Predictive factors for lung dose reduction by respiratory gating at radiotherapy for lung cancer.

This study aimed to identify factors predictive of the benefit of respiratory-gated radiotherapy. Three plans were created for 25 patients with non-small cell lung cancer, simulating the following 3 treatment scenarios. Protocol 1 was non-gated and the lung dose was calculated using 4-s slow CT (PnA), protocol 2 was also non-gated and the lung dose was calculated by CT at the end-expiration phase (PnE), and protocol 3 applied phase-based gating around end-expiration (PgE). We correlated possible predictive factors with the estimated lung dose reduction achieved by respiratory gating. The 3D clinical target volume (CTV) motion, craniocaudal CTV motion, and the craniocaudal CTV position were correlated with the reduction in V20 and the mean lung dose (p < 0.01). CTV was not significantly correlated with the estimated lung dose reduction. The area under the ROC curve (AUC) for 3D- and craniocaudal CTV motion, and craniocaudal CTV position was 1.000, 0.997, and 0.943, respectively, when the threshold for selecting patients was set at a 1% reduction of V20 and at a 0.5 Gy reduction in the mean lung dose. The results of the present study suggest that 3D CTV motion, craniocaudal CTV motion, and the craniocaudal CTV position are useful for predicting the benefit of respiratory-gated radiotherapy in lung cancer patients.