卵巢癌脂质组学诊断标志物的筛选及其效果研究  

Identification and validation of differential plasma diagnostic biomarkers of ovarian cancer based on lipidomics

作　　者：谢宏宇[1] 王文杰[1] 李贞子[1] 夏白荣 李康[1] 侯艳[1] XIE Hong-yu WANG Wen-jie LI Zhen-zi XIA Bai-rong LI Kang HOUYan(Department of Medical Statistics, Harbin Medical University, Harbin 150081 ,P. R. China Department of Gynecologic Oncology, Tumor Hospita , Harbin Medical University, Harbin 150081 ,P. R. China)

机构地区：[1]哈尔滨医科大学卫生统计学教研室,黑龙江哈尔滨150081 [2]哈尔滨医科大学附属肿瘤医院妇科肿瘤科室,黑龙江哈尔滨150081

出　　处：《中华肿瘤防治杂志》2016年第23期1550-1555,共6页Chinese Journal of Cancer Prevention and Treatment

基　　金：国家自然科学基金(81573256;81473072;81302511);黑龙江省自然科学基金青年项目(QC2015098)

摘　　要：目的卵巢癌是女性生殖器官常见的恶性肿瘤之一,90%的卵巢癌患者诊断时已进展到晚期(Ⅲ/Ⅳ期)。尽管治疗方式的不断改进,但晚期卵巢癌患者5年总生存率为20%,而早期卵巢癌患者5年总生存率>90%。本研究旨在识别卵巢癌脂质组学诊断生物标志物,提高卵巢癌早期诊断的准确性。方法采用稀疏组lasso变量筛选方法和单因素分析相结合的方法,对2011-03-01-2013-07-31哈尔滨医科大学附属肿瘤医院的139例卵巢癌患者和76例对照患者的血浆脂质组学数据进行分析,筛选可用于诊断卵巢癌脂质组学的生物标志物,在差异代谢物中选择与CA125相关性小的物质作为最终诊断标志物,通过七折交叉验证方法来评价其预测效果及与CA125联合的诊断准确性。进一步采用cytoscape软件研究差异脂质物质间的相互作用。通过Wilcoxon秩和检验方法筛选出能够区分早晚期卵巢癌的生物标志物。结果共筛选出20种可用于区分卵巢癌和对照的差异脂质生物标志物,其中Stearamide、Stearic acid、Arachidic acid和PI(42∶9)与CA125不相关(P<0.05),这4个脂质与CA125联合AUC值为0.94,大于CA125单独的诊断性能。另外,其中8个差异脂质在早期与晚期上皮性卵巢癌患者中有差异,分别为PC(35∶4)、PC(38∶6)、PC(46∶4)、PC(P-35∶2)、PE(P-36∶6)、PG(34∶2)、Cer(d18∶1/16∶0)和3-Deoxyvitamin D3,均P值<0.05。结论卵巢癌血浆脂质组学筛选出的物质提高了CA125单独诊断的准确性,其中8个差异物质可以作为卵巢癌早期诊断潜在的生物标志物。OBJECTIVE Ovarian cancer is one of the most common gynecologic maligancies.About 90% of ovarian cancer patients were diagnosed at an advanced stage（Ⅲ/Ⅳ）.Despite improvements in mulitimodality treatment,the5-year overall survival for advanced ovarian cancer is less than 20%,while patients recognized at early stage（Ⅰ/Ⅱ）,the5-year survival rates are more than 90%.This study aimed to identify the differential plasma lipids biomarkers to improve the accuracy of ovarian carcinoma diagnosis.METHODS We combined the results from multivariate sparse-group lasso method and univariate Wilcoxon rank sum test to identify the potentially differential lipids biomarkers of the plasma from139 ovarian carcinoma and 76 benign tumor patients which collected from Department of Gynecology of Harbin Medical University Tumor Hospital between March 1,2011 and July 31,2013.We identified the biomarkers that are not related with CA125 and assessed the diagnostic accuracy of differential lipids and the combined diagnostic accuracy of those complementary biomarkers with CA125 using the 7-fold cross-validation method.In addition,cytoscape software was used to construct the network of these differential lipids to study the relationship of them.Finally,we identified the differential lipids biomarkers to distinguish early-stage from late-stage patients.RESULTS We identified 20 differential lipids between ovarian cancer and controls,of which Stearamide,Stearic acid,Arachidic acid,PI（42∶9）were not related with CA125（P〈0.05）,the AUC value of the combination of these lipids with CA125 was 0.94,which was larger than that with CA125 alone.Of the 20 lipids,8lipids were selected as potential biomarkers to distinguish those early-stage from late-stage patients,they are PC（35：4）,PC（38∶6）,PC（46∶4）,PC（P-35∶2）,PE（P-36∶6）,PG（34∶2）,Cer（d18∶1/16∶0）,3-Deoxyvitamin D3,respectively,and all P〈0.05.CONCLUSIONS Studying on the plasma lipidomics of ovarian cancer has improved the diagnostic accura

关 键 词：稀疏组lasso 脂质组学 卵巢癌 诊断标志物 

分 类 号：R737.31[医药卫生—肿瘤]