文章：

确定结直肠腺癌分级的定性转录标记

A qualitative transcriptional signature for determining the grade of colorectal adenocarcinoma

原文发布日期：2019-10-09 

英文摘要：

Histological grading (HG) is an important prognostic factor of colorectal adenocarcinoma (CRAC): the high-grade CRAC patients have poorer prognosis after tumor resection. Especially, the high-grade stage II CRAC patients are recommended to receive adjuvant chemotherapy. Due to the subjective nature of HG assessment, it is difficult to achieve consistency among pathologists, which brings patients uncertain grading outcomes and inappropriate treatments. We developed a qualitative transcriptional signature based on the within-sample relative expression orderings (REOs) of gene pairs to discriminate high-grade and low-grade CRAC. Using the stage II–III CRAC samples, we detected gene pairs with stable REOs in the  high-grade samples and reversal stable REOs in the low-grade samples, and retained the gene pairs whose specific REO patterns were significantly associated with the disease-free survival of patients by univariate Cox regression model. Then, we used a forward-backward searching procedure to extract gene pairs with the highest concordance index as the final grading signature. Finally, 9 gene pairs (9-GPS) were developed to divide CRAC patients into high-grade and low-grade groups. With the signature, there were more differential expression characteristics between reclassified high-grade and low-grade groups. Significant difference of prognosis between the classified two group patients could be seen in four independent datasets. Additionally, genomic analyses showed that the classified high-grade groups were characterized by hypermutation while classified low-grade groups were characterized by frequent copy number alternations. In conclusion, the 9-GPS can provide an objective and robust grading assessment for CRAC patients, which could assist clinical treatment decision.

摘要翻译： 

组织学分级（HG）是结直肠腺癌（CRAC）的重要预后因素：高级别CRAC患者在肿瘤切除后预后较差。特别是，临床建议II期高级别CRAC患者接受辅助化疗。由于HG评估存在主观性，病理学家难以达成一致性判断，这导致患者面临分级结果不确定和治疗方案不当的风险。我们基于基因对在样本内的相对表达排序（REOs）开发了一种定性转录特征，用于区分高级别和低级别CRAC。通过分析II-III期CRAC样本，我们筛选出在高级别样本中具有稳定REOs而在低级别样本中呈现反向稳定REOs的基因对，并采用单变量Cox回归模型保留其特定REO模式与患者无病生存期显著相关的基因对。随后，我们通过前向-后向搜索程序提取具有最高一致性指数的基因对作为最终分级特征。最终构建的9个基因对（9-GPS）可将CRAC患者划分为高级别和低级别组。该特征使重分类后的高低级别组间呈现更显著的表达差异特征，并在四个独立数据集中显示出分组患者预后的显著差异。基因组分析表明，分类后的高级别组以高突变率为特征，而低级别组则表现为频繁的拷贝数变异。综上所述，9-GPS能为CRAC患者提供客观可靠的分级评估，有助于临床治疗决策的制定。

原文链接：

A qualitative transcriptional signature for determining the grade of colorectal adenocarcinoma