A novel method for comparing 3D target volume delineations in radiotherapy

Area of Science:

• Medical imaging analysis
• Computational geometry
• Biomedical engineering

Background:

• Accurate comparison of 3D volume surfaces is crucial in medical imaging.
• Existing local distance measures between delineations suffer from overestimation and asymmetry.
• These drawbacks limit the reliability of quantitative analysis in applications like radiotherapy planning.

Purpose of the Study:

• To introduce a new, reliable, and symmetric method for calculating local distances between two delineations.
• To address the limitations of current distance measurement techniques.
• To provide a more robust tool for comparing volumetric data in medical imaging.

Main Methods:

• A novel method is proposed that identifies corresponding points between two delineations.
• This is achieved by traversing a vector field derived from the combined gradient of distance transforms.
• The approach ensures a fundamentally more reliable and symmetric distance measure.

Main Results:

• The new method provides a symmetric and more reliable local measure of distance between volume surfaces.
• Illustrative examples, including observer variation in prostate delineation, demonstrate the method's effectiveness.
• The proposed technique overcomes the overestimation and asymmetry issues of previous methods.

Conclusions:

• The developed method offers a significant improvement for quantifying differences between 3D delineations.
• Its reliability and symmetry make it suitable for applications sensitive to accurate surface comparisons, such as medical image segmentation analysis.
• The method provides a valuable advancement in the field of medical image analysis and comparison.