Volume-based algorithm of lung dose optimization in novel dynamic arc radiotherapy for esophageal cancer

This study aims to develop a volume-based algorithm (VBA) that can rapidly optimize rotating gantry arc angles and predict the lung V₅ preceding the treatment planning. This phantom study was performed in the dynamic arc therapy planning systems for an esophageal cancer model. The angle of rotation of the gantry around the isocenter as defined as arc angle (θ_A), ranging from 360° to 80° with an interval of 20°, resulting in 15 different θ_A of treatment plans. The corresponding predicted lung V₅ was calculated by the VBA, the mean lung dose, lung V₅, lung V₂₀, mean heart dose, heart V₃₀, the spinal cord maximum dose and conformity index were assessed from dose-volume histogram in the treatment plan. Correlations between the predicted lung V₅ and the dosimetric indices were evaluated using Pearson's correlation coefficient. The results showed that the predicted lung V₅ and the lung V₅ in the treatment plan were positively correlated (r = 0.996, p < 0.001). As the θ_A decreased, lung V₅, lung V₂₀, and the mean lung dose decreased while the mean heart dose, V₃₀ and the spinal cord maximum dose increased. The V₂₀ and the mean lung dose also showed high correlations with the predicted lung V₅ (r = 0.974, 0.999, p < 0.001). This study successfully developed an efficient VBA to rapidly calculate the θ_A to predict the lung V₅ and reduce the lung dose, with potentials to improve the current clinical practice of dynamic arc radiotherapy.