中文摘要：

目的利用多重定量抗体芯片同时检测肝细胞癌（HCC）及肝硬化患者血清中肿瘤相关血清学标志物的水平，建立HCC早期诊断模型，以提高HCC早期诊断的准确度。方法运用双抗体夹心原理建立了相关指标的多重定量抗体芯片检测系统，8种肿瘤相关血清学标志物作为芯片检测目标：甲胎蛋白（AFP）和7种细胞因子包括肝细胞生长因子、胰岛素样生长因子、白细胞介素6、白细胞介素8、白细胞介素10、转化生长因子β1和血管内皮生长因子。应用该系统检测临床确诊的160例HCC和58例肝硬化（LC）患者血清，随机抽取其中60%为训练集（HCC96人；LC36人），40%为测试集（HCC64人；LC22人），通过对结果和临床资料进行回顾性研究，应用SPSS软件做logistic回归分析，利用训练集创建诊断模型，获得受试者工作曲线（ROC曲线下面积以及cutoff值，并在测试集中验证模型的诊断价值。结果AFP联合7种细胞因子在训练集中诊断的敏感度为93．3％，特异度为83．3％，准确度为90．9%；而在测试集中敏感度为89％，特异度为77．3％，准确度为86%，传统血清AFP值（cut off值为20ng／ml）诊断的敏感度为70%，特异度为59％，准确度为64%。结论多重定量抗体芯片检测系统具有较高敏感度和特异度，AFP联合7种细胞因子诊断模型对肝癌早期诊断优于传统AFP，具有潜在的临床应用价值。

英文摘要：

Objective To develop an early and accurate detection method for hepatocellular carcinoma （HCC） based on detection of tumor-associated serum markers using a multiplex quantitative antibody array. Methods The double-antibody sandwich principle was used to establish an antibody array composed of eight cancer-related serum markers, including alpha-fetoprotein （AFP）, hepatocyte growth factor （HGF）, insulin-like growth factor （IGF）, interleukin-6 （IL-6）, interleukin-8 （IL-8）, interleukin-10 （IL-10）, transforming growth factor-beta 1 （TGF-β1）, and vascular endothelial growth factor （VEGF）. Serum samples from 160 cases of clinically diagnosed HCC and from 58 cases of liver cirrhosis （LC; controls） were obtained to test the array. Sixty percent of the samples were randomly selected for use as the traiming set （HCC, n = 96; LC, n = 36）, and the remaining 40% was used as the test set （HCC, n = 64; LC, n = 22）. The SPSS statistical software was used to perform logistic regression analysis and to create a diagnostic model. Results When used with the training set, the model had sensitivity of 93.3%, specificity of 83.3%, and accuracy of 90.9%. When used with the test set, the model had sensitivity of 89.0%, specificity of 77.3%, and accuracy of 86.0%. The traditional serum AFP value （cut-off value of 20 ng/mL） showed 70.0% diagnostic sensitivity, 59.0% specificity, and 64.0% accuracy.Conclusion The newly developed multiplex quantitative antibody detection system has high sensitivity and specificity. The diagnostic model with AFP and seven other cancer-related factors was superior to the traditional AFP only approach for early diagnosis of liver cancer, indicating its potential clinical value.