Breast Tumor Detection and Classification Using Intravoxel Incoherent Motion Hyperspectral Imaging Techniques

Area of Science:

• Radiology and Medical Imaging
• Oncology
• Biomedical Engineering

Background:

• Breast cancer remains a leading global health concern for women.
• Traditional imaging like mammography and ultrasonography have limitations.
• Magnetic Resonance Imaging (MRI) offers superior resolution but requires contrast agents with potential toxicity and concerns over gadolinium deposition.

Purpose of the Study:

• To develop and evaluate an advanced imaging analysis technique for breast cancer assessment.
• To overcome the limitations and risks associated with contrast agents in MRI.
• To improve the accurate classification of breast tissue types using quantitative imaging analysis.

Main Methods:

• Development of an intravoxel incoherent motion (IVIM) MRI-based histogram analysis.
• Utilized hyperspectral techniques including band expansion process (BEP) and automatic target generation process (ATGP).
• Employed kernel constrained energy minimization (KCEM) for lesion detection, followed by decision tree and histogram analysis for classification.

Main Results:

• The proposed IVIM-MRI histogram analysis effectively differentiated between malignant tumors, cysts, and normal breast tissues.
• The method demonstrated quantitative analysis capabilities for lesion characterization.
• Successful classification of three distinct breast tissue categories was achieved.

Conclusions:

• The IVIM-MRI-based histogram analysis offers a promising, contrast-agent-free approach for breast cancer assessment.
• This technique enhances diagnostic accuracy and patient safety by avoiding toxic contrast agents.
• The study highlights the potential of advanced MRI analysis in improving breast cancer detection and characterization.