Constrained non-rigid registration for use in image-guided adaptive radiotherapy

Area of Science:

• Medical Physics
• Radiotherapy
• Image-guided therapy

Background:

• Accurate patient positioning and organ alignment are crucial for effective radiotherapy.
• Inter-day variations in patient anatomy, including bone motion and organ deformation, pose challenges in adaptive radiotherapy.

Purpose of the Study:

• To present a constrained non-rigid registration (CNRR) algorithm for prostate image-guided adaptive radiotherapy.
• To evaluate the performance of CNRR in aligning critical organs and bones for dose adaptation.

Main Methods:

• Developed a CNRR algorithm combining global rigid and local non-rigid Free Form Deformation (FFD) transformations.
• Utilized control points for non-rigidly constraining transformations to the prostate, rectum, and bladder.
• Applied rigid constraints to pelvic and femur bones to improve registration accuracy.
• Tested the algorithm with 3D conformal radiotherapy (3DCRT) and intensity-modulated radiotherapy (IMRT) datasets.

Main Results:

• CNRR significantly outperformed rigid and non-rigid registration methods in aligning bones and organs.
• Tested CNRR with various anatomical constraint combinations, showing consistent improvements.
• Adapted dose plans using CNRR results reduced radiation to the rectum and bladder.
• Maintained or increased radiation dose to the prostate in adapted plans.

Conclusions:

• The CNRR algorithm provides a robust framework for accurate image registration in prostate radiotherapy.
• CNRR enables effective dose adaptation by accounting for patient positioning errors and anatomical variations.
• This approach enhances treatment efficacy and safety by optimizing radiation delivery.