In vivo NIR Raman and autofluorescence spectroscopies of skin neoplasms

In this in vivo pilot clinical trial, an estimation of NIR spectral features of malignant and benign tumors and normal skin was carried out. The human tissue spectra simultaneously including Raman spectroscopy and autofluorescence signals were registered using the portable spectroscopic equipment with 785 nm excitation laser and spectra registration in the 800-950 nm area. Spectral data were analyzed using PLS-DA method extracted the relevant information that allow for classification of skin neoplasm type. The most important spectral bands were found with the variable importance in projection analysis. We analyzed four models: basal cell carcinoma versus normal skin, basal cell carcinoma versus benign tumors, basal cell carcinoma versus melanoma and malignant tumors versus normal skin and benign neoplasms. The results of the neoplasm groups differentiating for each model were presented by using boxplot diagrams and receiving operating characteristic (ROC) curve approach. Achieved results demonstrate high accuracy of Raman spectroscopy and autofluorescence combination in NIR for the classification of malignant and benign skin lesions.