Label-free diagnosis of breast cancer based on serum protein purification assisted surface-enhanced Raman spectroscopy

Serum protein is generally used to assess the severity of disease, as well as cancer progression and prognosis. Herein, a simple and rapid serum proteins analysis method combined with surface-enhanced Raman spectroscopy (SERS) technology was applied for breast cancer detection. The cellulose acetate membrane (CA) was employed to extract human serum proteins from 30 breast cancer patients and 45 healthy volunteers and then extracted proteins were mixed with silver nanoparticles for SERS measurement. Additionally, we also mainly assessed the use of different ratios of proteins-silver nanoparticles (Ag NPs) mixture to generate maximum SERS signal for clinical samples detection. Two multivariate statistical analyses, principal component analysis-linear discriminate analysis (PCA-LDA) and partial least square-support vector machines (PLS-SVM) were used to analyze the obtained serum protein SERS spectra and establish the diagnostic model. The results demonstrate that the PLS-SVM model provides superior performance in the classification of breast cancer diagnosis compared with PCA-LDA. This exploratory work demonstrates that the label-free SERS analysis technique combined with CA membrane purified serum proteins has great potential for breast cancer diagnosis.