Purpose: Historically for Cyberknife, long computation time was a barrier to using Monte Carlo (MC) instead of Ray Tracing (RT) for repeat and final dose calculations during plan optimization. With current computing power, MC can be used during optimization for more accurate dose calculations. In this study, the impact of changing the primary engine from RT to MC was evaluated for radiosurgery treatment planning of vestibular schwannoma with varying target sizes.
Methods: Retrospective analysis was performed in Precision v184.108.40.206 for 23 vestibular schwannoma patients treated at our institution to an Rx of 18Gy in 3 fractions. The mean GTV volume was 2.6+/-3.7cc. From the MC and RT calculations for each patient, we extracted the minimum dose to the GTV, the percent PTV volume receiving the full Rx, and the maximum dose to the PRV of the ipsilateral cochlea. Results were evaluated against 3 GTV sizes: >5cc, 1.5-5cc, and <1.5cc.
Results: For the minimum dose to the GTV, MC vs RT results showed minimal sensitivity to target sizes >5cc. For the second (1.5–5cc) and third (<1.5cc) bins, MC was higher by an average of 2% and 3.5%, respectively. This implies that with MC-based optimization that maintains the nominal Rx, future patients will, on average, receive a lower actual dose to the GTV by the amounts above. Dose to the neighboring OARs will be lower (an aside: for the PRV of the ipsilateral cochlea, MC results were higher by 1%, 2.5% and 5% for the above bins, respectively). The change in the percent PTV receiving full Rx changed by -1.0 +/-3.3%, implying the need for a small scaling in either direction with MC-based optimization.
Conclusion: Switching from RT- to MC-based optimization is not expected to affect the continuity of historical clinical outcomes for vestibular schwannoma patients, including small targets.