Analysis and classification of tissue with scatterer structure templates.

Area of Science:

• Medical Imaging
• Ultrasound Signal Processing
• Biomedical Engineering

Background:

• Back-scattered ultrasound signals reveal scatterer structure information.
• Conventional intensity images highlight large-scale structures like tissue and tumor boundaries.
• Small-scale scatterer structures manifest as textures in ultrasound images.

Purpose of the Study:

• To introduce and apply the generalized spectrum (GS) for characterizing small-scale scatterer structures in breast masses.
• To analyze scatterer structures in malignant and benign breast masses using the GS.
• To develop methods for scaling and normalizing the GS to minimize system and tissue variability.

Main Methods:

• Described the generalized spectrum (GS) for characterizing small-scale scatterer structures.
• Applied GS analysis to radiofrequency (rf) A-scans from 41 breast mass regions in 26 patients.
• Developed classifiers based on GS-identified structural differences between fibroadenoma and carcinoma masses.

Main Results:

• The generalized spectrum (GS) was used to analyze structural properties of breast masses.
• GS analysis facilitated discrimination between benign (fibroadenoma) and malignant (carcinoma) breast masses.
• Classifiers utilizing GS analysis achieved approximately 82% true-positive and 10% false-positive rates.

Conclusions:

• The generalized spectrum (GS) is a valuable tool for characterizing small-scale scatterer structures in ultrasound.
• GS analysis demonstrates potential for discriminating between benign and malignant breast masses.
• This approach offers a promising method for improving breast cancer diagnosis accuracy.