Mitigating the effects of breast and excitation-source misalignment in low-frequency breast cancer detection

Accurate breast-tumor detection requires precise positioning of both the excitation source and the breast, as motion or misalignment can degrade imaging performance. This work presents a novel technique to compensate for this effect and to reliably identify tumor presence and 2-D localization in breast impressions.

A numerical alignment approach was developed to estimate source-breast misalignment via cross-correlation and compensate for the resulting displacement using Fourier-domain phase compensation. An electrically small 200-MHz loop antenna was used as a non-ionizing excitation source to provide adequate penetration. In addition, a method was developed to generate MRI-derived contralateral breast phantoms based on enantiomorphic breast anatomy, thereby enhancing simulation realism.

Comprehensive simulations on a realistic heterogeneously dense breast model demonstrate that the proposed technique consistently detects tumors of varying sizes, depths, and locations under excitation-source and breast misalignment in the presence of AWGN with an SNR of 30 dB. The robustness of the method was further assessed under different noise levels. Quantitative positioning-error analysis confirms high localization accuracy, yielding an average positioning-error reduction of 85.75 % across the evaluated simulation scenarios.

The proposed alignment method effectively mitigates excitation-source and breast misalignment in low-frequency breast imaging, enabling non-ionizing tumor detection with accurate 2-D localization.