Evaluation of single-isocenter multiple-target IMRT using 1.5 T Elekta Unity MR-guided linear accelerator for stereotactic radiosurgery of multiple brain metastases

Purpose: To assess the dosimetric performance of using the 1.5 T Elekta Unity magnetic resonance (MR)-guided linear accelerator (MR-Linac) for intracranial single-isocenter, multiple-target stereotactic radiosurgery (SRS). Materials and methods: A total of 33 consecutive patients, each with 2-7 brain metastases (BMs), who had undergone SRS using a conventional Linac with single-isocenter, multiple-target, non-coplanar volumetric modulated arc therapy (NC-VMAT), were retrospectively chosen for this study. For each patient, a new treatment plan was created using intensity-modulated radiotherapy (IMRT) on the MR-Linac (MR-Linac-IMRT) with 13-15 coplanar beams. A 2 mm planning target volume expansion was applied for the MR-Linac-IMRT plans, consistent with the expansion used in the original NC-VMAT plans. The two plans were compared regarding homogeneity index (HI), gradient index (GI), normal brain tissue (NBT) dose exposure, conformity index (CI), monitor units (MUs), beam-on time (BOT) and the absolute volumes receiving = X% (VX%) of the prescription dose. Results: Overall, the MR-Linac-IMRT plans produced comparable result to the NC-VMAT plans regarding CI, but with a significantly higher GI (6.59 vs. 5.10, p < 0.001) and a lower HI (1.30 vs. 1.33, p = 0.046). The V10%-V75% for MR-Linac-IMRT were significantly larger than those for NC-VMAT (p = 0.001), while NC-VMAT exhibited higher V5% values. Regarding NBT sparing, the MR-Linac-IMRT plan resulted in a larger V12Gy (31.27 vs. 25.29 cm³, p < 0.001) but delivered a comparable mean dose to the NBT. While the MR-Linac-IMRT plan required less MUs compared to NC-VMAT, it had a significantly longer BOT (20.47 vs. 6.32 min, p < 0.001) due to the lower dose rate. Conclusion: For SRS of multiple BMs, MR-Linac-IMRT does not provide significant dosimetric advantages over conventional Linac-based NC-VMAT, but needs more BOT. As a result, MR-Linac-IMRT may not be the optimal choice for intracranial single-isocenter, multiple-target SRS