Author: Shen Fu, Zhangmin LI, Zuofeng LI, Yuanshui Zheng 👨🔬
Affiliation: GuangZhou Concord Cancer Center, Guangzhou Concord Cancer Center 🌍
Purpose: To investigate the influence of different robustness optimization strategies and parameters in proton therapy plans for head and neck tumors on target coverage and OAR (organ at risk) sparing, as well as plan robustness and consequent plan adaptation under subsequent treatment setup errors and tissue changes.
Methods: Ten head and neck cancer patients with bilateral cervical lymph nodes and clavicle preventive irradiation were retrospectively selected. The prescribed doses were 5412-6996 cGy(RBE) delivered in 33-35Fx. Treatment plans were generated on the original planning CT images using 5 optimization strategies: robustness optimization with uncertainty parameters of 1) 2mm/3%; 2) 3mm/3%; 3) 5mm/3%, 4) Head region 3mm/3%, neck region 5mm/3%; 5) Traditional PTV and PRV optimization without robustness. Robustness optimization weights were highest for 1st level OARs such as brainstem and spinal cord. Evaluation plans were generated on weekly repeated CTs to compare target area coverage and OAR doses for each optimization strategy and the need for plan adaptation.
Results: a) In nominal plans, target coverage was the highest with strategy 2, and 1st level OARs sparing was the best in strategy 3, with strategy 4 was in between strategies 2 and 3；b) using 3mm/3% uncertainty parameters for plan robustness evaluation and in the worst case scenario, target coverage (CTV V95%) with strategy 5 was the lowest at 74.8%, followed by strategy 2 at 92.4%；c) For evaluation plans based on repeated weekly CTs, strategy 4 seemed to be the best with good balance of target coverage and OARs sparing, and strategy 5 was not robust enough in target coverage.
Conclusion: For proton therapy plans for head and neck tumors, it is necessary to evaluate various plan optimization strategies based on clinical goals and practical setup and range uncertainty parameters.