Room: Exhibit Hall
Purpose: To provide proof of concept for a script-based method of automatically producing a region of interest (ROI) for spot placement and dose optimization in spatially fractionated radiation therapy (SFRT) techniques such as pencil beam scanning (PBS) proton grid therapy.
Methods: An in-house Python script for RayStation (RaySearch Laboratories, Stockholm, Sweden) was developed to create an array of narrow cylindrical ROIs with user-specified spacing and size, and a choice of either square or hexagonal packing arrangement. Another in-house script was then used to automatically reposition this array of cylinders over the treatment volume and rotate the array to align with the central axis of the beam. Cylinders could be repositioned individually or as a group to achieve optimal alignment with the target. The Boolean intersection of the cylinder array and the target volume produced an ROI for columnar spot placement and dose optimization. The successive Boolean intersection of a pair of orthogonal cylinder arrays and the target volume produced lattice nodes for lattice-style spot placement and dose optimization.
Results: Columnar proton grid therapy dose distributions with high uniformity parallel to the beam axis and high peak-to-trough ratios perpendicular to the beam axis were produced. Lattice-style proton grid therapy plans featured 3 dimensional arrays of high-dose islands in a highly inhomogeneous dose field. Dose distribution characteristics were comparable to clinically-delivered photon grid therapy plans.
Conclusion: The ability to customize the size, spacing, and packing arrangement of ROIs for optimization will facilitate inverse planning for proton grid therapy to control dose distribution characteristics such as peak-to-trough ratios and dose cluster arrangements.