The Gamma Knife Check software is an FDA approved second check system for dose calculations in Gamma Knife radiosurgery. The purpose of this study was to evaluate the accuracy and the stability of the commercial software package as a tool for independent dose verification. The Gamma Knife Check software version 8.4 was commissioned for a Leksell Gamma Knife Perfexion and a 4C unit at the University of Pittsburgh Medical Center in May 2012. Independent dose verifications were performed using this software for 319 radiosurgery cases on the Perfexion and 283 radiosurgery cases on the 4C units. The cases on each machine were divided into groups according to their diagnoses, and an averaged absolute percent dose difference for each group was calculated. The percentage dose difference for each treatment target was obtained as the relative difference between the Gamma Knife Check dose and the dose from the tissue maximum ratio algorithm (TMR 10) from the GammaPlan software version 10 at the reference point. For treatment plans with imaging skull definition, results obtained from the Gamma Knife Check software using the measurement-based skull definition method are used for comparison. The collected dose difference data were also analyzed in terms of the distance from the treatment target to the skull, the number of treatment shots used for the target, and the gamma angles of the treatment shots. The averaged percent dose differences between the Gamma Knife Check software and the GammaPlan treatment planning system are 0.3%, 0.89%, 1.24%, 1.09%, 0.83%, 0.55%, 0.33%, and 1.49% for the trigeminal neuralgia, acoustic neuroma, arteriovenous malformation (AVM), meningioma, pituitary adenoma, glioma, functional disorders, and metastasis cases on the Perfexion unit. The corresponding averaged percent dose differences for the 4C unit are 0.33%, 1.2%, 2.78% 1.99%, 1.4%, 1.92%, 0.62%, and 1.51%, respectively. The dose difference is, in general, larger for treatment targets in the peripheral regions of the skull owing to the difference in the numerical methods used for skull shape simulation in the GammaPlan and the Gamma Knife Check software. Larger than 5% dose differences were observed on both machines for certain targets close to patient skull surface and for certain targets in the lower half of the brain on the Perfexion, especially when shots with 70 and/or 110 gamma angles are used. Out of the 1065 treatment targets studied, a 5% cutoff criterion cannot always be met for the dose differences between the studied versions of the Gamma Knife Check software and the planning system for 40 treatment targets.
© 2016 The Authors.