Department of Respiratory and Critical Care Medicine, Affiliated Hospital of Guizhou Medical University, Guiyang, 550004, China.
BMC Cancer. 2020 Dec 7;20(1):1198. doi: 10.1186/s12885-020-07657-9.
Lung cancer is the most frequently diagnosed carcinoma and the leading cause of cancer-related mortality. Although molecular targeted therapy and immunotherapy have made great progress, the overall survival (OS) is still poor due to a lack of accurate and available prognostic biomarkers. Therefore, in this study we aimed to establish a multiple-gene panel predicting OS for lung adenocarcinoma.
We obtained the mRNA expression and clinical data of lung adenocarcinoma (LUAD) from TCGA database for further integrated bioinformatic analysis. Lasso regression and Cox regression were performed to establish a prognosis model based on a multi-gene panel. A nomogram based on this model was constructed. The receiver operating characteristic (ROC) curve and the Kaplan-Meier curve were used to assess the predicted capacity of the model. The prognosis value of the multi-gene panel was further validated in TCGA-LUAD patients with EGFR, KRAS and TP53 mutation and a dataset from GEO. Gene set enrichment analysis (GSEA) was performed to explore potential biological mechanisms of a novel prognostic gene signature.
A four-gene panel (including DKK1, GNG7, LDHA, MELTF) was established for LUAD prognostic indicator. The ROC curve revealed good predicted performance in both test cohort (AUC = 0.740) and validation cohort (AUC = 0.752). Each patient was calculated a risk score according to the model based on the four-gene panel. The results showed that the risk score was an independent prognostic factor, and the high-risk group had a worse OS compared with the low-risk group. The nomogram based on this model showed good prediction performance. The four-gene panel was still good predictors for OS in LUAD patients with TP53 and KRAS mutations. GSEA revealed that the four genes may be significantly related to the metabolism of genetic material, especially the regulation of cell cycle pathway.
Our study proposed a novel four-gene panel to predict the OS of LUAD, which may contribute to predicting prognosis accurately and making the clinical decisions of individual therapy for LUAD patients.
肺癌是最常见的癌种，也是癌症相关死亡的主要原因。尽管分子靶向治疗和免疫治疗取得了很大进展，但由于缺乏准确和可用的预后生物标志物，总体生存率（OS）仍然较差。因此，在这项研究中，我们旨在建立一个用于预测肺腺癌 OS 的多基因panel。
我们从 TCGA 数据库中获取了肺腺癌（LUAD）的 mRNA 表达和临床数据，以进行进一步的综合生物信息学分析。使用 Lasso 回归和 Cox 回归建立了基于多基因panel 的预后模型。基于该模型构建了列线图。使用接收者操作特征（ROC）曲线和 Kaplan-Meier 曲线评估模型的预测能力。在 TCGA-LUAD 患者中具有 EGFR、KRAS 和 TP53 突变以及 GEO 数据集的情况下进一步验证了多基因panel 的预后价值。进行基因集富集分析（GSEA）以探索新的预后基因特征的潜在生物学机制。
建立了用于 LUAD 预后指标的四个基因panel（包括 DKK1、GNG7、LDHA、MELTF）。ROC 曲线显示在测试队列（AUC=0.740）和验证队列（AUC=0.752）中均具有良好的预测性能。根据模型，每位患者根据四个基因panel 计算风险评分。结果表明，风险评分是一个独立的预后因素，高危组的 OS 明显差于低危组。基于该模型的列线图显示出良好的预测性能。该四基因panel 仍然是 LUAD 患者中 TP53 和 KRAS 突变患者 OS 的良好预测因子。GSEA 显示这四个基因可能与遗传物质的代谢，特别是细胞周期途径的调节密切相关。
我们的研究提出了一个新的四基因panel 来预测 LUAD 的 OS，这可能有助于准确预测预后，并为 LUAD 患者的个体化治疗做出临床决策。
