A 3-D Level Set Method for Microwave Breast Imaging.

Area of Science:

• Biomedical Engineering
• Medical Imaging
• Computational Electromagnetics

Background:

• Conventional microwave breast imaging algorithms yield moderate resolution and blurred tissue boundaries.
• Level set methods in 2-D studies preserve dielectric boundaries for higher resolution, but are computationally expensive for 3-D.
• Existing 3-D algorithms are computationally impractical.

Purpose of the Study:

• To develop a computationally tractable 3-D microwave imaging algorithm based on level sets.
• To improve the accuracy and resolution of dielectric property reconstructions in 3-D breast imaging.
• To provide a viable alternative to existing breast imaging modalities.

Main Methods:

• Reduced computational cost of level set method using Jacobian matrix for Frechet derivatives.
• Demonstrated feasibility with simulated 3-D array measurements from numerical breast phantoms.
• Compared 3-D reconstructions against conventional microwave imaging techniques.
• Quantitatively assessed algorithm's efficacy in breast density evaluation.

Main Results:

• Reconstructions of 3-D numerical breast phantoms showed improvement over conventional microwave imaging.
• Level set algorithm provided more accurate breast density estimates than conventional methods.
• Accuracy of density estimation surpassed reported mammographic values.

Conclusions:

• The level set method achieves feasible computational complexity for 3-D imaging.
• Accurate reconstruction of heterogeneous dielectric properties of the breast is achieved.
• 3-D microwave breast imaging with level sets is a promising, nonionizing alternative.