Feature-based rectal contour propagation from planning CT to cone beam CT

Area of Science:

• Medical Imaging
• Radiotherapy
• Image Registration

Background:

• Accurate delineation of organs at risk, such as the rectum, is crucial for effective radiotherapy planning.
• Variations in rectal filling and anatomy between planning CT (pCT) and cone beam CT (CBCT) pose challenges for adaptive radiotherapy.

Purpose of the Study:

• To develop and validate a novel feature-based registration strategy for automatic rectal contour mapping from pCT to CBCT.
• To enhance the precision of organ segmentation in adaptive radiotherapy workflows.

Main Methods:

• A feature-based registration algorithm utilizing Scale Invariance Feature Transform (SIFT) to match control points between pCT and CBCT.
• A thin plate spline transformation was employed to warp a narrow band around the rectum, facilitating contour propagation.
• The method involved manual outlining of rectal contours on pCT, followed by automated mapping to CBCT.

Main Results:

• The contour propagation technique demonstrated clinically acceptable results in both digital phantoms and clinical cases.
• The method proved robust to significant rectal content changes between pCT and CBCT scans.
• Over 90% accordance was achieved between registered rectal volumes and manual segmentations.

Conclusions:

• The developed feature-based registration strategy offers a reliable and automated solution for rectal contour mapping in adaptive radiotherapy.
• This technique can significantly improve the accuracy and efficiency of adaptive radiotherapy for prostate, rectal, and gynecological cancers.
• The method provides a valuable tool for real-time treatment adjustments in radiotherapy.