Four subgroups by gene expression profile correlate with biological and clinical features in colorectal cancer.

(Purpose) Biological classification of colorectal cancer (CRC) can help to understand its heterogeneous background. The purpose of this study is to classify CRC based on gene expression profiles using formalin-fixed paraffin-embedded (FFPE) samples and to correlate subgroups of CRC with biological features and clinical outcomes. (Results) CRC was clustered into four subgroups by unsupervised hierarchical clustering method. These subgroups show different biological and clinical features. (Conclusion) Gene expression profiles of CRC using FFPE samples distinguish four subgroups that had different biological features and clinical outcomes. These subgroups may explain heterogeneity of CRC and be useful biomarker for clinical. Patients and Methods: One hundred patients with unresectable and advanced or recurrent CRC who underwent the surgical resection from 1998 to 2010 were enrolled in this study. RNA extracted from FFPE samples was subjected to gene expression microarray. After comprehensive gene expression analysis, CRC were classified by an unsupervised hierarchical clustering and a principle component analysis (PCA). Mutation analysis of KRAS, BRAF, PIK3CA and TP53 genes were performed by direct DNA sequencing. Correlation between the biological information, clinicopathological factors and clinical outcomes were analyzed.

【研究目的】结直肠癌（colorectal cancer, CRC）的生物学分类有助于解析其异质性背景。本研究旨在基于福尔马林固定石蜡包埋（formalin-fixed paraffin-embedded, FFPE）样本的基因表达谱对结直肠癌进行分型，并探究结直肠癌亚组与生物学特征及临床结局的相关性。【研究结果】通过无监督层级聚类法将结直肠癌划分为四个亚组，各亚组展现出不同的生物学与临床特征。【研究结论】基于FFPE样本的结直肠癌基因表达谱可区分出四个具有不同生物学特征及临床结局的亚组。这些亚组可阐释结直肠癌的异质性，有望成为具有临床应用价值的生物标志物。【患者与研究方法】本研究纳入1998年至2010年间接受手术切除的100例不可切除性晚期或复发性结直肠癌患者。从FFPE样本中提取的RNA用于基因表达微阵列检测。经全面基因表达分析后，通过无监督层级聚类与主成分分析（principal component analysis, PCA）对结直肠癌进行分型。采用直接DNA测序法对KRAS、BRAF、PIK3CA及TP53基因开展突变分析。分析生物学信息、临床病理因素与临床结局之间的相关性。