Background Diagnosis and tissue-of-origin identification of malignant pleural mesothelioma (MPM) is currently assessed using a panel of positive/negative markers; however, there remains a subset of cases that are not definitively identified. The key role of miRNAs in cancer biology has encouraged their evaluation as clinical markers. Recent evidence indicates that the human genome encodes more miRNAs than are currently annotated, most of them displaying tissue-specific patterns. Here, we hypothesized that MPM tumors express a specific set of previously-uncharacterized miRNA sequences, and these species can distinguish MPM from other thoracic diseases. Methods We conducted a de novo search for novel miRNAs sequences in a cohort of MPM tumors (n=87), using smallRNA sequencing data from The Cancer Genome Atlas (TCGA). Using the newly-identified miRNA species, we performed a genomewide 3’UTR target prediction analysis using the miRanda algorithm. To investigate the ability of novel miRNAs to distinguish MPM from other thoracic cancers, we assessed their expression in 1,093 lung tumors from four independent sample cohorts from TCGA and the BC Cancer Agency (BCCA): two adenocarcinoma (LUAD) cohorts (TCGA n=497, BCCA n=94) and two squamous cell carcinoma (LUSC) cohorts (TCGA n=467, BCCA n=35). Finally, a classifier model was built using the weighted voting class prediction method. Results We identified 424 predicted novel miRNA-like sequences, which were subsequently filtered by RNA structure, abundance, and genomic location to identify a high-confidence set of 154 previously unannotated miRNA sequences. Some of the most highly expressed novel miRNAs were predicted to establish thermodynamically stable interactions (based on sequence homology) with 3’UTR region of genes relevant to MPM biology, such as the Ataxia Telangiectasia Mutated (ATM) and BRCA1 Associated Protein 1 gene (BAP1).A principal component analysis revealed that combined expression of the 154 newly-discovered miRNAs unambiguously distinguished MPM from LUAD and LUSC. To explore their clinical potential, we developed a 10-novel miRNA classifier by comparing MPM and LUAD cases from TCGA and validated by comparing MPM against LUAD cases from the BCCA cohort. Remarkably, this classifier successfully identified 86 out of the 87 MPM cases (98.8%) and 100% of LUAD cases (true positive rate = 98.85%, false positive rate = 1.15%). Conclusion Here, we provide evidence for the presence of 154 previously unidentified miRNA species in MPM.These miRNAs not only significantly expand the miRNA repertoire but also unveil specific roles in MPM biology. Furthermore, the specificity of these novel miRNAs could supplement current clinical practices and help to address unmet clinical needs in MPM.