Sarcomas are a diverse group of childhood and adult tumors that arise from mesenchymal tissue. In contrast to epithelial tumors, most of which are defined by a specific organ system, sarcomas can arise virtually anywhere in the body, such that their characteristic histopathology and clinical presentation form the core diagnostic criteria.Precise identification by differential diagnosis is the first element of a successful treatment, since these tumors show wide variation in response to specific therapies and misdiagnosis can lead to suboptimal therapy. However, due to overlapping histopathologic features among the sarcomas, as well as the multiple subtypes or variants within a single histologic group, pathologists and clinicians are increasingly reliant on molecular diagnostic approaches to aid in the differential diagnosis. Gene expression profiling or microarray analysis is now being used to develop expression signatures that appear to be better than histological features or any single biomarker at discriminating tumor types, identifying clinical variants, and modeling complex tumor behavior.This review examines the current progress in identifying diagnostic and prognostic expression signatures for four sarcomas: rhabdomyosarcoma, Ewing's family of tumors, synovial sarcoma, and osteosarcoma. In this context, we discuss the current status and future potential for using expression signatures to improve tumor classification, outcome prediction, and therapeutic response in patients with these sarcomas.