KEY POINTS
- Investigators performed a paired planning comparison in 105 patients, generating scenarios containing 1, 5, 10, 15 and 20 brain metastases from the same underlying patient geometry. Every plan prescribed 27 Gy in three fractions with a target goal of 98% PTV coverage.
- Six delivery strategies were evaluated: CyberKnife, ring-gantry helical radiotherapy, coplanar single- and dual-isocenter VMAT, and non-coplanar single- and dual-isocenter VMAT. Primary endpoints included normal-brain V5Gy, V10Gy, V12Gy, V15Gy and V20Gy, alongside hippocampal dose and treatment efficiency.
- Low-dose exposure increased with lesion number on every platform, but not at the same rate. At 10 lesions, V12Gy was 196 cc with coplanar single-isocenter VMAT, compared with approximately 50–59 cc for the other VMAT geometries, 32 cc with CyberKnife and 62 cc with ring-gantry delivery.
- The differences became particularly large at 20 lesions. V12Gy was approximately 312–376 cc across the four VMAT strategies, 172 cc with CyberKnife and 522 cc with the ring-gantry technique. Corresponding V10Gy was approximately 461–554 cc with VMAT, 285 cc with CyberKnife and 783 cc with ring-gantry delivery.
- Within C-arm VMAT, geometry mattered substantially. Coplanar single-isocenter delivery produced disproportionately high intermediate-dose spread at several lesion burdens, whereas introducing non-coplanar beams or a second isocenter often reduced spillage by diversifying beam paths and reducing excessive off-axis geometry.
- Hippocampal mean and maximum dose also increased as lesion burden rose. CyberKnife showed a more gradual increase, while C-arm and ring-gantry techniques increased more steeply; across platforms, hippocampal dose was strongly influenced by how close the metastases were to the hippocampi, not merely by lesion count.
- There was no universally superior platform. CyberKnife generally achieved better low-dose containment but had the longest beam-on times, ring-gantry treatment offered intermediate efficiency, and single-isocenter VMAT was generally faster but could increase low-dose spread. Phantom measurements of representative 20-lesion plans passed 3%/2-mm verification, with low-dose measurements within ±3% of treatment-planning-system calculations.
CLINICAL TAKEAWAY
For one or a few metastases, platform differences may be modest; at 10–20 lesions, the choice of delivery architecture can materially alter the volume of normal brain receiving low and intermediate doses. Treatment speed and normal-brain sparing pull in opposite directions, so technique selection should account for lesion number, spatial distribution and proximity to critical structures rather than relying on a single preferred platform. This was a planning study using platform-specific planning systems and has no neurocognitive or clinical-outcome data.