342 - Dynamic Individual Spot Collimation In Proton Beam Scanning with a Robotic Arm-Driven Pinhole Aperture

Purpose/Objective(s): Range shifters (RS) in pencil beam scanning (PBS) proton therapy are essential to treat shallow tumors at the cost of increased spot sizes, resulting in non-optimal dose conformality for shallow targets and poor protection of nearby organs-at-risk. Dynamic apertures partially mitigate this effect, but they are bulky, costly, and not very reliable due to the complicated mechanical design. This work proposes dynamic individual spot collimation (DISC) using a robotic arm–driven pinhole aperture to enable simple, flexible, spot-by-spot beam shaping with a small footprint.

Materials/Methods: A custom tungsten pinhole aperture (6-mm diameter) was mounted on a robotic arm (Universal Robots) and calibrated to the proton gantry coordinate system using a kV imaging–based workflow. Positioning accuracy was evaluated by delivering spots to predefined locations with kV imaging. Dosimetric accuracy was assessed using radiochromic film measurements compared with Monte Carlo simulations. Experiments included single-spot measurements for quantitative comparison of lateral dose profiles between collimated and uncollimated beams. A nine-spot square array (2 cm spacing) was delivered to assess robotic positioning reproducibility and tracking accuracy. A 1D dose line was further delivered to film twice, once with and once without edge-spot collimation while preserving central dose uniformity in both cases, in order to quantify improvements of DISC in the lateral penumbra.

Results: Aperture positioning accuracy was within 0.1 mm based on kV image verification. Collimated single-spot profiles demonstrated reduced spot size and sharper lateral dose falloff. Film measurements showed strong agreement with Monte Carlo simulations, with a gamma passing rate of 97.8% (2 mm/3%). Nine-spot array measurements demonstrated mean positional deviation of 0.1 mm from the planned 2 cm center-to-center spacing during sequential robotic delivery, indicating stable robotic tracking during sequential spot delivery. Line-beam experiments demonstrated a 4 mm reduction in lateral penumbra when DISC was applied to the edge spots, consistent with Monte Carlo predictions.

Conclusion: DISC addresses a fundamental limitation of PBS delivery with a RS for shallow tumors and enables sub-millimeter accurate, spot-by-spot beam shaping. This approach has potential to further improve normal tissue sparing for shallow and organ-at-risk–adjacent targets and enable stereotactic and small-field PBS.