基于质谱技术应用CLINPROT系统建立乳腺癌诊断模型

Establishment of diagnostic model of breast cancer by using CLINPROT system based on MALDI-TOF/TOF MS

目的应用液体蛋白芯片-飞行时间质谱技术(CLINPROT)系统筛选乳腺肿瘤患者血清中潜在的肿瘤标志物,并基于差异蛋白建立乳腺癌诊断预测模型。方法收集病理确诊的乳腺癌患者血清标本43例及健康女性血清标本24例,并随机分为建模组和验证组。应用ClinProt Tools系统对建模组中的乳腺癌患者和健康女性血清进行差异蛋白的筛选,建立差异蛋白质图谱,应用ClinProTools v3.0软件建立乳腺癌诊断预测模型,而后利用该诊断模型对验证组血清标本进行分组验证,初步评估该模型的诊断效能。结果通过分析病例组和对照组质谱峰信息,得到24个具有明显差异的蛋白峰(P<0.05)。根据遗传算法选取表达差异最显著的3个蛋白质建立乳腺癌诊断预测模型,该模型对建模组乳腺癌的识别能力为91.67%。后利用验证组标本对该模型的诊断效能进行评估,结果显示其对验证组乳腺癌识别准确率为85.7%。结论利用ClinProt Tools v3.0软件成功建立并评估了乳腺癌诊断预测模型,为乳腺癌诊断提供了新的途径。

Objective To use the liquid protein chip-MALDI-TOF/TOF MS CLINPROT system for screening the serum potential tumor markers in the patients with breast tumor and to establish the breast cancer diagnostic prediction model basic on differential protein.Methods Forty-three serum samples from the patients with pathologically diagnosed breast cancer and 24 serum samples from healthy female controls were collected and randomly divided into the construction model group and verification group.The differential protein screening in the breast cancer patients of the constructing model group and healthy female serum was performed by using the CLINPROT system.The differential protein atlas was established.The breast cancer diagnotic prediction model was established by using the Clin Prot Tools v3.0software.Then the diagnostic model was applied to perform the grouping verification in the serum samples of verification group.The diagnostic efficiency of this model was preliminarily evaluated.Results Twentyfour protein peaks with significant difference were obtained by analyzing the mass spectrum peak information of the patients group and control group（P〈0.05）.Three proteins expressing the difference most significantly were selected to build a breast cancer diagnostic prediction model according to genetic algorithm,and the recognition ability of this model for breast cancer in the construction model group was 91.67%.Then the diagnostic efficiency of this model was evaluated by using the samples of verification group,and the results showed that its recognition accuracy rate to breast cancer in the verification group was 85.7%.Conclusion Using the ClinProt Tools v3.0software successfully constructs and evaluates the diagnostic prediction model of breast cancer,which provides a new pathway for the diagnosis of breast cancer.