Purpose: We have previously evaluated the dose calculation accuracy for SRS/fSRS with ion chamber-measured dosimetry in a homogeneous phantom, which found that the measured dose in high dose regions, calculated with BrainLab and Eclipse TPS algorithms, agreed within <5%. However, in clinical cases, >5% discrepancy was sometimes observed between these two TPS. We therefore performed an anthropomorphic phantom study to evaluate dosimetric differences and the dependency on target size and location.
Methods: Single-target SRS plans were generated on a head phantom with three target locations at the center of the brain and two near-skull locations to account for heterogeneity of the phantom. Four different target sizes were investigated from 5 to 20 mm PTV diameter. Plans were generated using BrainLab Elements 3.0 with the dose recalculated in 1mm resolution using Acuros (Eclipse v.15) for comparison. For each of 12 plans, radiochromic film dosimetry was measured using Gafchromic EBT3, scanned with Epson Expression 11000XL.
Results: Dose differences >5% were observed between the two TPSs for PTVs ≤7mm at the center and ≤12mm near the skull (≤0.18cm3 and ≤0.90cm3, respectively). For PTVs at the center, the difference decreased as the PTV size increased; for PTVs near the skull, there was no clear trend. The two TPSs agreed within 5% for larger PTVs. Film measurement showed closer agreement with Elements in field size than Acuros.
Conclusion: Elements and Acuros showed >5% SRS dose calculation differences for smaller targets. The magnitude of these differences are affected by both target size and location. Film measurement showed closer agreement to Elements Monte Carlo calculations than to Eclipse Acuros calculations.
Stereotactic Radiosurgery, In Vivo Dosimetry, Treatment Planning
TH- External Beam- Photons: portal dosimetry, in-vivo dosimetry and dose reconstruction