Serum metabolic characteristics associated with the deterioration of colorectal adenomas

Abstract

Colorectal cancer (CRC) can evolve from colorectal adenomas, which can be further classified into non-advanced adenomas (NAAs) and advanced adenomas (AAs) based on their clinical characteristics. Their prognoses are vastly different, with patients with NAAs having significantly lower recurrence and CRC-related mortality rates than those with AA or CRC. Although serum metabolomics has shown promise for the early diagnosis of CRC, the differences in serum metabolite composition between NAA and AA still need to be further elucidated. This study aimed to explore the mechanism of CRC occurrence and development based on the unique serum metabolic fingerprints of different stages of CRC and to discover a new non-invasive diagnostic method based on serum metabolomics. A clinical CRC progression cohort containing healthy control (NC; n = 40), NAA (n = 40), AA (n = 40), and CRC (n = 22) groups was constructed, and untargeted metabolomic analysis based on liquid chromatography/mass spectrometry was performed to analyze the serum metabolite characteristics of each group. A semi-quantitative analysis of intergroup metabolite differences was conducted, focusing on specific metabolites that differed in the NAA and AA groups. Finally, variable metabolites were selected based on least absolute shrinkage and selection operator (LASSO) regression analysis, and receiver operating characteristic curves were plotted to evaluate the efficacy of the serum metabolite-based diagnostic model in distinguishing NC/NAA populations from AA/CRC populations. Metabolomic analysis revealed significant differences in the composition of metabolites in the NC and NAA groups compared to the CRC group, whereas the serum metabolites of the AA group were similar to those of the CRC group. The levels of 33 metabolites were significantly different in the serum of AA/CRC patients compared to that of NAA patients, and their functions included glycerophospholipid, sphingolipid, and caffeine metabolism. LASSO regression analysis identified 57 differential metabolite variables between the NC/NAA and AA/CRC groups. The diagnostic model constructed using the random forest algorithm had the best discrimination effect, with areas under the curve of 1.000 (95% confidence interval [CI] 1.000-1.000) and 0.685 (95% CI 0.540-0.830) for the training and testing sets, respectively. The composition of serum metabolites is specific to the different stages of CRC development. The serum metabolite composition of patients with AAs was similar to that of patients with CRC. Auxiliary diagnostic measures based on serum metabolites have the potential to guide the follow-up and treatment of patients with adenoma.

Keywords: Colorectal adenoma; Colorectal cancer; Diagnostic model; Machine learning; Serum metabolites.

Fig. 1

The composition of serum metabolites changes during the occurrence and development of colorectal cancer. (A) Principal component analysis; (B) partial least squares discriminant analysis. NC: normal control; NAA: non-advanced adenoma; AA: advanced adenoma; CRC: colorectal cancer; PERMANOVA: permutational multivariate analysis of variance.

Fig. 2

Pathway and functional prediction of serum metabolites associated with non-advanced adenoma deterioration. (A) Functional annotation of differential metabolites based on KEGG database; (B) functional annotation of differential metabolites based on GO database.

Fig. 3

Construction of a diagnostic model for patients with advanced adenomas/colorectal cancer. (A) Receiver operating characteristic curves of the training set; (B) receiver operating characteristic curves of the test set; (C–D) Permutation importance. LR: logistic regression; SVMS: support vector machine; GBM: gradient boosting machine; NNET: neural network classifier; RF: random forest; XGB extreme gradient boosting; AUC: area under curve; CI: confidence interval.