Project Overview

Conventional methods of tumor diagnosis, including histology, immunohistochemistry, cytogenetics, and molecular genetics, are capable of resolving many of the diagnostic issue encountered in the clinical evaluation of tumor biopsy specimens.  However, two areas in which these methods are often inadequate relate to (1) the classification of tumors having ambiguous histogenesis and (2) the determination of prognosis among tumors having similar diagnostic features but diverse clinical outcomes.  Various forms of human soft tissue sarcoma illustrate these problems exceptionally well.  In the research described in this application, we will address both types of problems in sarcomas by transcriptional profiling of tumors using hybridization of probes for total tissue RNA to microarrays of cDNAs.  To establish technical and biologic variables that may influence interpretation of these analyses, we will first perform control studies on the reproducibility and reliability of transcriptional profiles and on the effect that cellular heterogeneity both within and between tumors may have on such profiles.  We will then use transcriptional profiling to determine whether malignant fibrohistiocytomas, a large group of sarcomas of unclear histogenesis and often diagnose using loose criteria and by exclusion from other categories, can be definitively reclassified based on transcriptional profiles.  Finally, we will determine whether distinctive transcriptional profiles of tumors are correlated with different clinical outcomes in malignant fibrohistiocytomas, synovial sarcomas, and liposarcomas – three types of sarcomas associated with variable and unpredictable clinical behavior.  Ultimately, we believe that research proposed in this application will resolve whether comprehensive analysis of gene expression can provide objective methods that will complement or even replace conventional procedures for the classification and determination of prognosis in sarcomas.