Abstract
BACKGROUND: Bladder urothelial carcinoma (BLCA) is a highly heterogeneous cancer with a wide range of prognoses, ranging from low-grade non-muscle-invasive bladder cancer (NMIBC), which has a good prognosis but a high recurrence rate, to high-grade muscle-invasive bladder cancer (MIBC), which has a poor prognosis. Glycosylation dysregulation plays a significant role in cancer development. Therefore, this study aimed to investigate the role of glycosyltransferases (GT)-related genes in the prognosis of BLCA and to develop a prognostic model based on these genes to predict overall survival (OS) and assess its clinical application. METHODS: The Cancer Genome Atlas (TCGA)-BLCA dataset, comprising 411 tumor and 19 normal samples. The validation set, GSE13507 from the Gene Expression Omnibus (GEO) database, included 165 primary bladder cancer samples with survival data. Differentially expressed GT-related genes (DEGRGs) in BLCA were identified in the training set. Predictive DEGRGs were used to construct risk score models by univariate Cox regression, least absolute shrinkage and selection operator (LASSO) and multivariate Cox regression. The predictive value of the models was assessed using Kaplan-Meier survival analysis and receiver operating characteristic (ROC) analysis in the training and validation sets. A nomogram was developed and its performance was evaluated with calibration curves. In addition, the relationship between the risk score and the tumor immune microenvironment was explored, and tumor immune dysfunction score (TIDE) and immune signature scores were used to predict the response to immunotherapy in BLCA patients. RESULTS: Thirty-three DEGRGs were identified in the comparison of BLCA patients with control samples. A risk score model was constructed based on 11 of these genes (GYS2, GALNTL6, GLT8D2, PYGB, B3GALNT2, GALNT15, ST6GALNAC3, ST8SIA6, CHPF, ALG9 and B3GALT2). The model performed well in predicting 3-, 5-, and 7-year overall survival (OS), with areas under the curve (AUC) of 0.65, 0.67, and 0.68, respectively. In addition, patients in the high-risk group had significantly lower survival than those in the low-risk group, and there were significant differences in immune status between the two groups. Based on age, tumor stage, T stage, and risk score, a Nomogram was constructed to predict the probability of OS, and the results of the calibration curves showed that the model had high predictive accuracy. Further analysis showed that the rejection score and TIDE were higher in the high-risk group, while the GT-related pathway was significantly upregulated in the high-risk group. CONCLUSIONS: The 11 GT-related genes identified were associated with OS in BLCA patients, suggesting that the model has potential predictive value. At the same time, further research is needed to explore its role in clinical practice.