This study investigated small-field dosimetry for a Cyberknife stereotactic radiosurgery system using Monte Carlo simulations. The EGSnrc/BEAMnrc Monte Carlo code was used to simulate the Cyberknife treatment head, and the DOSXYZnrc code was implemented to calculate central axis depth-dose curves, off-axis dose profiles, and relative output factors for various circular collimator sizes of 5 to 60 mm. Water-to-air stopping power ratios necessary for clinical reference dosimetry of the Cyberknife system were also evaluated by Monte Carlo simulations. Additionally, a beam quality conversion factor, kQ, for the Cyberknife system was evaluated for cylindrical ion chambers with different wall material. The accuracy of the simulated beam was validated by agreement within 2% between the Monte Carlo calculated and measured central axis depth-dose curves and off-axis dose profiles. The calculated output factors were compared with those measured by a diode detector and an ion chamber in water. The diode output factors agreed within 1% with the calculated values down to a 10 mm collimator. The output factors with the ion chamber decreased rapidly for collimators below 20 mm. These results were confirmed by the comparison to those from Monte Carlo methods with voxel sizes and materials corresponding to both detectors. It was demonstrated that the discrepancy in the 5 and 7.5 mm collimators for the diode detector is due to the water non-equivalence of the silicon material, and the dose fall-off for the ion chamber is due to its large active volume against collimators below 20 mm. The calculated stopping power ratios of the 60 mm collimator from the Cyberknife system (without a flattening filter) agreed within 0.2% with those of a 10 X 10 cm2 field from a conventional linear accelerator with a heavy flattening filter and the incident electron energy, 6 MeV. The difference in the stopping power ratios between 5 and 60 mm collimators was within 0.5% at a 10 cm depth in water. Furthermore, kQ values for the Cyberknife system were in agreement within 0.3% with those of the conventional 6 MV-linear accelerator for the cylindrical ion chambers with different wall material.

A radiophotoluminescent (RPL) glass rod dosimeter (GRD) and a small active volume p-type silicon diode detector are used for the measurement of the output factors from Gamma-Knife fields. The GRD system consists of small rod-shaped glass chip detectors and an automatic readout device. The output factors measured with the GRD from the 14, 8 and 4 mm helmets relative to the 18 mm helmet are 0.981, 0.942 and 0.877, respectively. Similarly, the corresponding output factors measured with the p-type silicon diode detector are 0.980, 0.949 and 0.867, respectively. The output factors are corrected for the end effect for each helmet. The output factors obtained from both detectors are in good agreement with the values in a recent publication and the values recommended by Elekta, the manufacturer. The directional dependence of these detectors is also measured. For the Gamma-Knife angle ranging from 6 to 36 degrees in the y-z plane of the stereotactic space, the measured angular dependence of the GRD is approximately 1.0% at a 4 MV x-ray beam. The response of the silicon diode detector indicates approximately 3-4% directional dependence for the same angular range for a 6 MV x-ray beam. The Gamma-Knife helmet output factors measured with the silicon diode detector are corrected for angular dependence.

A fully automatic radiophotoluminescent glass rod dosimeter (GRD) system has recently become commercially available. This article discusses the dosimetric properties of the GRD including uniformity and reproducibility of signal, dose linearity, and energy and directional dependence in high-energy photon beams. In addition, energy response is measured in electron beams. The uniformity and reproducibility of the signal from 50 GRDs using a 60Co beam are both +/- 1.1% (one standard deviation). Good dose linearity of the GRD is maintained for doses ranging from 0.5 to 30 Gy, the lower and upper limits of this study, respectively. The GRD response is found to show little energy dependence in photon energies of a 60Co beam, 4 MV (TPR20(10)=0.617) and 10 MV (TPR(20)10=0.744) x-ray beams. However, the GRD responses for 9 MeV (mean energy, Ez = 3.6 MeV) and 16 MeV (Ez = 10.4 MeV) electron beams are 4%-5% lower than that for a 60Co beam in the beam quality dependence. The measured angular dependence of GRD, ranging from 0 degrees (along the long axis of GRD) to 120 degrees is within 1.5% for a 4 MV x-ray beam. As applications, a linear accelerator-based radiosurgery system and Cyber-Knife output factors are measured by a GRD and compared with those from various detectors including a p-type silicon diode detector, a diamond detector, and an ion chamber. It is found that the GRD is a very useful detector for small field dosimetry, in particular, below 10 mm circular fields.