Session Item

The adaptive aperture (AA) is a unique feature of Mevion Hyperscan S250i proton beam delivery system. In contrary to a static aperture, the AA allows for dynamic trimming of the proton beam (e.g. similar to a photon MLC) with the shape of AA adjusted for every energy layer. In its current implementation, only one AA setting per energy layer is available. As a consequence, only spots on the edge of the spot map are trimmed. Trimming of the inner spots could be beneficial in systems with relatively big spot sizes. Here we investigated the added value of a forward AA optimization to create steeper lateral dose fall-off and show the potential for reduction of organs at risk (OARs) dose.

The simulation part of the study was conducted using single treatment beam and a numerical phantom (simulated water tank with 10x10x1cm CTV at 4.8cm depth and OARs of 10x2x1cm placed on the side of the CTV, with CTV-OAR distance ranging from 0 to 10mm). The clinical investigation included three patients with brain tumors, with variable location within the brain. Treatment plans were created in RayStation v10A.

First, a plan was optimized with standard AA settings according to clinical protocol, using robust optimization with 1mm setup and 3% range uncertainty. Next, three additional AA positions were added per energy layer using RayStation scripting. To create the new AA shapes, original AA was shifted towards the field center, so that it closely followed the first and the second outermost contour of the spots (Fig 1). Finally, the modified AA plans were reoptimized and objectives were adjusted if needed to ensure a clinically acceptable plan. The plans were evaluated for their robustness computing voxel-wise minimum and maximum (VWmin/max) dose distributions. The CTV coverage was analyzed in VWmin, the D0.03cc in VWmax and mean doses in the nominal dose distribution. For the clinical cases, the reduction in 80% isodose between standard AA and modified AA plans was evaluated.

In the phantom study, the lateral beam penumbra was reduced by 2.6mm (from 10.3mm to 7.7mm). Furthermore, relevant reduction of dose to simulated OARs was observed. Reduction in mean dose to OARs was around 7% of the prescribed dose. Reduction in D1% ranged from 0% to 16% and was larger for the OARs at larger CTV-OAR distance.

For all 3 clinical cases, CTV coverage remained comparable to the standard AA plan. The modified AA plans resulted in reduction of dose to all OARs (Fig 2) with the exception of small increase of D0.03cc to brainstem and skin. However, this increase was below 1.5Gy and remained within the OAR tolerance. For every patient, at least one OAR had dose reduction >2Gy. Volume of isodose 80% (normalized to CTV volume) decreased by 16%, 12% and 44% for patients 1, 2 and 3, respectively.

This study presented a proof of concept describing the optimization of AA. The lateral dose fall-off was improved and led to clinically relevant reduction in OAR doses with preserved CTV coverage.