3195 - Background Equivalent Physical Dose for Optimization of Cumulative Equivalent Dose In 2Gy Fractions (EQD2) In Proton Therapy

Purpose/Objective(s): We investigate the feasibility of a voxel-wise cumulative EQD2 optimization method in proton reirradiation (reRT) settings, based on background equivalent physical dose (EPD) on a commercial treatment planning system (TPS).

Materials/Methods: A TPS capable of background dose optimization and voxel-wise EQD2 calculation was used. An in-house script converted prior physical dose distributions to cumulative EQD2 (a/ß=2-3) then back to EPD, a physical dose distribution with equivalent EQD2 in the current treatment’s fractions, n. ReRT patient plans were reoptimized retrospectively to minimize current plan + the background EPD (IRB #00114349). Cumulative physical dose D_total corresponding to physician-desired EQD2_goal varies with prior EQD2, so maximum allowable cumulative physical dose (D_total,max) was derived instead and used as a starting point for optimization, where D_total,max = -n(a/ß)+ v( n² (a/ß)² + 2n (2 + a/ß) EQD2_goal). A variety of reRT scenarios were examined, including parallel organ reRT and hypofractionation. All modified plans maintained the original proton beam parameters (e.g. spot/layer spacing, beam angles), except addition of cumulative dose constraint or substitution of EQD2 hotspot dose-painting structures with said constraints. Resulting plan coverage and cumulative EQD2 were compared to the original clinical plans.

Results: Table 1 compares nominal CTV D95% coverage and OAR EQD2-volume metrics for clinical versus background dose planning. Improvements over clinical plan in either coverage or cumulative EQD2 have been observed for all patients studied. The benefit was especially striking for a plan created without sub-volume EQD2 painting to reduce EQD2 hotspots (patient 1, current 5fx SBRT), though notable benefits persist even for plans with EQD2 painting. Meaningful reduction in EQD2 was achieved while maintaining equal D95% in a multiple-OAR scenario (patient 2, 5fx SBRT) where the target abuts carotids and great vessels. In a Type 2 reRT scenario (patient 3, 30fx conv.) with ultra-central lung reRT, optimization of lung V20Gy EQD2 was achieved using D_total,max over sub-volume EQD2 painting without D95% reduction.

Conclusion: The proposed framework enables direct optimization for any cumulative EQD2-volume goal using TPS with background dose planning capabilities. This represents a breakthrough especially for parallel organs. A limitation is the proposed D_total,max mathematically does not guarantee satisfaction of EQD2 goals, though it focuses on high-risk scenarios where prior and current EQD2 contributions are similar.