A Feasibility Study to Use Gamma-Histogram Analysis in Assisting Quality Assurance Criteria for Evaluating Volumetric Modulated Arc Therapy Treatment Plans
K Hisey*, M Morales-Paliza, G Ding, Vanderbilt University, Nashville, TNSU-E-T-168 Sunday 3:00PM - 6:00PM Room: Exhibit Hall
Purpose: To investigate if gamma-histogram statistical analysis can be used as criteria in establishing clinical action limits for pre-treatment quality assurance (QA) agreement between planned and delivered dose distributions in volumetric-modulated-arc therapy (VMAT).
Methods: Radiotherapy patient-treatment plans were calculated using AAA dose-calculation algorithm implemented in Varian Eclipse system (v.10). 6 and 10 MV photon beams from a Varian-Trilogy Linac were investigated. Pre-treatment QA dosimetry on 102 VMAT plans was performed using an electronic-portal-imaging device (EPID) and an ion-chamber-based array (MatriXX). Evaluation was done by comparing planned and measured dose distributions using gamma-histogram analysis for plans grouped by site (head and neck, prostate, brain and lung). Different dose-difference/distance-to-agreement gamma criteria from 3%/1mm to 5%/3mm were evaluated considering only dose levels higher than 10% of the maximum dose. Statistical analysis of gamma histograms included mean γ values (γmean), γ values corresponding to Δ = mean + 1.5 SD (γΔ), and areal percentages with gamma values less than 1.0, 1.5, and greater than 2.0. Average values and standard deviations for each of these five parameters were calculated per site to determine specific action limits.
Results: The γmean over all analyzed plans with EPID was 0.42 +/- 0.10 for 3%/1mm and 0.29+/- 0.06 for 3%/3mm, while the corresponding average percentage of points with gamma less than 1 was 95.0% +/- 5.0 and 99.1% +/- 1.6, respectively. The results showed that parameters under the 3%/1mm and 3%/3mm criteria are plan-site specific while the 4%/3mm and 5%/3mm criteria did not provide statistically significant differences among sites for EPID. Similar trends were observed for analysis using the MatriXX.
Conclusion: Statistical analysis of γ histogram parameters is useful to perform pre-treatment QA for VMAT plans. A specific selected criterion can be used to establish clinical action levels based on clinical acceptability and machine delivery accuracy.