Automatic breast lesion segmentation in phase preserved DCE-MRIs.

Area of Science:

• Medical Imaging
• Image Processing
• Computational Biology

Background:

• Accurate segmentation of breast lesions in Dynamic Contrast Enhanced (DCE) MRI is crucial for diagnosis and treatment planning.
• Existing segmentation methods may struggle with noise and preserving fine details in MRI images.

Purpose of the Study:

• To develop and validate a novel framework for automatic and accurate segmentation of breast lesions from DCE-MRI.
• To enhance lesion delineation by integrating denoising and continuous max flow/min cut algorithms.

Main Methods:

• The framework employs a three-stage approach: image subtraction and registration, phase-preserved denoising with adaptive Wiener filtering, and continuous domain max flow/min cut for lesion detection.
• Morphological operations are utilized for post-processing to refine segmentation results.
• The method operates on phase-preserved denoised images to maintain important image features like edges.

Main Results:

• The proposed method demonstrated high-quality segmentation of breast lesions in DCE-MRI.
• Qualitative and quantitative evaluations using nine performance metrics across 21 cases confirmed the method's effectiveness.
• The framework successfully handled different MR image resolutions.

Conclusions:

• The developed framework provides an effective and accurate solution for automatic breast lesion segmentation in DCE-MRI.
• The integration of denoising techniques with continuous max flow/min cut algorithms offers a robust approach for medical image analysis.
• The method shows significant potential for improving diagnostic accuracy in breast cancer detection.