Figure 1
Statistical partial-least square discriminant analysis model used for the blind test analysis of colon biopsies. The model was developed and validated using a cohort of 74 colorectal biopsies (control n = 35, adenocarcinoma n = 39) and histopathological criteria using the spectral region corresponding to the taurine, glutamate, aspartate, myo-inositol and glucose resonances. Cancerous and control colon biopsy specimens are represented by red and black boxes respectively. The two component partial-least square discriminant analysis model is characterized by the following parameters: R2Y = 0.80 and Q2 = 0.76. The high Q2 value indicates the high predictive capabilities for the model. Representative one-dimensional 1H Carr-Purcell-Meiboom-Gill high resolution magic angle spinning nuclear magnetic resonance spectra of control and adenocarcinoma colorectal tissues originating from the same patient are shown on the left side. Partial metabolite assignment in the 4.7 ppm to 0.5 ppm region is indicated. The numbers refer to the following metabolites: (1) isoleucine, (2) leucine, (3) valine, (4) ethanol, (9) lactate, (10) lysine, (11) alanine, (12) acetate, (13) glutamate, (14) methionine, (15) glutamine, (16) aspartic acid, (17) phosphoethanolamine, (18) choline, (19) phosphorylcholine, (20) glycerophosphocholine, (21) arginine, (22) taurine, (23) proline, (24) scyllo-inositol, (25) myo-inositol, (26) glycine, (27) threonine, (28) glycerol, (29) beta-glucose, (30) alpha-glucose, (31) serine, (32) creatine, (33) asparagine, (34) tyrosine, (35) phenylalanine, (36) ascorbic acid and (37) succinic acid.