早期肺腺癌的预后表型
本研究对200例手术切除的早期肺腺癌患者的大组织切片进行免疫表型分析,随访中位2.34年,发现两个分子亚型:增殖型(60%)5年总生存率50%,凋亡型(40%)5年总生存率70%,风险比2.23(95%CI 1.33-3.80;p=0.0069)。这些表型在外部数据中得到验证,并提供独立于TNM分期和WHO组织学分类的预后信息。
要点
- 早期肺腺癌存在增殖型和凋亡型两个分子亚型,预后差异显著。
- 增殖型(TP53、NF1、CD45、PCNA高表达)5年生存率50%,凋亡型(TUNEL高)70%。
- 基于Cox回归和机器学习构建的模型可预测总生存和肿瘤表型。
核心结论
- 早期肺腺癌存在两个分子亚型,其识别提供重要的预后信息。
研究设计与样本
- 前瞻性队列研究,纳入200例手术切除的肺腺癌患者,进行大组织切片免疫表型分析,随访中位2.34年(95%CI 1.71-3.49)。
主要发现
- 无监督层次聚类揭示两个患者亚组,临床病理特征和基因型相似但生存显著不同。
- 增殖型患者(60%)TP53、NF1、CD45和PCNA表达升高,5年总生存率50%。
- 凋亡型患者(40%)TUNEL高表达,5年总生存率70%(风险比2.23,95%CI 1.33-3.80;p=0.0069)。
- Cox回归和机器学习算法(包括随机森林)构建了预测总生存的评分以及预测肿瘤表型的公式和列线图。
- 不同表型在TCGA和KMplotter数据中得到验证,预后能力补充于IASLC TNM分期和WHO组织学分类。
临床怎么用
- 对早期肺腺癌切除标本进行TP53、NF1、CD45、PCNA和TUNEL等免疫表型检测,可辅助识别预后不同的分子亚型。
- 基于机器学习的列线图可用于预测肿瘤表型和总生存,辅助临床决策。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
BACKGROUND: Survival after curative resection of early-stage lung adenocarcinoma (LUAD) varies and prognostic biomarkers are urgently needed. METHODS: Large-format tissue samples from a prospective cohort of 200 patients with resected LUAD were immunophenotyped for cancer hallmarks TP53, NF1, CD45, PD-1, PCNA, TUNEL and FVIII, and were followed for a median of 2.34 (95% CI 1.71-3.49) years. RESULTS: Unsupervised hierarchical clustering revealed two patient subgroups with similar clinicopathological features and genotype, but with markedly different survival: "proliferative" patients (60%) with elevated TP53, NF1, CD45 and PCNA expression had 50% 5-year overall survival, while "apoptotic" patients (40%) with high TUNEL had 70% 5-year survival (hazard ratio 2.23, 95% CI 1.33-3.80; p=0.0069). Cox regression and machine learning algorithms including random forests built clinically useful models: a score to predict overall survival and a formula and nomogram to predict tumour phenotype. The distinct LUAD phenotypes were validated in The Cancer Genome Atlas and KMplotter data, and showed prognostic power supplementary to International Association for the Study of Lung Cancer tumour-node-metastasis stage and World Health Organization histologic classification. CONCLUSIONS: Two molecular subtypes of LUAD exist and their identification provides important prognostic information.
引用
- Lamort AS, Kaiser JC, Pepe MAA, et al. Prognostic phenotypes of early-stage lung adenocarcinoma. Eur Respir J. 2022. PMID: 34887322. doi:10.1183/13993003.01674-2021
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。