Impacts of Calculation Grid Spacing and Statistical Uncertainty of Monte Carlo Algorithm on Stereotactic Radiosurgery Planning With Volumetric-Modulated Arcs for Single Brain Metastases Using the Monaco® System.

Purpose In linac-based stereotactic radiosurgery (SRS) utilizing a multileaf collimator (MLC) for brain metastases (BMs), a volumetric-modulated arc (VMA) technique is indispensable for generating a suitable dose distribution with efficient planning and delivery. However, the optimal calculation grid spacing (GS) and statistical uncertainty (SU) of the Monte Carlo algorithm for VMA optimization have yet to be determined. This planning study aimed to examine the impacts of GS and GU settings on VMA-based SRS planning and to find the optimal combination for templating. Materials and methods Thirty clinical BMs with a gross tumor volume (GTV) of 0.08-48.09 cc (median 9.81 cc) were included. The treatment platform included a 5-mm leaf-width MLC Agility (Elekta AB, Stockholm, Sweden) and a planning system Monaco (Elekta AB). The prescribed dose was uniformly assigned to the GTV , the minimum dose of GTV minus 0.01 cc, i.e.,  for GTV >0.20 cc or to the GTV for GTV ≤0.20 cc, to minimize the uncovered GTV to the equivalent of a 3 mm diameter lesion. Five combinations of GS and SU per plan were examined for 12 selected GTVs (median 17.41 cc): GS of 2 mm and SU of 3% (G2U3), 2 mm and 2% (G2U2), 2 mm and 1% (G2U1), 1 mm and 2% (G1U2), and 1 mm and 1% (G1U1). Otherwise, the same arc arrangement and optimization method were uniformly used to prioritize the GTV dose conformity and the steepness of the dose gradient outside the GTV without dose constraints inside the GTV boundary. Further comparisons were conducted using 30 GTVs between the two groups with the highest plan quality. Results The G2U3 and G2U2 resulted in the equivalent total calculation time (tCT) and exactly the same plan quality. The overall plan quality was significantly superior in the G1U2 and G1U1 than in the G2U1 and G2U2, although the tCT was significantly longer in the G1U1 and G1U2 than in the G2U1 and G2U2. In the comparison of the G1U2 and G1U1, the concentric lamellarity of dose gradients 2 mm outside and 2-4 mm inside the GTV boundary was significantly superior in the G1U1 than in the G1U2, while there was no significant difference in the other parameters. The tCT tended to be longer in the G1U1 than in the G1U2. Conclusions The initial settings of GS and SU have significant impacts on the plan quality and tCT. The settings with GS of 1 mm and SU of 1% per plan are recommended to create the most suitable dose distribution for single BMs, especially for irregularly shaped and/or large lesions, although the tCT is long. In addition to common evaluation metrics, the coverage values of 2 mm outside and 2-4 mm inside the GTV surface by the , the minimum dose to cover the irradiated isodose volume equivalent to each target volume, are valuable for in-depth plan comparison.