The purpose of this study was to explore near-infrared (NIR) Raman spectroscopy for distinguishing cancer from normal
gastric tissue. In our study, a total of 236 Raman spectra of mucosa from 43 gastric cancer patients were obtained by NIR
Raman spectroscopy system with an excitation wavelength of 785 nm. After pretreatment, a comparison of the Raman
spectra between cancer and normal tissues occurred. It was found that the gastric cancerous mucosa showed lower
intensities at around 748, 944, and 1520cm-1, while higher at 807 and 1661cm-1, compared with normal tissue. And there
was only one peak at 1022cm-1 in the spectra of normal mucosa, while there were two peaks at 1022 and 1052cm-1 in the
spectra of cancerous mucosa. Support Vector Machine (SVM) was employed to classify Raman spectra between cancer
and normal gastric tissues. A sensitivity of 88.2%, a specificity of 91.9%, and an overall diagnostic accuracy of 90.3%
were achieved for discriminating gastric cancer from normal tissues with a Radial Basic Function (RBF) SVM algorithm.
The experimental results show that Raman spectra differed significantly between cancerous and normal gastric tissue,
which provides the experimental basis for the diagnosis of gastric cancer by Raman spectroscopy technology. And RBF
SVM algorithm can give the well generalized classification performance for the samples, which expands the application
of mathematical algorithms in the classification.