Efficient gamma index calculation using fast Euclidean distance transform.

Area of Science:

• Medical Physics
• Radiation Oncology
• Computational Imaging

Background:

• The gamma index is essential for comparing radiation dose distributions in radiotherapy.
• Current calculation methods are computationally intensive, limiting clinical application.
• Existing methods struggle to capture both dose difference and spatial agreement effectively.

Purpose of the Study:

• To develop a computationally efficient method for gamma index calculation.
• To reduce the complexity of gamma index evaluation from O(N(2k)) to O(N(k)M).
• To enable faster and more widespread clinical use of 3D gamma index analysis.

Main Methods:

• Embedding the reference dose distribution in a (k+1)-D spatial-dose space.
• Utilizing fast Euclidean distance transform for efficient gamma index table generation.
• Employing table lookup for rapid gamma index calculation for test dose distributions.

Main Results:

• The proposed method reduces computational complexity significantly, achieving a speedup of tens of thousands for 3D gamma index calculations.
• The generated gamma index table is reusable for multiple test dose distributions with the same reference.
• The method facilitates easy derivation of gamma index gradients for discrepancy analysis.

Conclusions:

• The novel method offers a substantial speedup, making clinical implementation of 3D gamma index feasible.
• The derived gradients provide valuable insights for treatment planning and optimization.
• This advancement enhances the utility of gamma index as a quality assurance tool in radiotherapy.