Evaluating Metabolomics for Prostate Cancer Diagnostics
MetadataShow full item record
Background: Prostate cancer is one of the most frequent cancers among men worldwide. The varying nature of the cancer and lack of diagnostic tools that can detect prostate cancer and separate the indolent cancers from the aggressive ones, makes it difficult to predict risk, set a diagnosis and determine the best suitable treatment for the individual patient. The malignant prostate cancer cells have altered metabolism that can give an altered metabolic phenotype. The purpose of this study was to investigate the metabolic phenotype, using Nuclear Magnetic Resonance to compare the metabolic profiles of secretions from prostate cancer cells and non-malignant prostate tissue. Methods: The samples of post prostatic-massage urine were obtained from 50 individuals (CaP = 29, controls =21). The samples were analyzed on a Bruker Avance III 600MHz/54 mm US-Plus operating at 600 MHz for proton. The resulting NMR spectra were examined by different approaches: Metabolite assignment was obtained studying NOESY- and HSQC spectra. Principal Component Analysis (PCA) was utilized to characterize small and compounded differences between the spectra. The Partial Least Squares Discriminant Analysis (PLS-DA) forces the data to confine to groups, and makes a predictive model for class discrimination. Results: From the analysis of the spectra, fifteen metabolites were assigned, whereas the identity of two of these remained uncertain. The results showed that the groups of CaP and controls could be separated with a sensitivity of 71% and a specificity of 65 %. The metabolites that contributed do these separations were Hippurate, Creatinine, Trigonelline and Trimethylamine-N-Oxide. Conclusion: The study shows that CaP patients can be separated from non-CaP controls with moderate sensitivity and specificity based on the metabolic profile of post prostatic-massage urine samples. The metabolites contributing to the results are of such a nature that one cannot be certain that they can be used as biomarkers for CaP. Further processing and studies of these data is necessary, as there may be other methods that might unveil useful aspects not yet discovered.