Proton PBS Plan Design and Robustness Evaluation for Head and Neck Cancer

Abstract

To evaluate the clinical application and robustness of a newly designed proton pencil beam scanning (PBS) planning technique for radiation therapy of patients with bilateral oropharyngeal cancer (OPC). We planned 10 patients with bilateral OPC using two techniques: proton PBS using 2 posterior oblique fields (group 1) and photon volumetric modulated arc therapy (VMAT) using 2 arcs (group 2). During treatment, we obtained weekly surveillance CT scans. To create verification plans, we forward calculated initial plans on verification scans. Post-treatment, we replanned all patients using 3 fields (2 posterior oblique and 1 anterior field, matched at the level of the thyroid notch) (group 3). We performed a dosimetric comparison of the 2 PBS plans (groups 1 and 3) with the VMAT plans (group 2) using an in silico study. To evaluate plan robustness for each group, we compared dosimetric indicators between initial and verification plans for both target volumes and organs at risk (OARs) using Student’s t-test. Target volumes included gross disease + 6 mm margin (CTV1), high risk microscopic disease (CTV2), and low risk microscopic disease (CTV3). We compared the robustness of target coverage in proton PBS plans (groups 1 and 3) with VMAT plans (group 2). All PBS plans resulted in lower integral dose and sparing of the oral cavity, optical apparatus and salivary glands compared to VMAT plans (p<0.05). We performed 39 verification scans; each patient underwent 3 to 6 verification scans. For target volumes, the 2-field proton PBS plan (group 1) did not meet our robustness threshold (p<0.05). For instance, the mean CTV2 D98% for initial plans was 100% compared to 96% for verification plans, with a maximum ΔD98% of -24%. As a result, 2 patients initially planned with 2-field proton PBS (group 1) were moved to treatment with photon VMAT (group 2). On evaluation of target volume dose, loss in CTV coverage in verification plans for group 1 was prominent in the low-neck region where loose skin folds and/or adipose tissues limited reproducibility of patient position. The 3-field proton PBS plan (which does not use beams traversing the posterior low-neck) and the VMAT plan met our robustness threshold (p>0.05) and demonstrated comparable robustness (p>0.05). For OARs, we found no difference in dosimetric indicators between initial and verification scans for all 3 plan groups (p>0.05). PBS has potential dosimetric advantages over IMRT for both 2- and 3-field irradiation geometries. We achieved PBS plan robustness for the entire cohort with 3-field, but not 2-field, irradiation geometry. The proton PBS plans with 3 field irradiation geometry achieved robustness comparable to photon VMAT plans. Future studies will focus on implementing and examining the 3-field proton PBS plans in clinical practice.