Novel Wavelet-Based Segmentation of Prostate CBCT Images with Implanted Calypso Transponders

Area of Science:

• Medical Imaging
• Radiotherapy Physics
• Computational Biology

Background:

• Accurate segmentation of organs at risk is crucial for real-time adaptive radiotherapy (ART).
• Prostate cancer patients with implanted Calypso transponders present segmentation challenges due to image artifacts.
• Existing segmentation methods struggle with the artifacts introduced by Calypso transponders in Cone Beam CT (CBCT) images.

Purpose of the Study:

• To develop and evaluate a novel wavelet-based segmentation algorithm for prostate, bladder, and rectum in CBCT images.
• To address the segmentation difficulties caused by Calypso transponder artifacts in CBCT imaging.
• To provide an effective tool for real-time adaptive radiotherapy (ART) in prostate cancer treatment.

Main Methods:

• A Moving Window-Based Double Haar (MWDH) transformation was applied to CBCT images.
• Adaptive thresholding techniques, including cluster-based and Lee filter-based methods, were applied to wavelet coefficients.
• Wavelet reconstruction was used to generate segmented images of the organs of interest.

Main Results:

• The proposed algorithm successfully segmented the prostate, bladder, and rectum in CBCT images from Calypso patients.
• Quantitative evaluation using DICE, Sensitivity, Inclusiveness, and ΔV demonstrated the algorithm's effectiveness.
• The method showed promise in overcoming segmentation challenges posed by Calypso transponder artifacts.

Conclusions:

• The novel wavelet-based segmentation algorithm offers a robust solution for segmenting organs in CBCT images of Calypso patients.
• This technique can potentially enhance the precision and efficiency of real-time adaptive radiotherapy (ART).
• Further validation on a larger cohort is warranted to confirm clinical applicability.