Surface enhanced Raman scattering (SERS) nanoparticles are an attractive alternative to fluorescent probes for biological labeling because of their photostability and multiplexing capabilities. However, nanoparticle size, shape, and surface properties are known to affect nanoparticle-cell interactions. Other issues such as the formation of a protein corona and antibody multivalency interfere with the labeling properties of nanoparticle-antibody conjugates. Hence, it is important to consider these aspects in order to validate such conjugates for live cell imaging applications. Using SERS nanoparticles that target HER2 and CD44 in breast cancer cells, we demonstrate labeling of fixed cells with high specificity that correlates well with fluorescent labels. However, when labeling live cells to monitor surface biomarker expression and dynamics, the nanoparticles are rapidly uptaken by the cells and become compartmentalized into different cellular regions. This behavior is in stark contrast to that of fluorescent antibody conjugates. This study highlights the impact of nanoparticle internalization and trafficking on the ability to use SERS nanoparticle-antibody conjugates to monitor cell dynamics.

As a rapid, label-free, non-destructive analytical measurement requiring little to no sample preparation, Raman spectroscopy shows great promise for liquid biopsy cancer detection and diagnosis. We carried out Raman analysis and mass spectrometry of plasma and saliva from more than 50 subjects in a cohort of head and neck cancer patients and benign controls (e.g., patients with benign oral masses). Unsupervised data models were built to assess diagnostic performance. Raman spectra collected from either biofluid provided moderate performance to discriminate cancer samples. However, by fusing together the Raman spectra of plasma and saliva for each patient, subsequent analytical models delivered an impressive sensitivity, specificity, and accuracy of 96.3%, 85.7%, and 91.7%, respectively. We further confirmed that the metabolites driving the differences in Raman spectra for our models are among the same ones that drive mass spectrometry models, unifying the two techniques and validating the underlying ability of Raman to assess metabolite composition. This study bolsters the relevance of Raman to provide additive value by probing the unique chemical compositions across biofluid sources. Ultimately, we show that a simple data augmentation routine of fusing plasma and saliva spectra provided significantly higher clinical value than either biofluid alone, pushing forward the potential of clinical translation of Raman spectroscopy for liquid biopsy cancer diagnostics.