Bladder cancer represents a frequently encountered urologic malignancy and is linked to considerable morbidity and mortality. Although immunotherapy has become an important therapeutic approach, patient responses vary widely. Altered glycosylation has been associated with cancer development and immune modulation. Nevertheless, an integrated view of how glycosylation contributes to bladder cancer progression and its clinical relevance remains insufficiently defined. A training cohort was generated using the TCGA-BLCA dataset. Additional datasets from Xiangya Hospital and public repositories (Xiangya cohort, GSE32894, GSE48075, GSE31684, GSE69795, and E-MTAB-1803) served as validation cohorts. To screen for glycosylation-associated genes linked to prognosis, we applied univariate Cox analysis followed by LASSO regression. A Cox proportional hazards model was then used to construct a risk-scoring system. Kaplan–Meier and ROC analyses evaluated the predictive performance of this score within the training set, and the model was subsequently examined across multiple external datasets.
Based on glycosylation-related gene expression, the training cohort patients were divided into two molecular categories, Cluster 1 and Cluster 2. Survival assessments showed that Cluster 2 had significantly worse outcomes. This cluster also displayed elevated immune-cell infiltration within the tumor microenvironment and stronger activation of major steps in the cancer-immunity cycle. We created an independent prognostic signature (p < 0.001) and incorporated it into a nomogram with robust predictive ability. Individuals with a high glycosylation-related risk score demonstrated increased immune infiltration, higher enrichment in immune-therapy-associated pathways, and a pattern consistent with a basal molecular phenotype. In contrast, the low-risk group showed sparse immune-cell infiltration and a luminal-type profile. These patterns were replicated in the real-world Xiangya cohort. This multi-omics glycosylation-based scoring system effectively captured tumor heterogeneity in bladder cancer, forecasted immunotherapy responsiveness and molecular subtype, and may support more precise therapeutic decision-making.
