Objective To build protein fingerprint models of early esophageal cancer and advanced esophageal cancer in high incidence area, and investigate their screening values. Methods The serum proteomic patterns of 38 cases of normal controls, 15 cases of early esophageal cancer and 36 cases of advanced esophageal cancer were detected by using surface enhanced laser desorption/ionization time of flight mass spectrometry ( SELDI - TOF - MS ). All subjects were from endoscopic screening population in high - risk areas of esophageal cancer. The data was analyzed and the protein fingerprint models were established by using support vector machine. The models were validated by leave one cross validation.Results The protein fingerprint model could differentiate early esophageal cancer from normal controls with a specificity of 81. 58% and a sensitivity of 80. 00%. The model could differentiate advanced esophageal carcinoma from normal controls with a specificity of 89.47% and a sensitivity of 83.33%. The model could differentiate early esophageal cancer from advanced esophageal cancer with a specificity of 73.33% and a sensitivity of 91.67%. The difference of protein expression peaks (4 291,4 392,4 823,4 975,5 644, 5 665,5 932, 8 473,8 776, 8 910Da ) was statistically significant between groups.Conclusion Establishment of three protein fingerprint models provides a new approach for diagnosing and screening early esophageal cancer in high incidence area. The protein peaks which had statistical significance in diagnostic models may be the potential biomarkers related to early esophageal cancer.%目的 建立高发区早期、进展期食管癌的蛋白指纹图谱模型并探究筛查价值.方法 使用飞行质谱技术检测高发区内镜筛查的38例正常对照、15例早期食管癌和36例进展期食管癌患者的血清蛋白质谱,支持向量机分别建立蛋白指纹图谱模型,留一法交叉验证判别效果.结果 区分早期食管癌和正常对照的蛋白指纹图谱模型特异性为81.58%,敏感性为80.00%;区分进展期食管癌和正常对照的模型特异性为89.47%,敏感性为83 33%;区分早期食管癌和进展期食管癌的模型特异性为73.33%,敏感性为91.67%.质荷比峰4 291、4 392、4 823、4 975、5 644、5 665、5 932、8 473、8 776、8 910Da的表达量组间比较差异有统计学意义(P<0.05,P<0.01).结论 3个蛋白指纹图谱模型的建立,为筛查高发区早期食管癌提供了一种新方法;构建模型的质荷比峰可能是早期食管癌的潜在生物学标记物.
展开▼