Deformable scintillation dosimeter I: challenges and implementation using computer vision techniques.

Area of Science:

• Medical Physics
• Radiation Oncology
• Detector Technology

Background:

• Plastic scintillation detectors offer advantages for radiotherapy dosimetry, including water-equivalence and real-time response.
• Embedding scintillators in deformable matrices can mimic anatomical changes, expanding their application range.
• Characterizing signal variations due to fiber movement is crucial for accurate dosimetry.

Purpose of the Study:

• To characterize signal variations in plastic scintillating fibers due to translation and rotation.
• To develop and validate correction methods for real-time deformable dosimetry.
• To assess the accuracy and precision of a prototype deformable dosimeter system.

Main Methods:

• Characterized signal variations of scintillating fibers with respect to camera position.
• Implemented stereo vision techniques with sCMOS and CCD cameras for corrections.
• Developed a prototype deformable dosimeter with 19 scintillating fibers.
• Analyzed signal-angle and signal-displacement relationships.

Main Results:

• Signal-angle relationship followed a Gaussian distribution (FWHM = 52°).
• Intensity variation from radial displacement followed the inverse square law.
• Position detection accuracy was 0.08 mm and precision was 0.3 mm.
• Angle measurement uncertainty was 2°.
• Corrections reduced intensity variations from ±10% to ±1% for depth displacement.
• Corrections improved lateral displacement intensity variations from 98±3% to 100±1%.

Conclusions:

• Accurate correction of camera-imaged signals from scintillating elements in 3D is achievable.
• The developed stereo vision techniques effectively correct for spatial dependencies in deformable dosimeters.
• This research enables the development of advanced real-time, scintillator-based deformable dosimeters for radiotherapy.