N-glycan fingerprint predicts alpha-fetoprotein negative hepatocellular carcinoma: a large-scale multicenter study
Menée à l'aide de données du projet "The Cancer Genome Atlas" et partir d'échantillons sériques prélevés sur 670 témoins et 670 patients présentant un carcinome hépatocellulaire AFP- ou une maladie hépatique chronique, cette étude met en évidence l'intérêt de déterminer le niveau de 13 structures de N-glycanes pour diagnostiquer un carcinome hépatocellulaire AFP- et établir un pronostic
AFP-negative hepatocellular carcinoma (ANHCC) patients account for more than 30% of the whole entity of HCC patients and are easily misdiagnosed. This three-phase study was designed to find and validate new ANHCC N-glycan markers which identified from TCGA database and noninvasive detection. Differentially expressed genes (DEGs) of N-glycan biosynthesis and degradation related genes were screened from TCGA database. Serum N-glycan structure abundances were analyzed using N-glycan fingerprint (NGFP) technology. Totally 1340 participants including ANHCC, chronic liver diseases and healthy controls were enrolled after propensity score matching (PSM). The Lasso algorithm was used to select the most significant N-glycan structures abundances. Three machine learning models [Random Forest (RF), Support Vector Machine (SVM) and Logistic Regression (LR)] were used to construct the diagnostic algorithms. All 13 N-glycan structure abundances analyzed by NGFP demonstrated significant and was enrolled by Lasso. Among the three machine learning models, LR algorithm demonstrated the best diagnostic performance for identifying ANHCC in training cohort (LR: AUC: 0.842, 95%CI: 0.784-0.899; RF: AUC: 0.825, 95%CI: 0.766-0.885; SVM: AUC:0.610, 95%CI: 0.527-0.684). This LR algorithm achieved a high diagnostic performance again in the independent validation (AUC: 0.860, 95%CI: 0.824-0.897). Furthermore, the LR algorithm could stratify ANHCC into two distinct subgroups with high or low risks of overall survival and recurrence in follow-up validation. In conclusion, the biomarker panel consisting of 13 N-glycan structures abundances using the best-performing algorithm (LR) was defined and indicative as an effective tool for HCC prediction and prognosis estimate in AFP negative subjects. This article is protected by copyright. All rights reserved.
International Journal of Cancer , résumé, 2021