Neuroblastoma (NB), recognized as the most prevalent solid tumor affecting children, is marked by a significant degree of heterogeneity and a notably poor prognosis, especially among those classified as high-risk patients. The existing diagnostic and treatment modalities currently available demonstrate limited efficacy for this particular group of patients, highlighting a critical need for more effective strategies. This study is designed to delve into the intricate molecular mechanisms that underpin NB, with a specific emphasis on the role of telomere-related genes (TRGs). By doing so, we aim to provide theoretical foundations and new predictive tools for precision medicine.

Using datasets from multiple cohorts, including bulk and single-cell sequencing from Gene Expression Omnibus (GEO), we conducted an in-depth multi-omics study to uncover the function of TRGs in NB. Differential genes were screened and intersected with 2089 TRGs in TelNet database, resulting in the identification of 103 telomere-related differentially expressed genes (TRDEGs) exhibiting survival differences. Through consensus clustering, we discovered distinct molecular subtypes linked to TRDEGs in NB, and further investigated them in the context of immune cell infiltration and pathway analysis. We developed the TRDEGs signature incorporating 8 TRGs, a prognostic model created using a combination of machine learning algorithms, and assessed its predictive power for patient outcomes and responsiveness to treatment. The TRDEGs signature expression level was confirmed using RT-PCR. Additionally, the functional significance of PSAT1 in cancer progression was further verified through cell viability assay, wound healing assay, and Annexin V/PI staining assay.

We successfully identified two distinct molecular subtypes of NB through consensus clustering linked to TRDEGs, ach associated with unique biological processes and varying prognostic outcomes. Furthermore, we developed a prognostic risk model for NB that incorporates eight specific TRDEGs—namely ARHGAP23, CHD5, E2F3, ELOVL6, FEN1, GMPS, LRR1, and PSAT1. This model demonstrated impressive predictive performance, achieving area under the curve (AUC) values of 0.885, 0.903, and 0.911 for predicting survival at 1, 3, and 5 years, respectively. The robustness of our findings was further validated in multiple independent cohorts, reinforcing the reliability of our model. Moreover, Multivariate Cox regression analysis identified risk score (HR=2.933, 95% CI: 2.127–4.045, P <0.001) as an independent risk factor for NB patients. The predictive nomogram that we constructed, which integrates TRDEGs riskscore with various clinical factors, significantly enhanced prognostic accuracy when compared to traditional risk stratification methods. High-risk patients exhibited reduced immune cell infiltration and heightened immune evasion markers, correlating with poor immunotherapy response. We also observed differences in chemotherapeutic drug sensitivity across different TRDEGs risk subtypes which further confirmed our model’s predictive efficacy for chemotherapy. Meanwhile, PSAT1 was validated as a potential therapeutic target promoting tumor cell proliferation and migration.

Our study emphasizes the significance of TRGs in the outlook of NB. The findings contributed to a better understanding of the link between TRGs and prognosis in NB, and identified potential therapeutic targets for patients with NB. Future research should focus on larger clinical validations and functional studies to further elucidate the role of telomere-related molecular markers in the precision medicine landscape for NB, ultimately aiming to improve patient outcomes.