Plasma microRNA panels to diagnose pancreatic cancer: Results from a multicenter study

摘要: // Zhe Cao 1, * , Chang Liu 2, Jianwei Xu 3, Lei You 1 Chunyou Wang 4 Wenhui Lou 5 Bei Sun 6 Yi Miao 7 Xubao 8 Xiaowo 2 Taiping Zhang Yupei Zhao Department of General Surgery, Peking Union Medical College Hospital, Chinese Academy Sciences and College, Beijing, 100730, China MOE Key Laboratory Bioinformatics, Bioinformatics Division Center for Synthetic Systems Biology, TNLIST/Department Automation, Tsinghua University, 100084, 3 Qilu Shandong Jinan, 250012, Pancreatic Disease Institute, Wuhan Tongji Huazhong University Science Technology, Wuhan, Hubei Province, 430022, Zhong Shan Fudan Shanghai, 200032, Hepatobiliary The First Affiliated Harbin Harbin, 150001, Nanjing Nanjing, 210029, Hepatopancreatobiliary West Sichuan Chengdu, Sichuan, 610041, These authors contributed equally to this work are co-first Correspondence to: Wang, email: xwwang@tsinghua.edu.cn Zhang, tpingzhang@yahoo.com Zhao, zhao8028@263.net Keywords: pancreatic cancer, microRNA panels, multicenter study, diagnosis Received: February 15, 2016      Accepted: May 04, Published: 19, 2016 ABSTRACT Biomarkers the early cancer (PC) urgent needed. Plasma microRNAs (miRNAs) might be used as biomarkers cancer. We analyzed 361 plasma samples from surgical centers in performed machine learning approach. gain insight association between aberrant miRNA expression disease. 671 were screened discovery phase 33 training 13 validation phase. After phase, diagnostic panels constructed comprising panel I (miR-486-5p, miR-126-3p, miR-106b-3p) II miR-106b-3p, miR-938, miR-26b-3p, miR-1285). Panel had high accuracy distinguishing chronic pancreatitis (CP) with area under curve (AUC) values 0.891 (Standard Error (SE): 0.097) 0.889 (SE: respectively, Additionally, we demonstrated that value discriminating PC CP comparable carbohydrate antigen 19–9 (CA 19–9) 0.775 0.053) ( P = 0.1 both). This study identified based on potential distinguish CP. patterns developed