Session Item

In clinical proton therapy a relative biological effectiveness (RBE) of 1.1 is used, although it is known that the RBE varies and depends on several factors as the LET, tissue type and dose. This has given rise to several different variable RBE models which aim to give a more accurate description of the effects of protons. Proton arc therapy (PAT) is an emerging treatment modality and has showed promise in improving both target coverage and reducing linear energy transfer (LET) in organs at risk (OARs). In the current study we further investigate the potential of RBE based optimization in PAT, employing both full and pruned arc configurations in treatment of complex tumours surrounding critical OARs.

Intensity modulated proton therapy (IMPT) and PAT plans for a paediatric ependymoma patient were created.  The PAT plans consisted of a 240° arc, where the front of the patient received no fields (Figure 1a). Additionally, the pencil beam spots in the arc plan (denoted full PAT plan) with the highest energies in each direction were removed to create a pruned PAT plan (Figure 1b and c), to see if the LET could be elevated in the target while still maintaining the same tumour coverage. The plans were optimized using a FLUKA Monte Carlo based optimization tool with respect to two different variable RBE models, the McNamara model (MCN), and an LET-weighted dose model (LWD), and additionally an RBE of 1.1. As the tumour was surrounding the brainstem, sparing this critical OAR had high priority during optimization. The PAT plans were compared to a reference IMPT plan in terms of their respective RBE weighted doses, physical dose and dose averaged LET (LETd).   

Tumour coverage was achieved for all plans in terms of (respective) RBE-weighted dose (Figure 1a). PAT resulted in lower mean physical doses in the brainstem by up to 0.9 Gy and 1.2 Gy for the full PAT plans and the pruned PAT plans compared to IMPT (Figure 2d), respectively, while there was an increase in LETd in the brainstem for the PAT plans compared to the IMPT (Figure 2f). This could be expected as the brainstem is located in the middle of the target. However, the RBE-weighted dose was lower for the PAT plans in general, compared to the IMPT plan. We also found the MCN plans providing an overall lower LETd to the brainstem compared to the LWD and the RBE_1.1 plans (Figure 2f). 

In this study we demonstrated methods for optimizing PAT plans with respect to variable RBE models, and the potential of using PAT plans in treatment planning, both with full arcs and pruned arcs. Although there was an increase of LET in the brainstem for the PAT plans, the physical doses and the RBE weighted doses were lower in the brainstem compared to the IMPT plan.