Related Experiment Video
Updated: Mar 3, 2026
08:17
Dynamic Lung Tumor Tracking for Stereotactic Ablative Body Radiation Therapy
Published on: June 7, 2015
16.3K
A priori patient-specific collision avoidance in radiotherapy using consumer grade depth cameras
Rex A Cardan1, Richard A Popple2, John Fiveash2
1Department of Radiation Oncology, University of Alabama at Birmingham, 2145 Bonner Way, Birmingham, AL, 35243, USA.
Medical Physics
|May 6, 2017
Summary
This study presents a low-cost, rapid framework for radiotherapy treatment planning, enabling patient-specific collision avoidance during simulation. The system accurately maps potential collisions, ensuring safer treatment delivery.
Area of Science:
- Medical Physics
- Radiotherapy Technology
- Computational Geometry
Background:
- Radiotherapy requires precise patient positioning to avoid collisions between the treatment unit and the patient or hardware.
- Current methods for collision avoidance can be time-consuming and may not be fully patient-specific.
- Accurate mapping of the treatment unit's (gantry and couch) potential movement is crucial for safety.
Purpose of the Study:
- To develop and evaluate a low-cost, fast, and accurate framework for radiotherapy collision avoidance.
- To create a patient-specific implementation integrated into the standard simulation process.
- To assess the system's ability to map feasible treatment positions and prevent collisions.
Main Methods:
- Consumer depth cameras were used to create 3D polygon meshes of patients and the treatment unit.
- A polygon interference algorithm was employed for virtual collision detection between patient and machine models.
- Physical collision mapping was performed in the treatment room to validate virtual predictions.
- A buffer geometry was applied to the gantry mesh to enhance collision detection accuracy.
Main Results:
- Optical scanning and virtual collision mapping were completed in under a minute per patient.
- The system achieved high accuracy (97.3% ± 2.4%) and negative prediction rate (96.9% ± 2.2%) in raw collision detection.
- A 6 cm polygon buffer on the gantry mesh improved the negative prediction rate to 100%, ensuring all collisions were detected.
- Increasing the buffer size slightly reduced overall accuracy due to increased false positives.
Conclusions:
- A fast, low-cost framework for a priori collision space mapping during radiotherapy simulation was successfully demonstrated.
- Polygon interference effectively identifies potential collisions, with a buffer geometry recommended for geometric uncertainties.
- This patient-specific system enhances safety by enabling collision avoidance during the simulation phase.
