Resonance Raman and Raman spectroscopy for breast cancer detection.

Raman spectroscopy is a sensitive method to detect early changes of molecular _composition and structure that occur in lesions during carcinogenesis. The Raman spectra of normal, benign and cancerous breast tissues were investigated in vitro using a near-infrared (NIR) Raman system of 785 nm excitation and confocal micro resonance Raman system of 532 nm excitation. A total number of 491 Raman spectra were acquired from normal, benign and cancerous breast tissues taken from 15 patients. When the 785 nm excitation was used, the dominant peaks in the spectra were characteristic of the vibrations of proteins and lipids. The differences between the normal and cancerous breast tissues were observed in both the peak positions and the intensity ratios of the characteristic Raman peaks in the spectral region of 700-1800 cm(21). With 532 nm excitation, the resonance Raman (RR) spectra exhibited a robust pattern of peaks within the region of 500-4000 cm(21). The intensities of four distinct peaks at 1156, 1521, 2854 and 3013 cm(21) detected in the spectra collected from normal breast tissue were found to be stronger in comparison with those collected from cancerous breast tissue. The twelve dramatically enhanced characteristic peaks, including the enhanced amide II peak at 1548 cm(21) in the spectra collected from cancerous breast tissue, distinguished the cancerous tissue from the normal tissue. Principal component analysis (PCA) combined with support vector machine (SVM) analysis of the Raman and RR spectral data yielded a high performance in the classification of cancerous and benign lesions from normal breast tissue.