PurposeTo evaluate cross-irradiation among targets in multiple isocenter (multiISO) plans with simultaneous optimization for multiple brain metastases in frameless stereotactic radiotherapy. MethodsThree spherical targets separated by 3 cm were drawn on a brain CT image set. Volumetric modulated arc therapy (VMAT) plans consisting of single or two-arc beams were created using the multiISO method, in which partial arcs and selective couch angles were included to avoid cross-irradiation among the targets. Plans were evaluated using conformity index, dose volumes, and beam weight from cross-irradiation. Single isocenter (singleISO) plans were also produced and compared. The entire simulation was repeated for targets separated by 5 cm. A single patient case with two targets was also simulated for application. ResultsFor target distance of 3 cm, both the multiISO and singleISO plans with two beams were equivalent and favored based on the conformity index and dose volumes. The cross-irradiation was considerable and was not reduced effectively by partial arc or selective couch angles. With a 5- cm target distance, multiISO two-arc plans made of full or partial arc and selective couch angles were equivalent and adequate for treatment. The cross-irradiation was small and made negligible by partial arc or selective couch angles. The two-arc singleISO plan was acceptable but slightly inferior with respect to conformity index and low-dose volumes. ConclusionsMultiISO radiation treatment for multiple brain metastases was prone to cross-irradiation among targets, which led to target undercoverage from potential rotational error, though it was less than with the singleISO method. For the multiISO VMAT plan to be robust to rotational error, the plan should be designed with beams minimizing cross irradiation by incorporating a partial arc or selective couch angle to avoid irradiating neighboring targets, which is more effective for targets with greater separation.