Several prognostic markers have been identified for soft tissue sarcomas. However, there reproducibility as regards both classification by different examiners and prognostic strength, as evaluated in different tumor series is poorly documented. Relevant factors and their strength differ in different subtypes of sarcoma. Thus, there is no generally accepted prognostic/staging system for soft tissue sarcoma. It seems possible that highly malignant soft tissue sarcomas behave like highly malignant bone tumors, such as osteosarcoma and Ewing sarcoma-i.e., almost all patients have metatastic disease when the primary tumor is diagnosed, although in most cases it is not detectable with today's imaging techniques. The highly malignant character of these tumors is defined by their histotype alone and the microscopic diagnosis is almost always easy. In contrast, no soft tissue sarcoma histotype is per se associated with such a poor prognosis: further prognostic markers are needed. The ideal prognostication system includes only two groups of patients; after removal of the primary tumor, one group has 100% survival and the other 0% survival. In the latter group, the tumor has metastasized and adjuvant therapy is necessary to improve survival. Meaningful chemotherapy trials requires reliable identification of high-risk patients; only these should be included. Inclusion of also low-risk patients is disadvantageous for two reasons: chemotherapy may have serious side-effects and positive effects may be difficult to detect when the trial is diluted by low-risk patients. Prognostication in soft tissue sarcoma has improved, but it may a long time before soft tissue sarcoma histotypes as highly malignant as osteosarcoma and Ewing sarcoma can be defined. Another approach would be to ignore efforts to characterize prognostic features of the primary tumor and instead, with some type of tracer technique (see pp x-y in this issue), focus on whether it has metastasized or not.