Abstract
Introduction:
Radiotherapy relies on accurate dose distribution comparison methods, but current approaches have limitations. This study introduces a novel algorithm based on image registration principles to address these limitations.
Materials & methods
The algorithm transforms an evaluated dose distribution into the coordinates of a reference dose distribution using an estimated transformation matrix. This matrix is obtained through a series of image registration steps, including detecting keypoints, constructing descriptors, matching keypoints, and estimating an affine transformation matrix. The resulting evaluated distribution has the same size as the reference, allowing for direct comparison using linear least squares regression.
Results
The algorithm was validated using 175 dose distribution pairs, showing its efficacy for distributions differing by a factor of 2 or less, with acceptable bias and precision. However, accuracy decreased for distributions with smooth dose contours lacking convexity defects (distinct features).
Conclusions
The algorithm validation yielded acceptable outcomes, supporting the potential use of this method in clinical practice. Nevertheless, further research is needed to improve its performance for smooth dose distributions.