Physics considerations for single-isocenter, volumetric modulated arc radiosurgery for treatment of multiple intracranial targets

Objective: Our purpose was to address challenges associated with single-isocenter radiosurgery for multiple intracranial targets (SIRMIT) including increased sensitivity to rotational uncertainties (resulting from distance of the targets from isocenter) as well as potential for decreased plan quality from larger multileaf collimator width >. 4 cm from isocenter. Methods and materials: We evaluated the effect that a 6 degrees-of-freedom couch correction had on localization uncertainty for SIRMIT using thermoplastic mask immobilization. Required setup margin was determined from rotation of the skull and mask (setup kV cone beam computed tomography relative to planning computed tomography). Intraoperational margin was determined from skull rotation within the mask (difference between pre- and posttreatment cone beam computed tomography). We also investigated 4 isocenter placement strategies: volume centroid, centroid of equally weighted points (1 per target), centroid of points weighted by inverse of volume, and Eclipse's built-in method. Results: When no 6 degrees-of-freedom couch correction is performed after initial setup, a 0.35-mm margin is required per centimeter of target-isocenter separation to account for 95% of rotational uncertainties at initial setup. This margin is reduced to 0.10 mm/cm of target-isocenter separation to account for intraoperative rotational uncertainties when the initial setup uncertainty is eliminated via image guided 6 degrees-of-freedom couch correction. Analysis of 11 multitarget plans (37 targets) showed that conformity index and gradient index improved with decreasing distance from isocenter, this trend being more pronounced for targets <. 1 mL. Alternative isocenters aimed at decreasing distance of small targets improved their gradient index, but resulted in poorer dose indices for large targets. Mean distance from isocenter was smallest for the centroid of equally weighted points (4.1 ± 1.6cm vs 4.2-4.5cm). Conclusions: Rotational corrections via image guidance are necessary for SIRMIT with a thermoplastic mask for immobilization. There is a clear tradeoff between dosimetric quality of small and large targets that should be considered carefully when placing the isocenter.