Empirical wavelet transform based photoacoustic spectral response technique for assessment of ex-vivo breast biopsy tissues

Abstract

Empirical wavelet transform (EWT) based Photoacoustic spectral response technique is proposed for quantitative assessment of human breast masses (e.g. normal, benign and malignant) based on tissue mechano-biological property. EWT derived parameters such as spectral magnitude and total energy of mono frequency components were utilised for quantitative analysis. Spectral magnitude of second frequency component clearly proves proliferation of epithelial cells in malignant tissues compared to normal. In addition, fibroadenoma illustrates identical energy values of both mono-components which indicates identical proliferation of epithelial and stromal cells. Subsequently, EWT features are proposed that can differentiate the breast masses with 97.5% efficiency, a significant improvement in non-invasive breast cancer diagnosis which also highly correlates with standard histopathology. © 2019 Elsevier Ltd