A mathematical model of radiation field edge localization

Area of Science:

• Medical Physics
• Radiotherapy Imaging
• Image Analysis

Background:

• Accurate radiation field edge localization is crucial for radiotherapy.
• Existing algorithms for automatic portal image segmentation require refinement.
• Portal imaging is essential for verifying radiation treatment delivery.

Purpose of the Study:

• To develop and evaluate a novel mathematical model for precise radiation field edge localization.
• To improve the performance of an automatic portal image segmentation algorithm.
• To enhance the accuracy of radiation field extraction from portal images.

Main Methods:

• Investigated a mathematical model of average edge slope across the radiation field border.
• Employed a global, hyperbolic function with four parameters to characterize the radiation field penumbra.
• Determined radiation field edge location from model parameters.
• Tested the model on portal images acquired under various clinical conditions.

Main Results:

• The proposed mathematical model demonstrated improved precision in edge localization.
• The global approach outperformed local methods for edge characterization.
• Enhanced performance of the automatic portal image segmentation algorithm was observed.
• The model effectively characterized the radiation field penumbra.

Conclusions:

• The developed mathematical model offers a precise method for radiation field edge localization.
• This approach significantly improves automatic segmentation of portal images in radiotherapy.
• The findings contribute to more accurate radiation therapy delivery and verification.