Preclinical evaluation of Raman nanoparticle biodistribution for their potential use in clinical endoscopy imaging.

Raman imaging offers unsurpassed sensitivity and multiplexing capabilities. However, its limited depth of light penetration makes direct clinical translation challenging. Therefore, a more suitable way to harness its attributes in a clinical setting would be to couple Raman spectroscopy with endoscopy. The use of an accessory Raman endoscope in conjunction with topically administered tumor-targeting Raman nanoparticles during a routine colonoscopy could offer a new way to sensitively detect dysplastic lesions while circumventing Raman's limited depth of penetration and avoiding systemic toxicity. In this study, the natural biodistribution of gold surface-enhanced Raman scattering (SERS) nanoparticles is evaluated by radiolabeling them with (64) Cu and imaging their localization over time using micropositron emission tomography (PET). Mice are injected either intravenously (IV) or intrarectally (IR) with approximately 100 microcuries (μCi) (3.7 megabecquerel (MBq)) of (64) Cu-SERS nanoparticles and imaged with microPET at various time points post injection. Quantitative biodistribution data are obtained as % injected dose per gram (%ID g(-1)) from each organ, and the results correlate well with the corresponding microPET images, revealing that IV-injected mice have significantly higher uptake (p < 0.05) in the liver (5 h = 8.96% ID g(-1); 24 h = 8.27% ID g(-1)) than IR-injected mice (5 h = 0.09% ID g(-1); 24 h = 0.08% ID g(-1)). IR-injected mice show localized uptake in the large intestine (5 h = 10.37% ID g(-1); 24 h = 0.42% ID g(-1)) with minimal uptake in other organs. Raman imaging of excised tissues correlate well with biodistribution data. These results suggest that the topical application of SERS nanoparticles in the mouse colon appears to minimize their systemic distribution, thus avoiding potential toxicity and supporting the clinical translation of Raman spectroscopy as an endoscopic imaging tool.