Alignment of multi-segmented anatomical features from radiation therapy images by using least square fitting

Area of Science:

• Medical Physics
• Radiation Oncology
• Image-Guided Therapy

Background:

• Accurate patient positioning is critical in radiation therapy to ensure precise dose delivery and minimize off-target exposure.
• Current methods for patient alignment may have limitations in accuracy and flexibility, particularly with complex anatomical structures.

Purpose of the Study:

• To develop and validate a novel least squares fitting algorithm for aligning multi-curved segments and point pairs of anatomical features.
• To enhance patient position verification in radiation therapy using a robust and adaptable alignment method.

Main Methods:

• Developed a least squares fitting algorithm integrating anatomical features from portal images and simulation radiographs.
• Constructed associated coordinate systems for curved segment pairs to quantify discrepancies.
• Incorporated user selection of multiple curve pairs and individual point pairs on anatomic landmarks.
• Employed a mathematical minimization method to determine optimal transformation for alignment.

Main Results:

• The algorithm successfully aligned multi-curved segments and point pairs of anatomical features.
• Misregistration was quantified by summing positional deviations of point pairs and segment pairs.
• Testing with phantom and clinical images demonstrated high reliability.
• Typical errors in translation and rotation were less than 1 mm and 1 degree, respectively.

Conclusions:

• The developed algorithm provides accurate and reliable patient position verification in radiation therapy.
• It offers flexibility in selecting multiple anatomical features for alignment.
• The method holds potential for improving the precision and safety of image-guided radiation therapy.