Post-implant computed tomography-magnetic resonance prostate image registration using feature line parallelization...

Area of Science:

• Medical Imaging
• Radiotherapy Physics

Background:

• Post-implant dosimetry for prostate brachytherapy relies on CT scans, which have poor soft tissue visualization.
• Integrating CT and MRI data can enhance the definition of anatomical structures and brachytherapy seeds.

Purpose of the Study:

• To develop and evaluate a novel two-stage rigid-body registration algorithm for CT and MR image volumes in prostate brachytherapy.
• To improve the accuracy and automation of image registration for post-implant dosimetry.

Main Methods:

• A two-stage registration algorithm combining parallelization of image features (seed trains, needle tracks) and normalized mutual information.
• Application to CT and 1.5T MR image sets from three patients treated with 125I brachytherapy seeds.
• Validation using simulated datasets and comparison with the Procrustes algorithm.

Main Results:

• Successful registration of clinical CT and MR image volumes with a root-mean-squares distance of <1.5 mm.
• The algorithm requires inclusion of pubic bone and anterior rectum and minimal motion artifact for optimal performance.
• Accuracy comparable to the Procrustes algorithm, with enhanced automation.

Conclusions:

• The novel two-stage registration algorithm effectively integrates CT and MR images for prostate brachytherapy.
• The algorithm offers improved accuracy and automation for post-implant dosimetry.
• Further development is warranted for clinical application in dosimetry.