Background: The combination of NMR spectroscopy and multivariate data analysis (MVDA), such as orthogonal partial least squares-discriminant analysis (OPLS-DA), has been collectively acknowledged as an excellent tool to investigate tissue metabolism and provide metabolite information for the diagnosis of disease, and become an important metabonomic platform for studies in biological tissues so far.
Methods: Both ex vivo high resolution magic-angle spinning1H NMR and in vitro1H NMR spectroscopy technique were synchronously employed to analyze the metabonomic characteristics of 102 lung tissues from 34 patients with lung cancer in hope to identify potential diagnostic biomarkers for malignancy detection in lung tissues.
Results: Significant elevations in the levels of lipids and lactate and significant reductions in the levels of myo-inositol and valine in the cancer tissues had been identified when compared with the adjacent non-involved tissues. Furthermore, the OPLSDA models calculated by two1H NMR spectra provided for relatively high sensitivity, specificity and good prediction accuracy in the identification of class membership regardless of the number of metabolites involved.
Conclusions: MVDA in combination with1H NMR spectra highlighted the potential of metabonomics in clinical settings so that the techniques might be further exploited for future lung cancer biomarker research or identification.
Keywords: Lung cancer; NMR spectroscopy; diagnosis; metabonomics; multivariate data analysis (MVDA).