Cross validated serum small extracellular vesicle microRNAs for the detection of oropharyngeal squamous cell carcinoma

• DOI: 10.1186/s12967-020-02446-1
• Publication Date: 2020-07-10T08:04:00Z
• Authors: G. C. Mayne; C. M. Woods; N. Dharmawardana; T. Wang; S. Krishnan; J. C. Hodge; A. Foreman; S. Boase; A. S. Carney; E. A. W. Sigston; D. I. Watson; E. H. Ooi; D. J. Hussey
• Keywords: Microvesicle; Biomarker Discovery

Abstract Background Oropharyngeal squamous cell carcinoma (OPSCC) is often diagnosed at an advanced stage because the disease causes minimal symptoms other than metastasis to neck lymph nodes. Better tools are required assist with early detection of OPSCC. MicroRNAs (miRNAs, miRs) potential biomarkers for head and cancer diagnosis, prognosis, recurrence, presence metastatic disease. However, there no widespread agreement on a panel miRNAs clinically meaningful utility cancers. This could be due variations in collection, storage, pre-processing, isolation RNA, but several reports have indicated that selection reproducibility has been widely affected by methods used data analysis. The primary analysis issues appear model overfitting incorrect application statistical techniques. purpose this study was develop robust approach identify miRNA signature can distinguish controls patients inflammatory from human papilloma virus positive (HPV +) Methods Small extracellular vesicles were harvested serum 20 control patients, gastroesophageal reflux (GORD), 40 locally HPV + purified, expression profiled OpenArray™. A novel cross validation method, using lasso regression, developed stabilise inclusion prediction model. named StaVarSel (for Stable Variable Selection), derive diagnostic biomarker signature. Results standard unable produce good validated predictive capacity. In contrast, produced regression containing 11 ratios clinical utility. Sample permutations estimated accuracy 11-miR-ratio not chance alone. Conclusions We StaVarSel, able miRNAs, present small derived blood serum, robustly as