A Digital Imaging and Communications in Medicine-based Collision Prediction System with an Angle Optimization...

The purpose of this study was to develop and validate a Digital Imaging and Communications in Medicine (DICOM)-based collision prediction system for electron beam therapy. This system enables a quantitative pretreatment assessment of mechanical interference between the electron applicator and the patient's body surface. It also automatically proposes safe alternative beam angle combinations.

We geometrically approximated a 20 cm × 20 cm electron applicator mounted on a TrueBeam STx linear accelerator as a 190-mm cube. We extracted body and table contours, as well as isocenter coordinates, from the RT Structure Set. For each combination of gantry, table, and collimator angles, the minimum clearance distance between the applicator model and the body contour was calculated. The discretized angle space was visualized using heatmaps of the minimum clearance distance. An optimization algorithm was implemented to calculate a score for each candidate angle set using a weighted sum of angular deviations from the original plan and to identify the best alternative angle configurations.

The system was evaluated using an adult female anthropomorphic phantom and four intentionally selected electron regions (superior, inferior, medial, and lateral around the left nipple). For the superior, inferior, and medial regions, the initial planned angles produced geometric intersections. However, the system-suggested alternative angle sets eliminated these intersections in three-dimensional visualization and were confirmed to be collision-free on the linear accelerator.

The proposed system requires no additional hardware, integrates with current workflows, and provides a practical tool for collision risk assessment and angle optimization in nearsurface electron therapy.