Wang Xia, Wang Zhiming
Department of Stomatology, Shengjing Hospital of China Medical University, Shenyang, China.
Medicine (Baltimore). 2025 Jan 17;104(3):e41316. doi: 10.1097/MD.0000000000041316.
In recent years, a notably heterogeneous malignant tumor, squamous cell carcinoma of the head and neck (HNSCC), has received increasing attention, with no significant improvement in its survival rate. The rapid increase in the number of prognostic models associated with HNSCC has been observed due to its accuracy, which offers crucial clinical benefits. The 10 genes were selected from 222 human genes associated with the basement membrane in the analysis of this article. The gene pool was narrowed through different classifications and intersections, followed by univariate Cox regression analysis. Genes with statistical significance underwent further Least Absolute Shrinkage and Selection Operator (LASSO) regression analysis, resulting in the final selection of 10 genes. The data and images extracted from the Human Protein Atlas database were utilized to confirm the differential expression of the corresponding genes. Multivariate Cox regression analysis was employed to develop a nomogram, and the nomogram was assessed by additional decision curve analysis (DCA). The Gene Expression Omnibus validation set was used to validate the established model. Finally, between the high- and low-risk score groups, Gene Set Enrichment Analysis, immune correlation analysis, and drug sensitivity analysis were conducted in this paper. ITGA5, SPOCK1, EVA1C, TINAGL1, LAMB4, ADAMTS1, EGFL6, GPC2, BGN, and ITGA2B were successfully developed as basement membrane-associated risk models. The time-dependent receiver operating characteristic (timeROC) curve illustrated that the risk score prediction accuracy outperformed indicators, which were commonly adopted in clinical practice, consisting of age, stage, gender, T-staging, and N-staging. The 3-year risk score timeROC area under the curve value was 0.679. This model demonstrates a reliable ability to assess the prognosis of HNSCC patients. In addition, the specific potential biomarkers associated with the basement membrane were explored in this research.
近年来，一种显著异质性的恶性肿瘤——头颈部鳞状细胞癌（HNSCC）受到了越来越多的关注，但其生存率并无显著改善。由于其准确性带来了关键的临床益处，与HNSCC相关的预后模型数量迅速增加。在本文分析中，从222个与基底膜相关的人类基因中选择了10个基因。通过不同的分类和交叉对基因库进行筛选，随后进行单变量Cox回归分析。具有统计学意义的基因进一步进行最小绝对收缩和选择算子（LASSO）回归分析，最终选择了10个基因。利用从人类蛋白质图谱数据库中提取的数据和图像来确认相应基因的差异表达。采用多变量Cox回归分析来构建列线图，并通过额外的决策曲线分析（DCA）对列线图进行评估。使用基因表达综合数据库验证集来验证所建立的模型。最后，在高风险和低风险评分组之间进行了基因集富集分析、免疫相关性分析和药物敏感性分析。成功构建了整合素α5（ITGA5）、富含半胱氨酸的分泌蛋白1（SPOCK1）、EVA1同源物C（EVA1C）、含tinag样结构域1（TINAGL1）、层粘连蛋白β4（LAMB4）、含血小板反应蛋白基序的解聚蛋白样金属蛋白酶1（ADAMTS1）、表皮生长因子样结构域6（EGFL6）、糖蛋白聚糖2（GPC2）、核心蛋白聚糖（BGN）和整合素α2β1（ITGA2B）作为与基底膜相关的风险模型。时间依赖性受试者工作特征（timeROC）曲线表明，风险评分预测准确性优于临床实践中常用的指标，这些指标包括年龄、分期、性别、T分期和N分期。3年风险评分的timeROC曲线下面积值为0.679。该模型显示出评估HNSCC患者预后的可靠能力。此外，本研究还探索了与基底膜相关的特定潜在生物标志物。
