缺乏无派性细胞凋亡（Anoikis）抵抗是骨肉瘤（OS）转移成功发展的先决条件，无论Anoikis相关基因（ARGs）的表达是否与OS预后相关仍不清楚。本研究旨在探究将ARGs作为OS风险分层的预后工具的可行性。癌基因组图谱（TCGA）和基因表达文库（GEO）数据库提供了与OS相关的转录组信息。使用GeneCards数据库确定ARGs。通过将ARGs与GSE16088、GSE19276和GSE99671数据集中OS和正常样本之间的共同差异表达基因（DEGs）重叠，确定差异表达的ARGs（DEARGs）。通过一致聚类获得Anoikis相关患者群，完成不同群集的基因集变异分析（GSVA）。然后，使用Cox回归分析创建风险模型。评估风险分数和临床特征的独立预后价值，并构建了一个预测模型。随后，对高风险组和低风险组进行功能富集分析。另外，比较高风险组和低风险组的OS样本的免疫特征，并探索其对治疗药物的敏感性。通过将501个ARGs与68个共同DEGs重叠，得到了OS和正常样本之间的7个DEARGs。BNIP3和CXCL12在两个群集之间明显差异表达（P<0.05），并被确定为与预后相关的基因。风险模型显示，风险分数和肿瘤转移是OS患者独立的预后因素。结合风险分数和肿瘤转移的预测模型有效地预测了预后。此外，高风险组患者免疫评分低，肿瘤纯度高。高风险组中的免疫细胞浸润水平、人类白细胞抗原（HLA）基因表达、免疫应答基因集和免疫检查点较低。低风险组对免疫检查点PD-1抑制剂敏感，而高风险组对包括AG.014699、AMG.706和AZD6482在内的24种药物展示更低的50%抑制浓度值。基于BNIP3和CXCL12等ARGs的OS患者预后分层框架可能会导致更高效的临床管理。版权所有©2023年张、文、王、任和赵。

Anoikis resistance is a prerequisite for the successful development of osteosarcoma (OS) metastases, whether the expression of anoikis-related genes (ARGs) correlates with OS prognosis remains unclear. This study aimed to investigate the feasibility of using ARGs as prognostic tools for the risk stratification of OS.The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases provided transcriptome information relevant to OS. The GeneCards database was used to identify ARGs. Differentially expressed ARGs (DEARGs) were identified by overlapping ARGs with common differentially expressed genes (DEGs) between OS and normal samples from the GSE16088, GSE19276, and GSE99671 datasets. Anoikis-related clusters of patients were obtained by consistent clustering, and gene set variation analysis (GSVA) of the different clusters was completed. Next, a risk model was created using Cox regression analyses. Risk scores and clinical features were assessed for independent prognostic values, and a nomogram model was constructed. Subsequently, a functional enrichment analysis of the high- and low-risk groups was performed. In addition, the immunological characteristics of OS samples were compared between the high- and low-risk groups, and their sensitivity to therapeutic agents was explored.Seven DEARGs between OS and normal samples were obtained by intersecting 501 ARGs with 68 common DEGs. BNIP3 and CXCL12 were significantly differentially expressed between both clusters (P<0.05) and were identified as prognosis-related genes. The risk model showed that the risk score and tumor metastasis were independent prognostic factors of patients with OS. A nomogram combining risk score and tumor metastasis effectively predicted the prognosis. In addition, patients in the high-risk group had low immune scores and high tumor purity. The levels of immune cell infiltration, expression of human leukocyte antigen (HLA) genes, immune response gene sets, and immune checkpoints were lower in the high-risk group than those in the low-risk group. The low-risk group was sensitive to the immune checkpoint PD-1 inhibitor, and the high-risk group exhibited lower inhibitory concentration values by 50% for 24 drugs, including AG.014699, AMG.706, and AZD6482.The prognostic stratification framework of patients with OS based on ARGs, such as BNIP3 and CXCL12, may lead to more efficient clinical management.Copyright © 2023 Zhang, Wen, Wang, Ren and Zhao.