Introducing a New Spot-Scanning Proton Arc Optimization Algorithm with a Variable Tolerance Window to Improve Plan Quality for Bilateral Breast Cancer
Abstract
Purpose
This study aims to introduce a new SPArc algorithm with a variable tolerance window (SPArc-variable) to further improve plan quality for patients receiving bilateral breast and regional nodal irradiation, by allowing extra energy layers across specific arc trajectories.
Methods
The SPArc-variable optimization algorithm is based on a 0-1 knapsack problem in dynamic programming (DP). Starting with the multifield-IMPT plan with 20 degrees apart, the algorithm iteratively selects a subset of energy layers with decreasing energy layer filtration factor. Then it resamples the energy layers based on the gantry’s movement and irradiation sequence. Finally, spot weighting optimization is applied. Five cases with bilateral breast and lymph nodes treatment are retrospectively selected. Three treatment planning groups were generated, including IMPT with 2 isos, SPArc planning with a fixed distance of 2.5 degrees between adjacent control points(SPArc-original), and SPArc planning with non-fixed distance(SPArc-variable), with a prescription of 5000cGy(RBE). DVH metrics are included to evaluate performance over OARs and targets, and dynamic delivery is simulated via a published dynamic arc system controller.
Results
Both SPArc-original and SPArc-variable plans showed slightly better target coverage than IMPT plans, but the difference was not statistically significant. In addition, SPArc-variable showed better performance in sparing the heart(P<0.01 for D1%) and better performance in sparing the left and right lungs(P<0.01 for V20Gy and V5Gy) than both IMPT and SPArc-original. In terms of delivery efficiency, the SPArc-variable shows a slightly longer treatment delivery time than SPArc-original and is superior to IMPT(SPArc-variable:434.99±52.95sec, IMPT: 506.47±14.65sec, P=0.02, SPArc-original:382.11±49.31sec, P=0.01).
Conclusion
This study introduces a novel SPArc optimization algorithm with a variable window to enhance dosimetric performance for bilateral breast cancer patient population. More specifically, it reduces the dynamic delivery time compared to IMPT, while achieving equivalent target coverage and significantly superior heart and lungs sparing compared to IMPT and SPArc-original.