Microwave Metasurface-Based Sensor with Artificial Intelligence for Early Breast Tumor Detection.

Area of Science:

• Biomedical Engineering
• Artificial Intelligence
• Microwave Sensing

Background:

• Early breast tumor detection is crucial for effective treatment and improved patient outcomes.
• Current detection methods have limitations in sensitivity and specificity, particularly for dense breast tissue.
• Advanced sensing technologies integrated with AI offer potential for enhanced diagnostic capabilities.

Purpose of the Study:

• To develop and validate a novel microwave metasurface sensor integrated with artificial intelligence (AI) for breast tumor detection.
• To assess the sensor's sensitivity and accuracy across different breast tissue densities and tumor sizes.
• To demonstrate the practical applicability of the combined system through experimental validation.

Main Methods:

• Utilized 137 realistic 3D numerical breast phantoms (Classes C1-C4) with varying tumor sizes as input data.
• Developed a custom neural network trained using cross-entropy loss and the AdamW optimizer.
• Analyzed shifts in the magnitude and phase of the reflection coefficient (S11) for tumor detection.
• Fabricated the sensor and validated its performance using physical breast phantoms.

Main Results:

• The AI model achieved 99% accuracy, with perfect precision, recall, and F1-score for individual breast tissue classes.
• Demonstrated reliable detection of a 10 mm tumor in experimental validation with physical breast phantoms.
• Accuracy varied for paired class combinations (71% for C1/C2, 65% for C2/C3) and decreased to ~50% for all four classes combined.

Conclusions:

• The integration of microwave metasurface sensing and AI presents a highly effective approach for breast tumor detection.
• The developed system shows significant promise for improving the accuracy and reliability of breast cancer diagnostics.
• Further research into optimizing the system for complex, multi-class scenarios is warranted.