BACKGROUND: Mass screening for gastric cancer (GC), particularly using endoscopy, may not be the most practical approach as a result of its high cost, lack of acceptance, and poor availability. Thus, novel markers that can be used in cost-effective diagnosis and noninvasive screening for GC are needed. METHODS: A total of 154 urine samples from GC patients and healthy individuals and 30 pairs of matched tumor and normal stomach tissues were collected. Multivariate analysis was performed on urinary and tissue metabolic profiles acquired using (1)H nuclear magnetic resonance and (1)H high-resolution magic angle spinning spectroscopy, respectively. In addition, metabolic profiling of urine from GC patients after curative surgery was performed. RESULTS: Multivariate statistical analysis showed significant separation in the urinary and tissue data of GC patients and healthy individuals. The metabolites altered in the urine of GC patients were related to amino acid and lipid metabolism, consistent with changes in GC tissue. In the external validation, the presence of GC (early or advanced) from the urine model was predicted with high accuracy, which showed much higher sensitivity than carbohydrate antigen 19-9 and carcinoembryonic antigen. Furthermore, 4-hydroxyphenylacetate, alanine, phenylacetylglycine, mannitol, glycolate, and arginine levels were significantly correlated with cancer T stage and, together with hypoxanthine level, showed a recovery tendency toward healthy controls in the postoperative samples compared to the preoperative samples. CONCLUSIONS: An urinary metabolomics approach may be useful for the effective diagnosis of GC.
BACKGROUND: Mass screening for gastric cancer (GC), particularly using endoscopy, may not be the most practical approach as a result of its high cost, lack of acceptance, and poor availability. Thus, novel markers that can be used in cost-effective diagnosis and noninvasive screening for GC are needed. METHODS: A total of 154 urine samples from GC patients and healthy individuals and 30 pairs of matched tumor and normal stomach tissues were collected. Multivariate analysis was performed on urinary and tissue metabolic profiles acquired using (1)H nuclear magnetic resonance and (1)H high-resolution magic angle spinning spectroscopy, respectively. In addition, metabolic profiling of urine from GC patients after curative surgery was performed. RESULTS: Multivariate statistical analysis showed significant separation in the urinary and tissue data of GC patients and healthy individuals. The metabolites altered in the urine of GC patients were related to amino acid and lipid metabolism, consistent with changes in GC tissue. In the external validation, the presence of GC (early or advanced) from the urine model was predicted with high accuracy, which showed much higher sensitivity than carbohydrate antigen 19-9 and carcinoembryonic antigen. Furthermore, 4-hydroxyphenylacetate, alanine, phenylacetylglycine, mannitol, glycolate, and arginine levels were significantly correlated with cancer T stage and, together with hypoxanthine level, showed a recovery tendency toward healthy controls in the postoperative samples compared to the preoperative samples. CONCLUSIONS: An urinary metabolomics approach may be useful for the effective diagnosis of GC.