Related Experiment Video
Updated: Sep 29, 2025
06:08
A Cognitive Fusion-guided Prostate Biopsy Using Multiparametric Magnetic Resonance Imaging and Transrectal Ultrasound
Published on: March 21, 2025
405
H-ProSeg: Hybrid ultrasound prostate segmentation based on explainability-guided mathematical model
Tao Peng1, Yiyun Wu2, Jing Qin3
1Department of Health Technology and Informatics, The Hong Kong Polytechnic University, Hong Kong, China.
Computer Methods and Programs in Biomedicine
|March 26, 2022
Summary
This study introduces H-ProSeg, a novel hybrid method for accurate prostate segmentation in ultrasound images. The method demonstrates robust performance, improving prostate cancer diagnosis and treatment planning.
Area of Science:
- Medical imaging
- Artificial intelligence in medicine
- Urology
Background:
- Accurate prostate segmentation in transrectal ultrasound (TRUS) images is critical for image-guided interventions and cancer diagnosis.
- Challenges include ambiguous boundaries, artifacts, and anatomical variations.
Purpose of the Study:
- To develop and evaluate a hybrid method (H-ProSeg) for accurate and robust prostate segmentation in TRUS images.
- To improve the precision of prostate boundary identification for clinical applications.
Main Methods:
- A hybrid approach (H-ProSeg) utilizing radiologist-defined seed points.
- Employs three subnetworks: principal curve-based model, differential evolution-based artificial neural network, and mathematical contour description.
- Validated on 55 brachytherapy patients.
Main Results:
- H-ProSeg achieved high segmentation accuracy: Dice Similarity Coefficient (DSC) 95.8%, Jaccard Index (Ω) 94.3%, and Accuracy (ACC) 95.4%.
- Demonstrated excellent robustness against Gaussian noise (σ=50), with minimal performance fluctuation (max ~2.5%).
Conclusions:
- The H-ProSeg method offers superior performance for prostate segmentation compared to existing techniques.
- Accurate prostate boundary delineation is vital for preserving critical structures and improving cancer treatment outcomes.
- The proposed models show potential for enhancing prostate cancer diagnosis and therapeutic results.
