Classification of normal and malignant human gastric mucosa tissue with confocal Raman microspectroscopy and wavelet analysis

Thirty-two samples from the human gastric mucosa tissue, including 13 normal and 19 malignant tissue samples were measured by confocal Raman microspectroscopy. The low signal-to-background ratio spectra from human gastric mucosa tissues were obtained by this technique without any sample preparation. Raman spectral interferences include a broad featureless sloping background due to fluorescence and noise. They mask most Raman spectral feature and lead to problems with precision and quantitation of the original spectral information. A preprocessed algorithm based on wavelet analysis was used to reduce noise and eliminate background/baseline of Raman spectra. Comparing preprocessed spectra of malignant gastric mucosa tissues with those of counterpart normal ones, there were obvious spectral changes, including intensity increase at similar to 1156 cm(-1) and intensity decrease at similar to 1587 cm(-1). The quantitative criterion based upon the intensity ratio of the similar to 1156 and similar to 1587 cm(-1) was extracted for classification of the normal and malignant gastric mucosa tissue samples. This could result in a new diagnostic method, which would assist the early diagnosis of gastric cancer. (C) 2007 Elsevier B.V. All rights reserved.