Aim: Finding reliable prognostic factors for osteosarcoma remains problematic. A systematic review [Davis AM, Bell RS, Goodwin PJ. Prognostic factors in osteosarcoma: a critical review. Journal of Clinical Oncology 1994; 12(2): 423-431.] showed chemotherapy response as only independent factor. We tried to identify evidence-based prognostic factors in the literature since 1992 and to establish pooled relative risks of factors.
Methods: MEDLINE and Embase search (1992-August 2006). Two reviewers independently selected papers addressing prognostic factors in localized extremity osteosarcoma, which were studied for methodological quality, and valuable new factors. An attempt was made to pool results.
Results: Of 1777 "hits", 93 papers were studied in depth. Several "new" prognostic factors were found. Only 7 papers were of sufficient quality to analyze. Chemotherapy response, tumor size and site, alkaline phosphatase level and p-glycoprotein expression seemed to be independent factors. Some new factors looked promising.
Conclusions: Although the literature is abundant, it is disappointing that only few papers are of sufficient quality to allow hard conclusions. Because of heterogeneity of the studies pooling results is hardly possible. There is a need for standardization of studies and reports.