Validation of Varian's AAA algorithm with focus on lung treatments

Area of Science:

• Medical Physics
• Radiation Oncology
• Dosimetry

Background:

• The Anisotropic Analytical Algorithm (AAA) is a pivotal tool in radiation therapy planning.
• Evaluating AAA accuracy is crucial for optimizing treatment delivery, especially in complex geometries.

Purpose of the Study:

• To rigorously assess the accuracy of the Anisotropic Analytical Algorithm (AAA).
• To compare AAA performance against the Pencil Beam Convolution (PBC) algorithm.
• To investigate AAA accuracy across various field configurations and media, including homogeneous and inhomogeneous (lung) environments.

Main Methods:

• Measured 2D dose distributions using ion chambers in solid water and a Thorax lung phantom with Gafchromic films.
• Evaluated dose distributions using a gamma criterion of 3% dose and 3 mm distance to agreement (DTA).
• Compared AAA calculations with experimental measurements and PBC algorithm results.

Main Results:

• AAA met criteria in homogeneous media, with minor deviations (4-5%) for specific fields (wedges, intensity-modulated plans).
• AAA showed good agreement in the Thorax phantom for 6 and 15 MV photons, with over 80% of points meeting the gamma criterion in high-dose regions.
• PBC algorithm achieved only 30-70% compliance with the gamma criterion, indicating significantly lower accuracy.

Conclusions:

• AAA is demonstrably superior to PBC in heterogeneous media, particularly at 15 MV.
• AAA generally provides accurate dose calculations (<3% deviation) in lung and mediastinum regions for most treatment plans.
• Potential underestimation of spinal cord dose by up to 7% necessitates careful clinical consideration.