Expression data from 87 complex sarcomas

Based on the transcriptome analysis of 555 sarcomas, we identified a group of tightly clustered leiomyosarcomas (LMS) due to their gene expression homogeneity. We named them “hLMS” and the other LMS “oLMS”. We derived a transcriptional signature able to identify each group and used it to classify patients from two independent cohorts. In all cohorts, hLMS were preferentially carried by women, located in the internal trunk, highly differentiated, and similarly altered at the genomic level. Based on integrative bioinformatic analysis, we show that hLMS originate from vascular smooth muscle cells presenting both contractile and synthetic characteristics, while oLMS could derive from fibroblasts. We found strong MYOCD expression to be an hLMS-specific driver and show that the MYOCD/SRF axis is essential only for hLMS survival. Identification of hLMS could become standard clinical practice, leading to the development of specific effective treatments with MYOCD/SRF inhibitors. These 87 samples from leiomyosarcoma primary tumors were analysed on Affymetrix and Agilent platforms and both datasets were used to define genes with consistent expression between the platforms. These genes were then selected in other datasets from the two plaforms before combination and normalization. 87 Complex Sarcomas hybridized to Agilent-014850 Whole Human Genome Microarray 4x44K G4112F