(Sunday, 7/14/2019)
Room: ePoster Forums
Purpose: To assess the robustness of MFO IMPT plans utilizing robustly-optimized (RO) bilateral fields and VMAT plans under systematic setup and range uncertainties for prostate cancer.
Methods: Ten previously treated prostate VMAT patients who received 74-76 Gy in 37-38 fractions were proton planned. IMPT plans were normalized such that 100% covers 99.99% of the CTV with worst-case scenario V95 ≥ 95% CTV. For robustly-optimized IMPT plans, combinations of 6 translational and 2 range uncertainty perturbations modeled the ± 0.5 rigid isocenter shifts and ± 3% CT calibration curve error, respectively. Identical rigid shifts of ±0.5 cm in the x, y, and z directions were manually applied to the VMAT plans. CTV coverage, OAR sparing, and dose envelope size differences was assessed for all scenarios.
Results: The VMAT constraint pass rates were either comparable to or slightly better than the proton scenarios for 80% of patients. Visual inspection shows proton OAR sparing in the lower dose region is consistently better than the photon OAR sparing. The anorectum violated the most constraints, with 18 proton and 13 photon violations whereas the bladder violated only once. The photon plans are more robust on the CTVs than they are in the proton plans, but at the cost of the robustness on the anorectum. Three patients had 5-7 overall scenario plans with insufficient OAR sparing and 10% of all scenarios were clinically insufficient, 91% of which were due to anorectum violations. The percentages due to CTV hot spots and bladder insufficiency are 6% and 3%, respectively.
Conclusion: RO provides a technique to generate IMPT plans that are less sensitive to errors. This work demonstrates that RO-IMPT plans produced comparable plans to the ones obtained with VMAT for prostate cancer. The anorectum, most sensitive OAR under different error scenarios, must have sufficient constraints for both modalities.
Keywords
Not Applicable / None Entered.
Taxonomy
Not Applicable / None Entered.
Contact Email
