SELDI技术预测恶性肿瘤患者化疗后血糖变化的临床研究

目的应用弱阳离子芯片结合表面增强飞行时间质谱（SELDI—TOF—MS）技术筛选与恶性肿瘤化疗后血糖变化情况、相关的血清蛋白质组指纹并建立预测模型。方法应用SELDI—TOF—MS、CM10蛋白质芯片技术检测182例恶性肿瘤患者化疗前血清样本的蛋白质谱，经过2年随访按化疗后的血糖情况分为血糖正常组（136例）、糖耐量异常组（27例）和糖尿病组（19例），利用Biomarker或Wizard软件回顾性地分析比较各组间化疗前的血清蛋白质指纹图谱，Biomarker Pattern软件建立预测模型。结果M／Z为5298和9608的两个蛋白质组成的诊断模型可将患者在化疗前准确分为糖尿病组与糖耐量异常组，灵敏度和特异度分别为81．481％（22／27）和100．00％（17／17），准确度为88．64％（39／44）；M／Z为10324、2761和4084的3个蛋白质组成的诊断模型可将糖尿病组与血糖正常组准确分组，灵敏度和特异度分别为62．35％（53／85）和88．24％（15／17），准确度为66．67％（68／102）；M／Z为5895、6010、6099、3930、5430和2495的6个蛋白质组成的诊断模型可将糖耐量异常组与血糖正常组准确分组，灵敏度和特异度分别为77．65％（66／85）和96．30％（26／27），准确度为82．14％（92／112）。结论SELDI—TOF—MS技术筛选出恶性肿瘤化疗后3组血糖情况的化疗前的蛋白质指纹，建立血糖正常组、糖耐量异常组和糖尿病组的诊断模型，用于肿瘤患者化疗后血糖变化的早期预测。

A clinical research about predicting the changing of malignant tumor patients serum glucose after chemotherapy by SELDI technology

Objective By surface-enhanced laser desorption / ionization time-of-flight mass spectrometry(SELDI-TOF-MS), the serum pmteomic fingerprints related with the changing of malignant tumor patients' serum glucose after chemotherapy was selected and constructed as an predictive model. Methods By SELDI-TOF-MS, the serum of 182 malignant tumor patients who had received chemotherapy were tested, and the pmteomic fingerprints were received. After 2 years follow-up, all the patients were divided into 3 groups: the euglycemia group(136 people), the carbohydrate tolerance abnormality group(27 people), and the diabetes mellitus group (19 people). The proteomic fingerprints were analyzed by Biomarker Wizard Software and the idio-proteomic fingerprint of protective models were constructed by BPS (biomarker pattern software). Results The diagnosis model composed with 2 proteins (M/Z values were 5298 and 9608) could classify the carbohydrate tolerance abnormality group, and the diabetes mellitus group correctly. In the test model, the sensitivity and specificity were 81.48 %(22/27) and 100.00 %(17/17) respectively, the accuracy was 88.64 % (39/44). The diagnosis model composed with 3 proteins (M/Z values were 10324, 2761 and 4084) could classify the diabetes mellitus group and the euglycemia group correctly. In the test model, the sensitivity and specificity were 62.35 %(53/85) and 88.24 %(15/17) respectively, the accuracy was 66.67 %(68/102). The diagnosis model composed with 6 proteins (M/Z values were 5895,6010,6099,3930,5430 and 2495) could classify the diabetes mellitus group and the the carbohydrate tolerance abnormality group correctly. In the test model, the sensitivity and specificity were 77.65 %(66/85) and 96.30 %(26/27) respectively, the accuracy was 82.14 %(92/112). Conclusion SELDI-TOF-MS could be utilized to analyze protein profiling in screening serum glucose changing-related biomarkers and developing diagnostic and predictive patterns, and the developed patterns may be used to predict the changing of serum glucose after chemotherapy in malignant tumor patients.