Soft tissue sarcomas (STS) are a group of malignant mesenchymal tumours that present a dilemma for clinical management. Current staging systems, based largely on morphological tumour characteristics, cannot accurately classify tumours nor predict an individual patient's risk for eventual metastases. Consequently, it is a challenge to identify patients with STS who should be considered for intensive adjuvant treatment protocols. Characterisation of specific gene expression patterns in STS will improve classification of these tumours and be of value in identifying patients at high risk for developing metastases.

Our aim is to generate comprehensive gene expression profiles of the molecular alterations that occur during the development and progression of human soft tissue sarcoma. We plan to accomplish this by taking advantage of two major resources — a national STS tumour bank with corresponding clinical information and a recently developed microarray facility.

STS specimens for this study will be obtained from the Canadian Sarcoma Group (CSG) Soft Tissue Sarcoma Tumour Bank that is a resource of tumour tissue and clinical correlative data situated at Mount Sinai Hospital, Toronto, Canada. The CSG tumour bank has been actively collecting tissue and pertinent patient information prospectively since 1993. As of December 1998, this bank had collected 701 tumour samples from 609 patients with STS, and is continuing to collect 110 STS cases per year. The CSG bank includes separate computer databases for clinical/pathologic and laboratory information, and stores frozen tumour and normal tissues, blood, as well as histological slides, paraffin embedded tissue and pathology reports from the collaborating centres. Most sarcomas in the tumour bank are large which generally allows for abundant frozen tumour tissue.

We plan to take advantage of the developing technology of cDNA microarrays that allows differential gene expression to be quantitated for thousands of genes simultaneously between multiple samples. A gene array facility, which is situated at the Ontario Cancer Institute, Toronto, is capable of producing high quality chips with up to 20,000 ESTs with an XYZ robot and generates data with a very fast and sensitive reader at low cost to collaborating members of the facility. Even without knowing the identity of the majority of the ESTs examined, patterns of expression can be identified that are characteristic of the behavioural aspects of a tumour and may be prognostically important. As well, studying these gene cohorts will help identify molecular pathways as well as novel genes important in tumourgenesis.