Immunohistochemical analysis

All selected MPM histological samples were reviewed by an experienced mesothelioma pathologist (F.B.) and divided into epithelioid, sarcomatoid (including desmoplastic) and biphasic MPM according to the 2015 World Health Organization (WHO) classification (21). To include the MPM sample in the b-MPM subgroup, the presence of both sarcomatoid and epithelioid components was required at least in the ten percent of the tumor. All samples were FFPE and a single 4-mm-thick paraffin section was cut from the sample with the greatest amount of tumor tissue for each patient. All sections were deparaffinized and rehydrated in graded concentrations of xylene and ethanol. Sections were coated with 1:50 mouse monoclonal BAP1 antibody (clone C4:sc-28383; Santa Cruz Biotechnology, USA) and incubated at 4 °C overnight. Then, automated IHC was performed on Omnis platform (Agilent, USA). BAP1 IHC status was considered as “positive/retained” if there was an unambiguous positive nuclear staining in any number of tumor cells, and “negative/loss” if the nuclear staining was absent in neoplastic cells. Tumor cells with cytoplasmic reactivity without a clear nuclear staining were considered negative. Non-neoplastic cells, such as vascular endothelium, fibroblasts or inflammatory cells, were considered as internal positive control.