Identification of novel markers for neuroblastoma immunoclustering using machine learning

Area of Science:

• Oncology
• Immunology
• Bioinformatics

Background:

• Neuroblastoma exhibits significant heterogeneity, impacting treatment and prognosis.
• The role of the tumor immune microenvironment (TME) in neuroblastoma requires further investigation.
• There is a critical need for biomarkers to assess the TME status in neuroblastoma.

Purpose of the Study:

• To investigate the impact of the tumor immune microenvironment on neuroblastoma.
• To identify potential biomarkers for assessing the neuroblastoma TME.
• To explore the relationship between TME characteristics and clinical outcomes.

Main Methods:

• Transcriptome data for neuroblastoma were obtained from the GEO Database.
• Immunity scores were calculated using ssGSEA, and samples were clustered into high/low immunity groups.
• Machine learning algorithms (LASSO, SVM-RFE, Random Forest) were employed to screen for biomarker genes.

Main Results:

• Significant differences in clinicopathological characteristics and treatment responses were observed between high and low immunity groups.
• Six characteristic genes (BATF, CXCR3, GIMAP5, GPR18, ISG20, IGHM) were identified as potential biomarkers.
• These identified genes are associated with various immune-related pathways and immune cells within the neuroblastoma TME.

Conclusions:

• The identified genes (BATF, CXCR3, GIMAP5, GPR18, ISG20, IGHM) show promise as biomarkers for the neuroblastoma immune microenvironment.
• These biomarkers may aid in predicting therapeutic responses and guiding personalized treatment strategies for neuroblastoma.
• Further validation is warranted to establish the clinical utility of these immune microenvironment biomarkers in neuroblastoma management.