Respiratory gated beam delivery cannot facilitate margin reduction, unless combined with respiratory correlated image guidance.

PURPOSE/
OBJECTIVE: In radiotherapy of targets moving with respiration, beam gating is offered as a means of reducing the target motion. The purpose of this study is to evaluate the safe magnitude of margin reduction for respiratory gated beam delivery. MATERIALS/
METHODS: The study is based on data for 17 lung cancer patients in separate protocols at Rigshospitalet and Stanford Cancer Center. Respiratory curves for external optical markers and implanted fiducials were collected using equipment based on the RPM system (Varian Medical Systems). A total of 861 respiratory curves represented external measurements over 30 fraction treatment courses for 10 patients, and synchronous external/internal measurements in single sessions for seven patients. Variations in respiratory amplitude (simulated coaching) and external/internal phase shifts were simulated by perturbation with realistic values. Variations were described by medians and standard deviations (SDs) of position distributions of the markers. Gating windows (35% duty cycle) were retrospectively applied to the respiratory data for each session, mimicking the use of commercially available gating systems. Medians and SDs of gated data were compared to those of ungated data, to assess potential margin reductions.
RESULTS: External respiratory data collected over entire treatment courses showed SDs from 1.6 to 8.1mm, the major part arising from baseline variations. The gated data had SDs from 1.5 to 7.7mm, with a mean reduction of 0.3mm (6%). Gated distributions were more skewed than ungated, and in a few cases a marginal miss of gated respiration would be found even if no margin reduction was applied. Regularization of breathing amplitude to simulate coaching did not alter these results significantly. Simulation of varying phase shifts between internal and external respiratory signals showed that the SDs of gated distributions were the same as for the ungated or smaller, but the median values were markedly shifted. The gated distributions could generally not be covered by margins derived from ungated data, if the phase shift was not accounted for.
CONCLUSIONS: Margins can only be reduced for respiratory gated radiotherapy, if respiratory baseline shifts and variations in external/internal motion correlation are accounted for. Gated beam delivery alone cannot facilitate margin reduction. In the worst case, margins must be increased to accommodate inter-fraction variations in respiration.