. Conclusions The proposed measurement procedure required a total time of approximately 1 h to obtain and analyze 20 transmission measurements. In addition, it was confirmed that the transmission curve of high‐Z materials influences the accuracy of spectrum reconstruction more than that of low‐