Session Item

Pelvic lymph node irradiation is recommended in patients with high-risk prostate cancer. Due to the large treatment volumes, these patients could benefit from the normal tissue sparing that might be achieved with proton therapy. However, the influence of the proton relative biological effectiveness (RBE) on normal tissue doses for this treatment site has not previously been studied. We are launching a multi-centre randomised clinical trial to compare front-line photon-based radiotherapy with pencil-beam scanning (PBS) proton therapy for high-risk prostate cancer. The aim of this study was to explore the implications of potentially variable RBE for the proton therapy class solution that will be used in the trial. 

The proton therapy class solution consisted of four PBS beams, including two lateral beams tilted slightly posterior to avoid the femoral heads (gantry angles 100°/260°) and two posterior beams angled slightly laterally to avoid the rectum (gantry angles of 170°/190°). The two nearly lateral beams covered the high-dose target volume and in addition the lymph nodes on either side. The two nearly posterior beams covered the entire target with prostate, seminal vesicles and lymph nodes on both sides. Treatment planning of eight test cases was done in Eclipse, while dose and linear-energy transfer (LET) distributions were re-calculated using Monte Carlo simulations (TOPAS v3.5/Geant4 v.10.6) for our beamline (Fig. 1). Biological doses with published variable RBE models (Carabe, Wedenberg, McNamara) as well as the LET-weighted dose (LWD) (Fig. 1c) were compared to the clinically applied constant RBE (RBE=1.1). The analysis focused on the bowel, bladder and rectum, also taking the uncertainty in the model input parameter α/β into account (±2 Gy around the point estimates). The differences between different RBE models and between different α/β values were assessed at dose levels corresponding to the dose/volume constraints of the trial.

With both LWD and even more so with the LET-based variable RBE models, the biological doses in the bowel, bladder and rectum increased compared to constant RBE (Fig. 2a). The Wedenberg and Carabe RBE models resulted in the highest doses. At the dose levels of the normal tissue constraints, volume increases of 0.3-8% were seen, largest for the bowel. For each of the three RBE models, lowering the α/β increased the biological doses, with the effect being in the same range as the differences between the models (Fig 2b).

Accounting for the potentially variable RBE in proton therapy of high-risk prostate cancer resulted in increases in relative volumes of up to 8% at the dose levels corresponding to the trial dose/volume constraints. The choice of model and the choice of input parameter had similar effects. The biological doses obtained with the LWD were closest to the constant RBE. Biological doses including variable RBE will be assessed prospectively in our trial.