Abstract

An automated multimodal spectral imaging system based on auto-fluorescence (AF) imaging and Raman micro-spectroscopy (RMS) was develop to allow label-free molecular diagnosis of surgical resections obtained during cancer surgery. The AF images of tissue can be analysed by optimized segmentation algorithms to provide sampling points for RMS. After acquisition of Raman spectra of the tissue, the Raman spectra are analysed using a multivariate classification model to obtain full diagnosis of the resection specimen. The advantage of the multimodal approach is that the number of sampling points for RMS is reduced ~100-fold compared to raster-scanning RMS, while spatial resolution is determined by the AF images. Here, we show the feasibility of this instrument to diagnose non-melanoma skin cancer in less than 10 minutes and evaluate the potential for extending the technique to lung cancer.

© 2016 Optical Society of America