[2.01] A comparative Raman and CARS imaging study of colon tissue

Abstract

Objective This paper presents an experimental evaluation of the information content, which can be obtained from using the complementary strengths of Raman and Coherent anti-Stokes Raman (CARS) microspectroscopy. Material and methods CARS images were recorded from thin colon tissue sections at various relative wave number detunings and compared with Raman images. Raman images were analyzed by chemical mapping, representing a univariate algorithm and by k -means clustering, representing a multivariate algorithm. The colon-tissue samples used in this study were obtained by resection from patients undergoing surgery (diverticulitis, colorectal carcinoma, Crohn's disease), and frozen at −80 °C without fixative or tissue embedding media. 10 μm-tissue sections were prepared using a cryotome, transferred onto quartz slides for Raman spectroscopy, and dried. The study protocol was approved by the local ethics commission. Results CARS images were recorded at different Stokes shifts allowing for selective protein and lipid bands to be probed resonantly. All CARS images correlate well with the photomicrographs because the non-resonant signals provided significant morphological information. The chemical contrast can be enhanced by compensation of the non-resonant background. All CARS images also correlate well with the Raman images. The amount of time of acquisition of Raman images was incomparably greater (by a factor of 1000) than for the acquisition of CARS images. The resulting Raman images probed only small areas in the 20×20 μm 2 range. However, complete spectral information could be extracted from Raman images. Extended spectral ranges are required to distinguish different tissue types in colon pathologies such as inflammation or colorectal cancer. Conclusion Based on the current results we suggest a complementary application of Raman and CARS imaging. Raman imaging defines spectral regions and spectral markers that are essential for tissue classification. CARS imaging at different Stokes shifts or in the multiplex mode, probes these spectral descriptors at video-rate speed. Such a combination offers very good prospects for clinical diagnoses.