文献速递淋巴瘤

LymphGen-Sig:整合遗传与转录状态预测弥漫大B细胞淋巴瘤治疗反应

LymphGen-Sig: Integrating Genetic and Transcriptional States to Predict Therapeutic Response in Diffuse Large B-Cell Lymphoma.
PubMed 文献 · J Clin Oncol · 2026-09-09

研究开发了基于基因表达的分类平台LGsig,利用764例配对基因组与转录组数据构建模型,并在POLARIX试验678例样本中验证。LGsig能准确识别LymphGen亚型,将未分类病例重新归类,并改善LymphGen作为生物标志物的性能,识别出从polatuzumab vedotin-R-CHP治疗中获益的亚型。

要点

  • LGsig基于294个基因的表达特征,将DLBCL分为MCDsig、BN2sig、ST2sig和EZBsig四个亚型。
  • LGsig可重新分类LymphGen未分类病例,其转录和临床特征与原分类对应亚型一致。
  • 在POLARIX试验中,LGsig识别出pola-R-CHP较R-CHOP有生存获益的独特亚型。

核心结论

  • LGsig通过整合遗传和转录维度扩展了DLBCL分子分类,能更好地指导亚型特异性治疗策略。

研究设计与样本

  • 研究基于美国国家癌症研究所/不列颠哥伦比亚癌症机构764例配对基因组和转录组数据构建LGsig模型,并在POLARIX试验678例存档样本的RNAseq中验证。

主要发现

  • LGsig仅用转录数据即可准确识别LymphGen亚型,并将分类扩展至所有先前LymphGen未分类病例。
  • LGsig重新分类的LymphGen未分类DLBCL在转录和临床特征上与其对应的LymphGen亚型一致。
  • LGsig将LymphGen A53亚型(以非整倍体和TP53改变为特征)重新归入更具生物学和治疗相关性的LGsig簇。
  • 在POLARIX研究中,LGsig改善了LymphGen作为生物标志物的性能,识别出在LymphGen分类和未分类病例中pola-R-CHP较R-CHOP有生存获益的独特DLBCL亚型。

关键数值与推荐

  • LGsig模型使用294个基因通过最近收缩质心分类器进行最优分类。
  • LGsig分类包括MCDsig、BN2sig、ST2sig和EZBsig。

临床怎么用

  • LGsig可基于转录数据对DLBCL进行分子分型,辅助识别可能从pola-R-CHP治疗中获益的患者。

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

PURPOSE: Genetic classification may advance precision medicine in diffuse large B-cell lymphoma (DLBCL), but existing tools like LymphGen (LG) are limited by complexity and incomplete classification and do not incorporate nongenetic features that affect disease biology and therapeutic outcomes. To address these limitations, we developed LG-sig (LGsig), a gene expression-based platform that classifies all DLBCLs and harmonizes both genetic and nongenetic dimensions of the disease. METHODS: LGsig was built on the distinct subtype-specific gene expression signature of each LG class using paired genomic and transcriptomic data (National Cancer Institute/British Columbia Cancer Agency; N = 764). Model development was restricted to DLBCLs classified into MYD88L265P and CD79B mutations (MCD), BCL6 translocation and NOTCH2 mutations (BN2), EZH2 mutations and BCL2 translocation (EZB), or SGK1 and TET2 mutations (ST2). Gene features were selected by differential gene expression, with 294 genes being optimal for classification using a nearest shrunken centroid classifier. LGsig classifications were designated as MCDsig, BN2sig, ST2sig, and EZBsig. The final model was applied to RNAseq from archival samples from the POLARIX trial (N = 678) to assess outcomes after polatuzumab vedotin-R-CHP (pola-R-CHP) or rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) for each LGsig subtype. RESULTS: LGsig accurately identified LG subtypes using transcriptional data alone and extended assignments to all previously LG-unclassified cases. Importantly, LG-unclassified DLBCLs reassigned by LGsig mirrored the transcriptional and clinical features of their corresponding LG counterparts, supporting their reclassification. In addition, LGsig reassigned LG A53 DLBCLs, characterized by aneuploidy and TP53 alterations, into more biologically and therapeutically relevant LGsig clusters. Finally, LGsig improved the performance of LG as a biomarker in the POLARIX study, by identifying distinct DLBCL subtypes exhibiting a survival benefit with pola-R-CHP over R-CHOP in both LG-classified and LG-unclassified cases. CONCLUSION: LGsig expands molecular classification beyond current genetic classifiers in DLBCL by integrating both genetic and transcriptional dimensions of the disease to better inform subtype-specific therapeutic strategies.

来源:J Clin Oncol,查看原文。版权归原作者所有。

引用

  1. Tumuluru S, Cooper A, Jiang Y, et al. LymphGen-Sig: Integrating Genetic and Transcriptional States to Predict Therapeutic Response in Diffuse Large B-Cell Lymphoma. J Clin Oncol. 2026. PMID: 42715516. doi:10.1200/JCO-26-00451

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。