Department of Thoracic Surgery, Jiangmen Central Hospital, Jiangmen, Guangdong, China.
Department of Medical Oncology, The First Affiliated Hospital of University of South China, Hengyang, Hunan, China.
BMC Cancer. 2024 Sep 12;24(1):1138. doi: 10.1186/s12885-024-12911-5.
Lung adenocarcinoma (LUAD) significantly contributes to cancer-related mortality worldwide. The heterogeneity of the tumor immune microenvironment in LUAD results in varied prognoses and responses to immunotherapy among patients. Consequently, a clinical stratification algorithm is necessary and inevitable to effectively differentiate molecular features and tumor microenvironments, facilitating personalized treatment approaches.
We constructed a comprehensive single-cell transcriptional atlas using single-cell RNA sequencing data to reveal the cellular diversity of malignant epithelial cells of LUAD and identified a novel signature through a computational framework coupled with 10 machine learning algorithms. Our study further investigates the immunological characteristics and therapeutic responses associated with this prognostic signature and validates the predictive efficacy of the model across multiple independent cohorts.
We developed a six-gene prognostic model (MYO1E, FEN1, NMI, ZNF506, ALDOA, and MLLT6) using the TCGA-LUAD dataset, categorizing patients into high- and low-risk groups. This model demonstrates robust performance in predicting survival across various LUAD cohorts. We observed distinct molecular patterns and biological processes in different risk groups. Additionally, analysis of two immunotherapy cohorts (N = 317) showed that patients with a high-risk signature responded more favorably to immunotherapy compared to those in the low-risk group. Experimental validation further confirmed that MYO1E enhances the proliferation and migration of LUAD cells.
We have identified malignant cell-associated ligand-receptor subtypes in LUAD cells and developed a robust prognostic signature by thoroughly analyzing genomic, transcriptomic, and immunologic data. This study presents a novel method to assess the prognosis of patients with LUAD and provides insights into developing more effective immunotherapies.
肺腺癌（LUAD）在全球范围内显著导致癌症相关死亡。LUAD 肿瘤免疫微环境的异质性导致患者的预后和对免疫治疗的反应各不相同。因此，需要并不可避免地制定一种临床分层算法，以有效区分分子特征和肿瘤微环境，从而促进个性化治疗方法。
我们使用单细胞 RNA 测序数据构建了一个全面的单细胞转录图谱，以揭示 LUAD 恶性上皮细胞的细胞多样性，并通过结合 10 种机器学习算法的计算框架确定了一个新的特征。我们的研究进一步调查了与该预后特征相关的免疫特征和治疗反应，并在多个独立队列中验证了该模型的预测效果。
我们使用 TCGA-LUAD 数据集开发了一个由六个基因组成的预后模型（MYO1E、FEN1、NMI、ZNF506、ALDOA 和 MLLT6），将患者分为高风险和低风险组。该模型在各种 LUAD 队列中预测生存率的表现稳健。我们观察到不同风险组之间存在不同的分子模式和生物学过程。此外，对两个免疫治疗队列（N=317）的分析表明，高风险特征的患者对免疫治疗的反应优于低风险组。实验验证进一步证实，MYO1E 增强了 LUAD 细胞的增殖和迁移。
我们已经确定了 LUAD 细胞中与恶性细胞相关的配体-受体亚型，并通过彻底分析基因组、转录组和免疫学数据开发了一个稳健的预后特征。本研究提供了一种评估 LUAD 患者预后的新方法，并为开发更有效的免疫疗法提供了思路。
