A novel couch-gantry trajectory based stereotactic treatment method

We propose a robust inverse planning solution for trajectory-based stereotactic (SRS) treatment of small lesions in the brain with a conventional c-arm linear accelerator (lin- ac). The linac’s beam trajectory is formed by simultaneously rotating the gantry and the treatment couch, thus enabling a wide selection of beam angles in a time-efficient manner. Dose is optimized by dynamically varying the dose rate and the multi-leaf collimators’ (MLC) leaf positions along the beam trajectory. The pre-defined beam trajectory is formed by the couch sweeping left to right (180°) while the gantry produces partial arcs in a sinusoidal pattern. An optimized beam intensity pattern is formed along this trajectory while the MLC aperture changes continuously. The optimization adds controls points using the progressive sampling algorithm and varies each control points’ MLC positions and dose rate with stochastic perturbations, keeping only the perturbations that lower the cost function. This method puts increased importance on dose rate modulation since the default trajectory samples the entire phase-space and unwanted portions of the trajectory need to be removed. Therefore, our method emphasizes beam weight perturbation and we are able to expedite the optimization by adding momentum to these perturbations. When applied to select patients, we observed a reduction of the mean dose to the organs at risk (OAR) and the doses to the planning target volume (PTV) were comparable for both cases. In addition, normal tissue dose was improved for both cases. Delivery time for a 15Gy fraction at the isocenter was approximately 8 minutes. The method we propose is viable on current conventional linacs and produces high quality, reproducible treatment plans for SRS patients.