结合生物信息学技术及机器学习技术筛选前庭神经鞘瘤（VS）相关的生物标志物，分析其在免疫微环境中的生物学功能与机制。

从基因表达综合数据库（GEO）获取VS患者的信息（GSE56597和GSE39645），使用差异表达基因（DEGs）分析、加权基因共表达网络分析（WGCNA）获取特征基因。采用最小绝对收缩和选择算子（LASSO）回归、随机森林（RF）和Boruta算法3种机器学习策略进一步精确鉴定VS免疫微环境的生物标志物，并使用受试者工作特征（ROC）曲线评估其诊断价值。采用拟时序分析探究生物标志物在VS免疫微环境中的细胞分布和作用。

（1）DEGs分析筛选出1331个VS的DEGs。（2）WGCNA分析筛选出7个与VS相关的基因模块，共2055个候选基因。（3）从DEGs和模块核心基因中筛选出了722个重叠基因，3种机器学习策略鉴定出4个特征基因，即C10orf11、NCAM2、MLLT4、PI16。（4）NCAM2、PI16、C10orf11和MLLT4在数据集（GSE56597）的ROC曲线下面积（AUC）分别为0.982、0.918、1、1，在验证集（GSE39645）的ROC的AUC分别为1、0.968、1、1。（5）单细胞数转录组测序数据集（GSE216783）分析结果显示，VS肿瘤微环境中具有不同的细胞簇，包括髓系细胞、自然杀伤细胞、非髓鞘化施万细胞等。NCAM2基因在非髓鞘化施万细胞中的表达水平与分布密度均较高。（6）对单细胞数据中NCAM2+和NCAM2-施万细胞进行拟时序分析，结果显示NCAM2+施万细胞具有更强的分化能力，而且细胞通信分析发现NCAM2+细胞通过MIF-（CD74+CD44）和MIF-（CD74+CXCR4）受体配体对与其他细胞类型建立稳定的联系。

C10orf11、NCAM2、MLLT4和PI16是VS的生物学标志物，在免疫微环境中发挥作用并具有诊断价值。

To identify biomarkers associated with vestibular schwannomas (VS) by integrating bioinformatics and machine learning techniques, and to investigate their biological functions and mechanisms within the immune microenvironment.

VS patient data (GSE56597 and GSE39645) were obtained from the Gene Expression Omnibus (GEO) database. Differentially expressed genes (DEGs) analysis and weighted gene co-expression network analysis (WGCNA) were employed to identify feature genes. Least absolute shrinkage and selection operator (LASSO) regression, random forest (RF) and Boruta algorithm were applied to precisely identify VS immune microenvironment biomarkers. Diagnostic value was evaluated using receiver operating characteristic (ROC) curve analysis. Pseudotime trajectory analysis was performed to explore the cellular distribution and functional roles of these biomarkers.

(1) DEGs analysis identified 1331 VS-associated DEGs. (2) WGCNA screened seven gene modules related to VS, yielding 2055 candidate genes. (3) From the overlapping DEGs and module hub genes, 722 genes were selected. LASSO regression, Boruta, and RF algorithms identified four signature genes (C10orf11, NCAM2, MLLT4, PI16). (4) In dataset GSE56597, the area under the ROC curve (AUC) values for NCAM2, PI16, C10orf11, and MLLT4 were 0.982, 0.918, 1, and 1, respectively. In validation set GSE39645, the AUCs were 1, 0.968, 1, and 1. (5) Single-cell RNA sequencing (GSE216783) revealed heterogeneous cell clusters in the VS tumor microenvironment, including myeloid cells, natural killer cells, non-myelinating Schwann cells, fibroblasts, myelinating Schwann cells, circulating cells, and B cells/plasma cells, all contributing to tumor progression. NCAM2 exhibited high expression and density in non-myelinating Schwann cells. (6) Pseudotime analysis of single-cell RNA sequencing data revealed that NCAM2+ Schwann cells exhibit enhanced differentiation capacity compared to NCAM2-counterparts. Further cell-cell communication analysis and pathway enrichment analysis demonstrated that NCAM2+ Schwann cells establish robust interactions with other cell types through MIF-(CD74+CD44) and MIF-(CD74+CXCR4) ligand-receptor pairs, indicating more active intercellular communication.

C10orf11, NCAM2, MLLT4, and PI16 were identified as biomarkers of VS, demonstrating significant roles in the immune microenvironment and high diagnostic value.