Principal Components Analysis of FT-Raman Spectra of ex vivo Basal Cell Carcinoma

摘要

FT-Raman spectroscopy is a modern analytical tool and it is believed that its use for skin cancer diagnosis will lead to several advantages for patients, e.g., faster results and a minimization of invasivity. This article reports results of an ex Vivo study of the FT-Raman spectra regarding differentiation between non-diseased and malignant human skin lesions, Basal Cell Carcinoma (BCC). A Nd: YAG laser at 1064nm was used as the excitation source in the FT-Raman, RFS 100/S Spectrometer, Bruker. Thirty-nine sets of human skin samples, 18 histopathologically diagnosed as non-diseased, and 21 as BCC, were obtained during routine therapeutic procedures required by the primary disease. No sample preparation was needed to promote the FT-Raman spectra collection. The main spectral features, which may differentiate the sample, were found in the shift region of Amide I (1640 to 1680 cm~(-1)), Amide III (1220 to 1330cm~(-1)) proteins and lipids (1400 to 1500 cm~(-1)), amino acids (939 to 940 cm~(-1)) and deoxyribonucleic acid (1600 to 1620cm ~(-1)). Principal Components Analysis (PCA) was applied to FT-Raman spectra of Basal Cell Carcinoma. Analysis was performed on mean-normalized and mean-centered data of the non-diseased skin and BCC spectra. The dynamic loading of PCA was expanded into 2D contour by calculating a variance-covariance matrix. PCA was used to verify the statistical differences in the sample. This technique applied over all samples identified tissue type within 83% of sensitivity and 100% specificity. The PCA technique proved efficient for analysis in skin tissue ex vivo, results were significant and coherent.