Fig. 2

3 miRNA panel can differentiate between cancerous and non-cancerous samples with very good sensitivity and specificity. (a) In tissue cohort, the combined measure of sensitivity and specificity of 3 miRNAs (miR-143, miR-145, and miR-140) was represented by AUC of 0.93 (p < 0.0001) with 93.02% sensitivity and 93.33% specificity. In the salivary exosome cohort, the combined AUC of 3 miRNA was 0.99 (p < 0.0001) with 98% sensitivity and 99% specificity. A diagonal reference line acts as a performance measure of the diagnostic test. Note: AUC-area under the curve, CI-confidence interval. (b) Representative images are 3D plots of PCA performed using the expression of 3 significantly expressed miRNAs. These plots suggest that expression of these 3 miRNAs can distinguish OSCC tissue samples and salivary exosomes from their normal counterparts with 81.12% variance and 89.98% variance respectively. (c) Kaplan–Meier plot of overall survival of patients with high and low expression of miR-143, miR-145, and miR-140 generated from data available on TCGA_HNSC and data obtained from the literature. (d-f) Association of identified 3-miRNA signature with various clinicopathological parameters. Representative box plots depict relative expression patterns of miRNAs in early stages (Stage I/II) vs. late stages (Stage III/IV) of OSCC patient-derived (d) tissue samples and (e) salivary exosomal samples. (f) Representative box plots depict relative expression patterns of tumour-derived miRNAs in recurrent vs. non-recurrent OSCC patients. The expression levels of miRNAs were estimated using real-time PCR. The data were normalised with U6 values, and the relative miRNA levels were analysed using the ddCt method. Error bars represent mean ± SD of three independent experiments (*p ≤ 0.05; **p ≤ 0.05; ***p ≤ 0.001)