Discrimination of Breast Cancer from Normal Tissue with Raman Spectroscopy and Chemometrics

Abstract

Conventional Raman spectra of normal and cancerous breast tissues were acquired at an excitation wavelength of 785 nm and subjected to a discrimination analysis. First the spectra were pretreated with wavelet transform and polynomial fitting; next, cancerous tissue was identified by applying an adaptive local hyperplane K-nearest neighbor (ALHK) method to the pretreated spectra. The best discrimination accuracy of the ALHK method was 93.2%. In summary, normal and cancerous breast tissue were accurately distinguished by a miniature laser Raman spectrometer and the chemometrics method (ALHK), which might prove to be a portable and accessible diagnostic system.