Implementation and experimental validation of the high dose rate stereotactic treatment mode at Varian accelerators.

Abstract

Background. Treating small tumours in the brain requires usage of small fields. However, it is a challenge for all treatment planning systems (TPS) to modulate small fields (≤3×3 cm²) for which lateral electronic equilibrium does not exits. Material and methods. Open field beam data was measured for the stereotactic mode at our Varian Trilogy accelerator and was implemented in the Eclipse TPS for both the Pencil Beam (PB) and the Anisotropic Analytical Algorithm (AAA). Additional output factors, profiles and percentage depth dose curves were measured in a water phantom for asymmetric fields and fields defined by the Varian 120 MLC and the BrainLAB m3 MLC. Dose distributions in transverse sections were measured with gafchromic films in a homogeneous phantom. These were compared to those calculated by the Eclipse TPS for 11 stereotactic treatment plans. The plans included non-coplanar, dynamic and arc fields. All comparisons for the film measurements in the phantom were made with a Gamma criterion of 2 mm in distance and 3% in dose. Results. The AAA and the PB algorithms broadened the penumbra of the profiles by approximately 1.5 mm and 2.0 mm, respectively. For static fields defined by the Varian 120 MLC, the measured field size was 1.0 mm broader than the calculated. The output factors were typically within 2% for field sizes ≥ 2×2 cm². More than 95% of the points passed the Gamma criterion. Up to 50% of the points in the low dose regions failed the criterion, but overall the match was satisfactory. Conclusion. The accuracy of calculations performed by the Eclipse TPS for the SRS mode was shown to be within a few percent for most clinical fields. AAA was found to be superior to the PB algorithm in modelling the dose in the penumbra region.