Towards Non-contact Macroscopic Imaging of Multiple Cancers using Mulit-Spectral Inelastic Scattering Detection

In situ spontaneous Raman spectroscopy at sub-millimeter detection scales has been reported previously to guide tumor resection. However, single-point spectroscopy is limited in its ability to sample whole surgical margins rapidly and avoid under-sampling. A retrospective analysis of Raman spectroscopy datasets acquired on four cancer types has shown that the spectral bands around 1004, 1302, 1440, and 1657 inverted centimeters were common cancer biomarkers. Linear SVM classification models were developed to detect cancer with only 3 spectral bands and showed a sensibility and specificity close to the one obtained with the original classification. These results led to the development of a multi-spectral imaging system that has the capability to map a Raman biomarker, over a >2 cm² field-of-view. A proof-of-principle study is presented imaging the wavenumber band centered at 1440 cm-1 associated lipids and proteins in porcine tissues. Wide-field single-band imaging was achieved in ~5 seconds and preliminarily demonstrated the identification of distinct tissue regions. This study paves the way for the development of a new Raman imaging technique that is rapid, label-free, and wide-field.