Label-free Raman spectral analysis of the cytotoxicity effect of carboplatin and cisplatin in human nasopharyngeal cancer cells

We have recently applied Raman spectroscopy for biochemical analysis of cell groups treated with 2.5, 5, 7.5 and 10 micromoles per liter of carboplatin and cisplatin, respectively, for apoptosis detection. The continuously decreased intensities of Raman bands assigned to the proteins and deoxyribonucleic acid can be observed as the concentrations of drugs increased, suggesting the cell damage induced by carboplatin and cisplatin. Principal component analysis combined with linear discriminate analysis and support vector machines analysis of the obtained spectral data was used to develop diagnostic algorithms. When cell groups were treated with the same concentration, both principal component analysis combined with linear discriminate analysis and support vector machines analysis executed between control group and groups treated with cisplatin, produce a lower error than those treated with carboplatin, suggesting the more cell damage induced by cisplatin. Among treated groups, an accuracy of 100% is only observed in the group treated with 10 micromoles per liter of cisplatin, demonstrating the apoptosis. Those Statistical analysis results are consistent with the biochemical and Raman spectral analysis. This study indicated that Raman spectroscopy in conjunction with principal component analysis combined with linear discriminate analysis and support vector machines analysis has great potential use in label-free detection of apoptosis in cells and label-free assessment of drug efficacy.