应用液体蛋白芯片飞行时间质谱技术研究原发性肝癌门静脉癌栓形成相关标志物

 目的 应用液体蛋白质芯片飞行时间质谱技术研究肝癌门静脉癌栓（portal vein tumor thrombus，PVTT）形成相关标志物，并建立预测模型。方法 采取非PVTT和PVTT原发性肝癌（hepatocellular carcinoma，HCC）患者血清各20例，采用磁珠分离系统纯化、基质辅助激光解析电离飞行时间质谱（matrix-assisted laser desorption/ionization time of flight mass spectrometry，MALDI-TOF-MS）分析及ClinProTools生物信息学方法研究其血清蛋白表达谱，通过神经网络算法（statistical neural networks，SNN）建立图谱诊断模型并进行验证。结果 经数据对比分析找到39个差异有统计学意义的质荷比（m/z）峰（P<0.05），从中筛选出最具划分能力的8个质荷比峰（m/z=4 420.43、4 055.19、2 988.80、4 091.46、4 072.24、6 935.04、4 109.53和5 321.17），并以此建立HCC的PVTT诊断模型，模型识别率达到97.5%，而预测率达到92.73%。非PVTT组和PVTT组的准确性分别为100%和95%。结论 筛选得到的蛋白可能成为原发性肝癌PVTT的相关标志物，诊断模型可能对预测PVTT的形成具有重要临床意义。

Serum protein biomarkers for portal vein tumor thrombus in hepatocellular carcinorma detected by ClinProt system

Objective To screen potential protein biomarkers associated with the formation of portal vein tumor thrombus (PVTT) in hepatocellular carcinoma (HCC) patients, and to build a predictive model. Methods The serum samples from 40 HCC patients (20 with PVTT and 20 free of PVTT) were enriched by the magnetic bead and analyze by matrix-assisted laser desorption/ionization time of flight mass spectrometry (MALDI-TOF-MS) and ClinProTools, respectively. Diagnostic model was established and identified by statistical neural networks (SNN). Results Thirty-nine peaks with significant statistical differences in mass charge ratio (m/z) were found (P<0.05). Eight protein peaks (m/z=4 420.43, 4 055.19, 2 988.80, 4 091.46, 4 072.24, 6 935.04, 4 109.53 and 5 321.17) among them were selected to establish the predictive model of PVTT. Specificity and accuracy of this model was 97.5% and 92.73%, repectively. Predictive values in patients with and without PVTT were 100% and 95%. Conclusions These 8 protein peaks may be potential biomarkers of PVTT in HCC patients, and the predictive model may have great clinical significance in predicting the formation of PVTT.