210P: Digital gene expression profiling to separate malignant pleural mesothelioma from benign reactive mesothelial hyperplasia

Background The diagnosis of malignant pleural mesothelioma (MPM) is based on the histological analysis of pleural lesions, however the morphological separation of benign mesothelial hyperplasia (MH) from MPM can be exceedingly difficult. Nowadays the most reliable indicator of malignancy is the mesothelial cells invasion of the chest wall soft tissue or of the underlying lung parenchyma. Several deregulated gene pathways have been described in MPM, we investigated how the over and down expressed genes work together in the differential diagnosis Methods We designed a custom NanoString Codeset including 113 genes with a crucial role in cancer and 6 reference genes. The gene expression analysis was performed by the nCounter System ®, without any amplification steps, directly on RNA purified from 48 formalin-fixed and paraffin-embedded tissues of epithelioid mesothelioma (32) and MH (16) samples. Results A total of 43 genes resulted deregulated in MPM in comparison with MH (Mann–Whitney U test; P < 0.005): 24 genes exhibited an over expression (EGFR, ITGA3, PAK4, MSLN, GLI2 and others) and 19 showed a down expression (BAP1, ITGA5, CD44, MMP9, PDGFRB and others). To model the gene expression profiles of MPM and MH samples we performed a cluster analysis of expressed genes using the Euclidean distance between samples. The cluster analysis exposed an evident difference between MPM and MH, particularly after we filtered the differentially expressed genes. Conclusions We analysed a panel of genes some of which known as deregulated in MPM, however none of these genes have yet to be used as a biomarker. To evaluate how transcriptomic data could be applied in the diagnosis of MPM we used an enzyme-free digital count of mRNA molecules to analyse simultaneously all the selected genes and to reduce the potential errors associated with multiple qPCR assays. We identified a specific panel composed of 43 genes, whose expression profile resulted clearly distinct between MPM and MH. Our genes panel may constitute, after further validation on a larger series of samples, a powerful tool to separate MPM from MH, improving the current diagnostic methods.