Validating That the Gamma Dose Distribution Comparison Defaults to the Distance-To-Agreement (DTA) Test in Steep Dose Gradients for Clinical Dose Distributions
T Dou*, D Morele, D Low, Department of Radiation Oncology, University of California, Los Angeles, Los Angeles, CASU-E-T-20 Sunday 3:00PM - 6:00PM Room: Exhibit Hall
Purpose: To test the validity of the central assumption that the gamma dose comparison test defaults to a distance-to-agreement (DTA) test in steep dose gradient regions of clinical dose distributions for typically employed dose difference and DTA criteria.
Methods: The angle θ between the dose axis and the gamma vector was computed using
θ = tan⁻¹(Λ), where Λ was the local dose gradient of the evaluated dose distribution. θ depended only on the ratio α between the two test criteria and its computation required only one dose distribution. While the most common test criteria were 3%/3mm (α = 1%/mm), we also evaluated 2%/3mm (α = 0.67%/mm) and 3%/2mm (α = 1.5%/mm). Clinical treatment plans were used to calculate θ using the three values of α. 50 cases of both prostate and head and neck and 27 cases for lung cancer were examined. The planning target volumes and selected critical structures were examined. The maximum values of θ were determined for these structures to determine if these maximum θ angles were close to 90 degrees for the selected α values.
Results: Most examined targets and critical structures exhibited large values of θ. The exception was for some parotid glands where the doses were very low and relatively homogeneous. The fact that we found that most structures had at least some voxels with sufficiently large dose gradients to yield θ values near 90 degrees validated our hypothesis that for steep gradients, the gamma test defaulted to distance-to-agreement test for clinical dose gradients.
Conclusion: The implicit assumption that the gamma dose comparison tool defaults to the DTA test in steep dose gradient regions was validated for prostate, had-and-neck, and lung cancer sites. While dose distributions from other sites have not been examined, it is likely that they too will yield similar results.
Add this talk to vcal | ical | Contact Email: