Non-invasive tests for the detection of nasopharyngeal carcinoma / Sivanesan Vijaya Mohan

Nasopharyngeal carcinoma (NPC) is a malignancy that arises in the epithelium of the nasopharynx. Diagnosis is often delayed due to inaccessibility of primary tumour site which leads to inaccurate biopsy. The pathogenesis of NPC is linked to the Epstein-Barr virus (EBV) and studies have shown the employability of plasma EBV DNA for the detection and monitoring of NPC. MicroRNAs (miRNA) are a class of small non-coding RNA that are dysregulated in various cancers and have been extensively studied as biomarker due to their stability in various biospecimens. The aim of this project is to evaluate the load of EBV DNA and dysregulated miRNAs as potential biomarkers to detect NPC in two non-invasive samples from nasal washings (NW) and saliva (SL). Thirty-five NPC and 64 non-NPC patients were included in this study. Two different quantitative polymerase chain reaction (qPCR) assays, which targeted the EBNA1 gene and the BAMHI-W region of the EBV genome, were evaluated to quantify EBV DNA load, and evaluate their accuracy. Twenty-seven human and EBV miRNAs were shortlisted from datasets deposited in the Gene Expression Omnibus (GEO). Expression of these miRNAs were first evaluated in 6 NPC and 6 non-NPC tissues samples and then validated in the NW sample set using reverse transcription qPCR (RT-qPCR). Receiver Operating Characteristic (ROC) analysis was used to calculate the area under the ROC curve (AUC) to evaluate these biomarkers as classifiers for NPC. Multiple Logistic Regression (MLR) was performed to determine if the combination of these biomarkers will lead to an NPC classifier panel with improved accuracy. EBNA1, BAMHI-W, miR-21, miR-26a, miR-29c, miR-93, miR-205, miR-375 and miR-421 were upregulated in NPC samples compared to controls (p<0.05). MLR showed that combination of EBNA1 and miR-21 produced the best AUC of 0.860 with 80.0% sensitivity, 78.1% specificity. Thirty-five NPC and 25 non-NPC saliva samples were analysed in this study. The load of EBNA1 together with 30 human and EBV miRNAs were evaluated in these samples. EBNA1 was not differentially expressed but miR-363-3p, miR-20a-5p, miR-222-3p, miR-126-3p, miR-361-5p, and miR-21-5p were significantly upregulated (p