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 ∼1156 cm−1 and intensity decrease at ∼1587 cm−1. The quantitative criterion based upon the intensity ratio of the ∼1156 and ∼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.