A quantitative framework for patient-specific collision detection in proton therapy

Area of Science:

• Medical Physics
• Radiation Oncology

Background:

• Proton therapy requires beam modifying accessories close to patients for optimal dosimetry.
• Large proton treatment units can lead to collisions between equipment and patients, hindering planned treatment geometry.

Purpose of the Study:

• To develop a quantitative framework for modeling patient-specific proton treatment geometry.
• To minimize air gaps and prevent collisions during proton therapy.

Main Methods:

• Patient contours converted to IEC gantry coordinates, considering orientation and beam angles.
• Snout components modeled as 3D geometric shapes; collision detection based on isocenter, snout type, and extension.
• 3D GUI for visualization; analytic algorithm quantifies collisions or minimum distances.

Main Results:

• Successfully modeled three snout designs and demonstrated collision avoidance through snout retraction.
• Validated for various patient orientations, sitting positions, and multiple isocenters.
• Demonstrated dosimetric advantages of reduced air gaps compared to standard plans.

Conclusions:

• Framework implementation reduces treatment room collisions.
• Enables planners to minimize air gaps for improved proton therapy plan dosimetry.