Radial Basis Function for Breast Lesion Detection from MammoWave Clinical Data.

Area of Science:

• Biomedical Engineering
• Medical Imaging
• Machine Learning in Healthcare

Background:

• Novel microwave apparatus (MammoWave) developed for breast lesion detection.
• MammoWave operates in air with rotating antennas (1-9 GHz band).
• Machine learning (ML) is utilized to analyze frequency spectrum data from MammoWave.

Purpose of the Study:

• To assess the efficacy of the MammoWave device augmented with ML for breast lesion identification.
• To evaluate the accuracy, sensitivity, and specificity of the proposed system.

Main Methods:

• Utilized MammoWave to record frequency spectrum from 61 breasts (35 patients).
• Applied Principal Component Analysis (PCA) for feature extraction from frequency response.
• Employed Support Vector Machine (SVM) with a radial basis function kernel for automated lesion identification.

Main Results:

• MammoWave data revealed distinct frequency spectrum behaviors for tissues with and without lesions.
• The SVM model achieved high performance metrics: 91% accuracy, 84.40% sensitivity, and 97.20% specificity.
• Demonstrated in-vivo feasibility validated by ethical committee approvals.

Conclusions:

• The ML-augmented MammoWave system demonstrates high accuracy in identifying breast lesions.
• This technology offers a promising non-invasive approach for breast cancer screening.
• Further validation and clinical integration of MammoWave are warranted.