Monte Carlo Investigation into Effect of Setup and Motion Uncertainties on Dosimetry of Trigeminal Neuralgia Treatments with Virtual Cone SRS
Abstract
Purpose
Estimate impact of setup and motion uncertainties on dosimetry of SRS with Virtual Cone (VC) and evaluate a simple method to account for such uncertainties in clinical dose calculations.
Methods
SRS treatment of trigeminal neuralgia using Virtual Cone (proposed by University of Alabama, UAB) involves multiple noncoplanar arcs and collimator rotations using HD MLCs with two leaf pairs opened during the treatment to create a fixed gap. The gap sizes of 1.4 – 2.6 millimeters were reported. Vancouver Island Monte Carlo (MC) system was used to model such treatments. Voxel size of 1mm was used as most common dose resolution for such calculations, and 0.5% statistical uncertainty was used. MC calculations were also re-run with 1.25, 1.5 and 2mm voxels. Fluence convolution method was used to model setup uncertainties.
Results
For the treatments with 1.4mm leaf gap, setup uncertainties with 0.5 and 1mm standard deviations (SD) reduced the values of maximum dose (Dmax) by 10.5% and 26% respectively. For the treatments with leaf gap of 2.1mm, Dmax reduction was 8% and 26%, and for the leaf gap of 2.6mm Dmax reduced by 7% and 22%. MC calculations using 1.5mm voxel size (no setup uncertainty) agreed with the dose from modeling 0.5% SD setup uncertainty, reducing Dmax differences to less than 1.5% for all studied leaf gaps. The dose profiles also agreed within less than 0.2mm below 90% isodose. Similarly, using calculations with 2mm voxel size (no setup uncertainty) reproduced 1mm setup uncertainty within 3% at Dmax.
Conclusion
Setup uncertainties can have considerable effect on dosimetry of VC SRS. Larger voxels sizes such as 1.5mm and 2mm may better reproduce dosimetry for uncertainties of 0.5% and 1% SD than commonly used 1 mm voxels. Prescribing to isodose level (such as 90%) may provide more consistent dosimetry than prescription to Dmax.