Discovery‐Replication Strategy Identifies Serum Metabolite Biomarkers for Colorectal Cancer in a Chinese Cohort

Early diagnosis of colorectal cancer (CRC) is difficult to achieve. The use of serum metabolites may be a noninvasive diagnostic method for CRC; however, few studies have examined this method in China. This study aimed to analyze serum metabolite alterations and their diagnostic value for CRC in a Chinese cohort. Serum metabolomics analysis was performed via liquid chromatography–mass spectrometry in a discovery cohort of 20 CRC patients and 20 healthy controls. The diagnostic value of differential serum metabolites was verified by LASSO regression and receiver operating characteristic (ROC) curve analysis. Furthermore, an independent validation cohort of 80 CRC patients and 80 controls was established. A total of 1299 serum metabolites, including 311 differentially abundant metabolites, were detected in the discovery cohort. LASSO regression revealed that 8 metabolites could distinguish CRC patients from controls: among which 2‐hydroxyhexadecanoic acid, 3‐hydroxypentadecanoic acid, 5(S)‐HETE, glutamine pyruvate, lactic acid, and lysoPC (0:0/16:0) were elevated in CRC patients, whereas DG (8:0/10:0/0:0) and cis‐muconic acid were decreased. Each of the elevated metabolites had a high diagnostic value for CRC, with an AUC exceeding 0.80. The panel of 8 metabolites was superior for diagnosing CRC, with an AUC of 0.968, a sensitivity of 0.95, and a specificity of 1.00. The validation cohort replicated the results of the discovery cohort and further confirmed that the metabolite panel was valuable for the early diagnosis and assessment of lymph node metastasis. In conclusion, our study revealed that serum metabolites could be used for CRC detection.