A Novel Method for the Delivery of 3-Dimensional High-Dose Spatially Fractionated Radiation Therapy With Pencil Beam Scanning Proton Therapy: Maximizing the Benefit of the Bragg Peak

Abstract

High dose spatially fractionated (GRID) radiotherapy has been utilized sparingly in clinical practice by means of traditional photon techniques. However, this approach has shown great potential to allow for safe dose-escalation in large tumors by mimicking brachytherapy distributions on an external beam platform. While both scattered and scanned particle beams have been proposed for this application, all previously reported models utilize spatially fractionated spread out Bragg peaks (SOBPs), that, while providing less than ideal valleys and peaks of GRID therapy, substantially increase entrance dose and fail to take full advantage of the enhanced radiobiologic effect (RBE) at end of range. We hypothesized that pencil beam scanning proton therapy (PBS-PT) based GRID plans could be developed utilizing only pristine Bragg peaks (without SOBPs), addressing each of these issues. Five patients with tumors greater than 10 cm in greatest dimension and in different anatomical sites (head/neck, lung, abdomen, pelvis, and extremity) were identified retrospectively. Gross tumor volume was delineated and contracted by a small safety margin to ensure that all Bragg peaks would fall within tumoral tissue. A trained physicist delineated a 3-dimensional lattice of 3 mm spherical targets, spaced center-to-center 3 cm apart, each intended to be treated by an individual “spot.” Two beam angles were utilized for each plan, and the target lattice was rotated in 3 dimensions to prevent spot stacking. Inverse optimization with manual spot map modifications were employed. Plan peak-to-valley ratios were optimized to be consistent with those achieved by photon-based GRID in our clinic. All plans achieved consistent maximum dose (Dmax) peaks approximating 15 Gy(RBE) with a confluent valley dose across targeted tumor of ∼1-2 Gy(RBE). The depth of valleys was able to be modulated based on target lattice spacing. No SOBPs were utilized in planning, instead, each spot entrance path was only utilized once, though laterally penumbra was allowed to overlap in the valleys. In even superficial targets, maximum skin dose was reduced to less than 50% Dmax to small islands of surface tissue, as is customary with GRID. All Bragg peaks terminated within gross tumor offering the potential for enhanced tumoral RBE with maximal sparing of tissues beyond the target. Plans incorporated between 11 and 28 energy layers and from 42 to 157 total spots. PBS-PT based GRID therapy can be achieved solely utilizing pristine Bragg peaks without need for SOBP or spot stacking. This technique offers the promise of dose escalation, maximal skin, and organ at risk sparing, and enhanced tumoral RBE. Efforts to develop algorithmic automation and inverse optimization of this methodology are underway.