A panel of microRNAs can be used to determine oral squamous cell carcinoma. (11th July 2017)

• Record Type: Journal Article
• Authors: Prasad, Gareema; Seers, Christine; Reynolds, Eric; McCullough, Michael J.
• Is Part Of: Journal of oral pathology & medicine. Volume 46:Number 10(2017)
• Page Start: 940
• Page End: 948
• Publication Date: 2017-07-11
• Subjects: algorithm -- diagnostic -- microRNAs -- oral cancer; Dentistry -- Periodicals; Teeth -- Diseases -- Periodicals; 617
• DOI: 10.1111/jop.12592
• Languages: English
• ISSNs: 0904-2512

Abstract

Abstract : Background: Subjective histopathology is currently used to diagnose oral squamous cell carcinoma (OSCC). We tested if abundances of a panel of microRNA could be an objective OSCC indicator. Method: Literature review enabled identification of 10 microRNAs associated with oral and head and neck malignancies. We extracted RNA from formalin‐fixed paraffin‐embedded biopsies; 20 each with OSCC, dysplasia, or histologically normal epithelium (HNE) and 10 with oral lichen planus (OLP). Relative abundances of microRNAs in HNE and OSCC were determined using reverse transcription and then real‐time PCR with global mean normalization. MicroRNAs differentially expressed (test microRNA, T‐miR) and non‐differentially expressed (normalization microRNA, N‐miR) were identified. The raw microRNA Cq data were incorporated in a developed algorithm that output a T‐miR expression value (T‐miREV) score. Raw Cq data from HNE, OSCC, dysplasia, and OLP samples were then used to test the algorithm scoring and OSCC classification. Results: Four test and normalization microRNAs were identified. Algorithm output of T‐mirEV >1 or <−1 indicated high and low OSCC probability score, respectively, and gave 88.9% sensitivity, 100% specificity, and 93.5% accuracy. Grouping high and intermediate T‐mirEV scores (T‐miREV ≥−1) resulted in sensitivity of 90%, specificity of 65%, and accuracy of 77.5% in OSCC classification. All 20 dysplasias and eight of 10 OLP had T‐miREV ≥−1 indicating intermediate to high probability of malignant changes. Conclusion: A microRNA panel combined with our algorithm can identify tissue with probable oncogenic changes. Impact: The developed algorithm serves as a baseline for prospective trials, which may result in potential clinical utility.