A Serum Microrna Biomarker Panel for Detection of Gastric Cancer.
Jimmy So,Ruiyang Zou,Lihan Zhou,Heng-Phon Too,Feng Zhu,Yik Ying Teo,Celestial T. Yap,Sun Young Rha,Koji Kono,Joanne Yoong,Khay-Guan Yeoh,Wei-Peng Yong
DOI: https://doi.org/10.1200/jco.2015.33.15_suppl.4060
IF: 45.3
2015-01-01
Journal of Clinical Oncology
Abstract:4060 Background: Gastric cancer is the 2nd most common cause of cancer deaths worldwide. Currently, endoscopy is the only reliable method for early diagnosis. However, the invasiveness and cost limit its usage as a screening test. MicroRNAs (miRNAs) have been shown to be important in the pathogenesis of cancers. They are exceptionally stable in body fluids, making them potential noninvasive biomarker for cancers. Methods: We screened 600 miRNAs using our MiRXES qPCR technology in the sera of 236 gastric cancer subjects and 236 matched high risk subjects, serving as the discovery set. 191 miRNAs were reliably detected in all the serum samples, out of which 75 informative ones were identified to be significantly (false discovered corrected P value lower than 0.01) altered between gastric cancer and high risk subjects. Multivariate miRNA biomarker panels were then formulated by sequence forward floating search and support vector machine using all the quantitative data obtained for the expression of 191 miRNAs. Results: Multiple iterations of two-fold cross-validation were performed in silico where the panels with 8 or more miRNAs consistently achieved high accuracy [areas under the curve (AUC) ≥ 0.87 in the receiver operating characteristic (ROC) curve]. A 24-miRNA model was optimized based on the discovery set [AUC = 0.92 ( 95% Confidence Interval [CI]: 0.88-0.95)] and validated in two blinded studies with fixed algorithm and threshold definition; Korean case-control cohort (n = 129, 74 cancer cases) and Singaporean Chinese case-control cohort (n = 89, 20 cancer cases). The 24-miR panel showed 90% sensitivity and 81% specificity in the Korean cohort (AUC = 0.91 (95% CI: 0.86-0.96)] and had 90% sensitivity and 75% specificity in the Singaporean Chinese cohort [AUC = 0.89 (95% CI: 0.79-0.99)]. Importantly, the 24-miR panel was able to distinguish stage 1 and 2 gastric cancer in both Korean [AUC = 0.88 (95% CI: 0.81-0.96)] and Singaporean Chinese [AUC = 0.91 (95% CI: 0.80-1.00)] cohorts. Conclusions: We have identified a serum miRNA panel which can confidently differentiate patients with gastric cancer including the early-staged cancers from controls. This may be able to serve as a non-invasive screening test for gastric cancer which warrants study in larger cohorts.
What problem does this paper attempt to address?