OBJECTIVE: To analyze the metabonomic (1)H-MRS of plasma samples from patients with esophageal cancer and healthy controls applying different pattern recognition methods, and to explore the potential of application of (1)H-MR-based metabonomics in clinical research. METHODS: (1)H-MR was performed on plasma samples from 109 EC patients and 50 health controls to analyze the metabonomic variation between EC patients and healthy subjects and the corresponding (1)H-MRS were recorded on Varian Unity ANOVA 600 MHz to perform principal components analysis (PCA), partial least squares discriminant analysis (PLS-DA), orthogonal partial least squares discriminant analysis (OPLS-DA), respectively. RESULTS: OPLS-DA analysis could correctly separate almost all the plasma samples from EC patients and health controls, better than both the PCA and PLS-DA. The plasma levels of leucine, alanine, isoleucine, valine, glycoprotein, lactate, acetone, acetate, choline, isobutyrate, unsaturated lipid, VLDL, LDL, 1-methylhistidine were significantly decreased in EC patients (r total > 0.27, P < 0.05), while that of dimethylamine, α-glucose, β-glucose, citric acid increased in the EC patients (r total < -0.27, P < 0.05). CONCLUSIONS: The analysis of metabonomic (1)H-MRS of plasma samples by OPLS-DA method can eliminate the influence of non-experimental factors and decrease the heterogeneity of samples. It is useful and of great potential for application in clinical diagnosis and research of esophageal cancer.
OBJECTIVE: To analyze the metabonomic (1)H-MRS of plasma samples from patients with esophageal cancer and healthy controls applying different pattern recognition methods, and to explore the potential of application of (1)H-MR-based metabonomics in clinical research. METHODS: (1)H-MR was performed on plasma samples from 109 EC patients and 50 health controls to analyze the metabonomic variation between EC patients and healthy subjects and the corresponding (1)H-MRS were recorded on Varian Unity ANOVA 600 MHz to perform principal components analysis (PCA), partial least squares discriminant analysis (PLS-DA), orthogonal partial least squares discriminant analysis (OPLS-DA), respectively. RESULTS: OPLS-DA analysis could correctly separate almost all the plasma samples from EC patients and health controls, better than both the PCA and PLS-DA. The plasma levels of leucine, alanine, isoleucine, valine, glycoprotein, lactate, acetone, acetate, choline, isobutyrate, unsaturated lipid, VLDL, LDL, 1-methylhistidine were significantly decreased in EC patients (r total > 0.27, P < 0.05), while that of dimethylamine, α-glucose, β-glucose, citric acid increased in the EC patients (r total < -0.27, P < 0.05). CONCLUSIONS: The analysis of metabonomic (1)H-MRS of plasma samples by OPLS-DA method can eliminate the influence of non-experimental factors and decrease the heterogeneity of samples. It is useful and of great potential for application in clinical diagnosis and research of esophageal cancer.