Cardiac substructure avoidance in lung cancer using photon vs proton radiotherapy: A planning study

Abstract

Purpose or Objective Higher radiotherapy dose to selected cardiac structures is associated with cardiac events and death in patients with lung cancer. Studies reveal a dose response relationship for structures at the base of the heart including the right atrium and ascending aorta, with a significant dose threshold of 23Gy (McWilliam et al, 2020). Our objective was to define a cardiac avoidance area and investigate the ability of optimised photon and proton plans to spare these structures compared to standard photon radiotherapy. Materials and Methods A cardiac avoidance area (CAA) was defined, based on previous studies, comprising the superior vena cava, right atrium, aortic root, and proximal coronary arteries (figure 1). 12 patients with stage 3 NSCLC treated at 2 UK centres who had a 4D planning CT and ITV within 5cm of the heart were selected. Plans were generated in Raystation v7R to achieve a dose of 66Gy in 33fractions using standard dose parameters. Cases were re-planned using an additional dose constraint of D1cc<23Gy to the CAA for optimised photon plans. For pencil beam scanning proton (PBSP) plans Varian Eclipse (v13.7) was used for replanning. Photon plans used VMAT with two 360° arcs. PBSP plans used single field optimisation with 3 or 4 fields. PBSP plans were robustly optimised for ITV as target using 5mm setup error and 5% range uncertainty. Plans were compared using analysis of variance and pairwise comparisons.Results Mean ITV volume 219cc (SD 66cc) and 6 ITVs were right sided. The median mean heart dose (MHD) was significantly lower in PBSP plans compared to photon plans (7Gy v 14Gy v 17Gy, p<0.001). CAA was an average of 15% (SD 2%) of the heart volume. Maximum D1cc to CAA was significantly lower with both photon and PBSP cardiac avoidance plans compared to standard plans without cardiac avoidance (19Gy v 22Gy v 41Gy, p=0.003) (figure 2). There was no difference in ITV D95% between the 3 planning strategies. Median mean lung dose (MLD) was 16Gy for standard and cardiac avoidance photon plans; MLD was significantly lower at 9Gy with PBSP (p<0.01). Ability to achieve D1cc<23Gy to CAA depended on the location of the ITV. If CAA>5mm from ITV then D1cc was already <23Gy in standard photon plans. No reduction to CAA D1cc was possible with either PBSP or photons if ITV and CAA overlapped. 9 plans where ITV was on the same axial plane but not overlapping the CAA were analysed. The mean dose reduction in D1cc to the CAA in these patients was 24Gy (SD10Gy) with photons and 26Gy (SD 13Gy) with PBSP. Conclusion In this cohort, dose reduction to a cardiac avoidance area can be achieved with photons and protons without compromising ITV coverage or increasing MLD. Protons did not allow more reduction to D1cc than photons. The location of the ITV determines the maximum dose reduction that can be achieved, allowing selection of patients that could benefit from this strategy.