结直肠癌Dukes B、C、D期患者血清比较蛋白质组学研究

Comparative proteomics analysis in serum of colorectal carcinomas in Dukes B,C,D stages

目的建立用于诊断结直肠癌患者Dukes分期的分类树模型。方法用表面增强激光解吸电离飞行时间质谱(SELDI-TOF-MS)技术检测32例Dukes B期、24例Dukes C期和26例Dukes D期结直肠癌患者血清差异蛋白,用BioMaker Pattern软件在学习模式下建立用于判断结直肠癌患者Dukes分期的分类树模型,并用该模型在双盲模式下对随机选取的30例结直肠癌血清标本(Dukes B、C、D期各10例)进行检测以验证其诊断的准确性。结果在捕获的31个差异蛋白中,建立了以14个差异蛋白组成的分类树模型,该模型在学习模式下的诊断准确率为89．0%(73/ 82),在双盲模式下的诊断准确率为76．7%(23/30)。结论该分类树模型对判断结直肠癌患者Dukes分期有一定的诊断价值,可以为结直肠癌患者治疗方案的选择提供依据。

Objective To create a decision classification tree model of diagnosing the Dukes stage in colorectal carcinoma. Methods The distinct proteins in serum were detected in 32 cases of stage B,24 cases of stage C and 26 cases of stage D of colorectal carcinoma by surface-enhanced laser desorption,/ionization time-of-flight mass spectrometry （SELDI-TOF-MS） and a decision classification tree model of diagnosing the Dukes stage was created in the learning mode by BioMaker Pattern software. The accuracy of the model was tested by detecting distinct proteins of 30 cases of colorectal carcinoma （each 10 cases of Dukes B,C and D stage were randomly selected） in double-blind mode. Results Among 31 distinct proteins which were stably detected in the serum protein, a decision classification tree model composed of 14 distinct proteins was creased. In the learning mode its accuracy of diagnosis was 89.0 % （73/82） and in double-blind mode the accuracy of diagnosis was 76.7 % （23/30）. Conclusion The mode can be used to judge the Dukes stage and offer selective evidence of treatment plan in colorectal carcinoma patients.