Chen Xiaoyan, Wang Jingnan, Guo Qianqian
National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
Department of Radiation Oncology, The First Affiliated Hospital of Bengbu Medical College, Bengbu, China.
Int J Genomics. 2025 Jul 17;2025:9346050. doi: 10.1155/ijog/9346050. eCollection 2025.
Pituitary adenomas (PAs) are common intracranial tumors, and their aggressive phenotype exhibits a poor prognosis. We aimed to explore the aggressive feature of PAs and discover novel diagnostic markers. The datasets of GSE260487 and GSE169498, which contained invasive and noninvasive samples, were downloaded from the Gene Expression Omnibus (GEO) database. Aggressive phenotype-related gene modules were classified using the "WGCNA" package. Differentially expressed genes (DEGs) in each module were identified by the "limma" package. Next, a protein-protein interaction (PPI) network was used in the construction and identification process of key genes, and the CytoHubba tool was utilized to analyze the subnetwork and select the top 10 genes. Diagnostic markers were selected using two machine learning algorithms: support vector machine (SVM) and Lasso. Finally, the ESTIMATE and "GSVA" were applied for immune infiltration assessment. WGCNA showed that the turquoise module was closely associated with the aggressive phenotype and enriched in neural differentiation and cell migration pathways. A total of 521 DEGs were intersected with the turquoise module genes to obtain 187 overlapping genes, from which 10 hub genes related to tumor proliferation were selected to develop a PPI network. Next, we determined as an accurate diagnostic marker, and the immune infiltration analysis revealed that expression was negatively correlated with stromal score and immune score but positively correlated with the infiltration of antitumor cells. We developed a novel marker with a strong diagnostic performance for PAs, providing novel insights for the detection and individualized treatment of PAs.
垂体腺瘤（PAs）是常见的颅内肿瘤，其侵袭性表型预后较差。我们旨在探索垂体腺瘤的侵袭性特征并发现新的诊断标志物。从基因表达综合数据库（GEO）下载了包含侵袭性和非侵袭性样本的GSE260487和GSE169498数据集。使用“WGCNA”软件包对与侵袭性表型相关的基因模块进行分类。通过“limma”软件包识别每个模块中的差异表达基因（DEGs）。接下来，在关键基因的构建和识别过程中使用蛋白质-蛋白质相互作用（PPI）网络，并利用CytoHubba工具分析子网并选择前10个基因。使用支持向量机（SVM）和套索两种机器学习算法选择诊断标志物。最后，应用ESTIMATE和“GSVA”进行免疫浸润评估。WGCNA显示，绿松石模块与侵袭性表型密切相关，并在神经分化和细胞迁移途径中富集。共有521个DEGs与绿松石模块基因相交，获得187个重叠基因，从中选择10个与肿瘤增殖相关的枢纽基因构建PPI网络。接下来，我们确定 作为一个准确的诊断标志物，免疫浸润分析显示 表达与基质评分和免疫评分呈负相关，但与抗肿瘤细胞浸润呈正相关。我们开发了一种对垂体腺瘤具有强大诊断性能的新型标志物，为垂体腺瘤的检测和个体化治疗提供了新的见解。
