Linac-based VMAT radiosurgery for multiple brain lesions: comparison between a conventional multi-isocenter approach and a new dedicated mono-isocenter technique

Linac-based stereotactic radiosurgery or fractionated stereotactic radiotherapy (SRS/FSRT) of multiple brain lesions using volumetric modulated arc therapy (VMAT) is typically performed by a multiple-isocenter approach, i.e. one isocenter per lesion, which is time-demanding for the need of independent setup verifications of each isocenter. Here, we present our initial experience with a new dedicated mono-isocenter technique with multiple non-coplanar arcs (HyperArc™, Varian Inc.) in terms of a plan comparison with a multiple-isocenter VMAT approach.

From August 2017 to October 2017, 20 patients with multiple brain metastases (mean 5, range 2–10) have been treated by HyperArc in 1–3 fractions. The prescribed doses (Dp) were 18–25 Gy in single-fraction, and 21–27 Gy in three-fractions. Planning Target Volume (PTV), defined by a 2 mm isotropic margin from each lesion, had mean dimension of 9.6 cm³ (range 0.5–27.9 cm³). Mono-isocenter HyperArc VMAT plans (HA) with 5 non-coplanar 180°-arcs (couch at 0°, ±45°, ±90°) were generated and compared to multiple-isocenter VMAT plans (RA) with 2 coplanar 360°-arcs per isocenter. A dose normalization of 100%Dp at 98%PTV was adopted, while D_2%(PTV) < 150%D_p was accepted. All plans had to respect the constraints on maximum dose to the brainstem (D_0.5cm3 < 18 Gy) as well as to the optical nerves/chiasm, eyes and lenses (D_0.5cm3 < 15 Gy). HA and RA plans were compared in terms of dose-volume metrics, by Paddick conformity (CI) and gradient (GI) index and by V₁₂ and mean dose to the brain-minus-PTV, and in terms of MU and overall treatment time (OTT) per fraction. OTT was measured for HA treatments, whereas for RA plans OTT was estimated by assuming 3 min. For initial patient setup plus 5 min. For each CBCT-guided setup correction per isocenter.

Significant variations in favour of HA plans were computed for both target dose indexes, CI (p < .01) and GI (p < .01). The lower GI in HA plans was the likely cause of the significant reduction in V₁₂ to the brain-minus-PTV (p = .023). Although at low doses, below 2–5 Gy, the sparing of the brain-minus-PTV was in favour of RA plans, no significant difference in terms of mean doses to the brain-minus-PTV was observed between the two groups (p = .31). Finally, both MU (p < .01) and OTT (p < .01) were significantly reduced by HyperArc plans.

For linac-based SRS/FSRT of multiple brain lesions, HyperArc plans assured a higher CI and a lower GI than standard multiple-isocenter VMAT plans. This is consistent with the computed reduction in V₁₂ to the brain-minus-PTV. Finally, HyperArc treatments were completed within a typical 20 min. time slot, with a significant time reduction with respect to the expected duration of multiple-isocenters VMAT.