结直肠癌并同时性肝转移化疗敏感预测模型的建立

A valuable biomarker model to predict the sensitivity of FOLFOX4 chemotherapy for colorectal cancer patients with simultaneous liver metastases

目的通过检测接受FOLFOX4方案化疗的结直肠癌并同时性肝转移（mCRC）患者的全基因表达谱，构建可预测化疗敏感性的生物标记基因模型。方法共人组30例mCRC患者，根据化疗后评估结果，将患者分为化疗敏感组及不敏感组，其中敏感组13例，不敏感组17例；应用组织芯片（Chip）技术获得结直肠癌组织的基因表达谱，采用聚类及微阵列显著性（SAM）分析法筛选出差异表达基因值IScore（d）1≥2，同时差异倍数（FoldChange）≥2或≤0．5的差异表达基因30个，根据各基因的探针信号对数比值，计算特征曲线下面积（AUC）、灵敏度、特异度等，最终筛选出若干预测能力非常显著的差异表达基因组成化疗敏感预测模型。结果7个预测靠前的正、负相关基因NKX2—3、FXYD6、TGFB111、ACTG2、ANPEP、同源盒基因（HOX）B8和KLK11组合成1个预测模型，总预测正确率为93．3％。结论将上述7个基因作为预测模型，正确率高，对FOLFOX4化疗敏感性评估具有重要指导作用。

Objective ＂To detect global gene expression in colorectal cancer patients with simulta- neous liver metastases （mCRC） and screen a valuable biomarker model to predict the sensitivity of FOL- FOX4 chemotherapy. Methods There are totally 30 cases from primary mCRC patients who have underg- one FOLFOX4 chemotherapy were collected. Their chemotherapy effects were evaluated and the patients were divided into chemotherapy sensitive group （13 cases） and non-sensitive group （17 cases）. Cancerous tissue gene expression profiles were detected by using chip technology. Two groups with differentially ex- pressed genes based on gene expression profiles were screened by cluster analysis and significance analysis of microarrays （SAM）. Thirty differentially expressed genes were screened out which had IScore（d） l~〉2 and Fold Change~〉2 or ~〈0. 5. The differentially expressed gene probe signals were transformed into log ra- tio values using Biweight algorithm to calculate area under curve （AUC）, sensitivity, and specificity. The most predictable and accurate differential expression genes would be selected to value the efficiency of chemotherapy. Results The first seven predictive ability genes, NKX2-3, FXYD6, transforming growth factor betal induced transcript 1 （TGFBII1）, ACTG2, ANPEP, KLKll and homeobox B8 （HOXB8） combined into a prediction model with the forecasting accuracy being 93.3%. Conclusion The combined seven genes prediction model may be classified as valuable and important model for assessing chemotherapy sensitivity of mCRC patients who will receive FOLFOX4 chemotherapy.