ABSTRACT
<p><b>OBJECTIVE</b>To investigate the Raman spectral characteristics of the pathological lip minor salivary glands affected in primary Sjögren's syndrome.</p><p><b>METHODS</b>Thirty pathological samples and 30 normal samples were collected in this study. The samples were examined by Raman microscope.Support vector machine(SVM) was employed to analyze the data and establish the classification model.</p><p><b>RESULTS</b>The spectra of pathological tissues was different from the controls in proteins, nucleic acids, lipids and glycogen skeleton. The sensitivity, specificity and accuracy of the model established by SVM on the training sets were all 92.0% (92/100), but the sensitivity, specificity and accuracy of the model established by SVM on the testing sets were 69.2% (37/53), 100.0% (37/37) and 82.0% (73/89) respectively.</p><p><b>CONCLUSIONS</b>There was significant difference in Raman spectra between the pathological and normal lip salivary glands, and the classification model established by SVM could discriminate the pathological glands from the normal ones.</p>