Utilizing Raman spectral imaging for distinguishing breast cancer from hyperplasia

Breast cancer is one of the most common malignant tumors among female cancer patients. The clinical diagnosis of breast cancer is mainly through biopsy, X-ray, color doppler ultrasound and nuclear magnetic resonance technology, which has a long-time examination and a high-rate misdiagnosis. In this work, we used confocal Raman microspectroscopy imaging (CRMI) to identify the spectra-pathological features of infiltrating ductal carcinoma (IDC) and the hyperplasia tissue. After point-scanned the lesion site, the obtained spectral set was reconstructed for further pathologic visualization by K-means clustering analysis (KCA). The main differences between the cancerous tissue and the hyperplasia tissue are existed at the spectral feature of lipid and protein. The peak intensities of protein at 748, 1000, 1320, 1618 cm^-1 in the cancerous tissue was higher than that in the hyperplasia tissue, whereas the protein Raman peaks at 859 cm^-1 was lower in the cancerous tissue than that in the hyperplasia tissue. While, the lipid content in the 1450, 2880, 2926 cm^-1 were lower than that in the hyperplasia tissue. The content of nucleic acid in 748 cm^-1 cancerous tissue was higher than that in hyperplasia tissue, and there was an additional peak of 1572 cm^-1 presented in hyperplasia tissue. The reconstructed pathological Raman image provided both compositional and structural information for IDC progression. The achieved results lay a foundation for understanding the pathological changes of breast cancer in vivo.