Multiplexed Molecular Imaging of Biomarker-Targeted SERS Nanoparticles on Fresh Tissue Specimens with Channel-Compressed Spectrometry

Biomarker-targeted surface-enhanced Raman scattering (SERS) nanoparticles (NPs) have been explored as a viable option for targeting and imaging multiple cell-surface protein biomarkers of cancer. While it has been demonstrated that this Raman-encoded molecular imaging (REMI) technology may potentially be used to guide tumor-resection procedures, the REMI strategy would benefit from further improvements in imaging speed. Previous implementations of REMI have utilized 1024 spectral channels (camera pixels) in a commercial spectroscopic CCD to detect the spectral signals from multiplexed SERS NPs, a strategy that enables accurate demultiplexing of the relative concentration of each NP flavor within a mixture. Here, we investigate the ability to significantly reduce the number of spectral-collection channels while maintaining accurate imaging and demultiplexing of up to five SERS NP flavors, a strategy that offers the potential for improved imaging speed and/or detection sensitivity in future systems. This strategy was optimized by analyzing the linearity of five multiplexed flavors of SERS NPs topically applied on tissues. The accuracy of this binning approach was then validated by staining tumor xenografts and human breast tumor specimens with a mixture of five NP flavors (four targeted NPs and one untargeted NP) and performing ratiometric imaging of specific vs. nonspecific NP accumulation. We demonstrate that with channel-compressed spectrometry using as few as 16 channels, it is possible to perform REMI with five NP flavors, with <20% error, at low concentrations (<10 pM) that are relevant for clinical applications.