A Retrospective Sensitivity Analysis of Dynamic Collimation System to Realistic Delivery and Mechanical Uncertainties
Abstract
Purpose
To experimentally quantify the mechanical delivery errors of a pre-clinical Dynamical Collimation System (DCS) prototype for intensity modulated proton therapy (IMPT) and evaluate the dosimetric impact of these errors to the target coverage and organ sparing.
Methods
Treatment delivery errors were quantified and simulated using log-based QA from 30 patient-specific quality assurance treatment fields delivered with the DCS as integrated within the IBA DN beam line at the Miami Cancer Institute. DCS treatments and delivery mechanics were optimized so that each treatment field could be delivered within 60 seconds of their uncollimated counterpart. The distribution of spot and collimation positioning error were analyzed to quantify the range of mechanical errors between the DCS and beam scanning controller. The robustness of DCS-collimated treatments were interrogated by simulating a distribution of 50 candidate delivery scenarios among three patient plans that simulated DCS isocentric mounting, trimmer and spot positioning errors, which were modeled from normal distributions with measured standard deviations of 0.2, 0.25, and 0.2 mm, respectively.
Results
Sub-millimeter delivery accuracy, with average spot positioning errors of −0.05 ± 0.20 mm and 0.04 ± 0.20 mm in the lateral x and y directions, respectively. Across all three treatment plans, target coverage remained robust, with the range of target D95 variation limited to ±0.91% of the intended treatment delivery. Variations in the target D2% remained within 67.2 Gy across all scenarios. For organs at risk, changes in D2% were generally below 65.3 Gy relative to nominal plans, with no systematic degradation of dose conformity or sparing observed.
Conclusion
The DCS technology can deliverable efficient treatments with sub-millimeter accuracy between collimator and spot positioning. As such, DCS treatments exhibit minimal sensitivity to the range of observed treatment delivery errors from the DCS and beam scanning controller.