Clinical and morphologic variables in soft-tissue sarcomas and lipomas were retrospectively analysed and compared in consecutive, unselected population-based series from southern Sweden 1964-1981. Among the sarcoma patients these variables and the treatment given were evaluated by multivariate analysis with regard to the clinical course. The results of aspiration cytology and its influence on the choice of surgical procedure were evaluated in two series of patients referred to the Orthopaedic Oncology Group (the Center) because of suspected malignant soft-tissue tumors. Changes in the surgical treatment over time was determined. The annual incidence of sarcoma was 1.4/10(5) with a 30 per cent male preponderance. The mean age was 58 years. The most common histologic types were malignant fibrous histiocytoma, liposarcoma and leiomyosarcoma. Twenty-three per cent were histologically classified as low-grade malignant (Grades I and II), 33 per cent as Grade III and 44 per cent as Grade IV tumors. Slightly more than one half of the tumors were deep and these had a median size of 8 cm compared to 4 cm for the superficial tumors. One third of the tumors were located in the thigh. The annual clinical incidence of solitary subcutaneous lipoma was estimated to 1/10(3). Four fifths of the lipomas were smaller than 5 cm and they were most common in the trunk, shoulder and upper arm. By comparing clinical data for benign tumors and sarcomas it was found that a tumor 5 cm or larger or a deep tumor is relatively more likely to be a sarcoma. Patients with tumors of that size and depth should be referred before surgery. The probability of a benign cytodiagnosis being correct was 0.97 while that of a malignant one was 0.85. For a cytodiagnostic report of sarcoma the probability of correct diagnosis was 0.84. In a prospective evaluation the planned surgery was changed by subsequent cytodiagnosis in one third of the patients. Negative prognostic variables as regards survival in the order of decreasing relative risks were: Malignancy Grades IV and III, pain, male sex, increasing age and tumor size, marginal surgery and extracompartmental tumor site. Marginal excision, extracompartmental tumor site and age exerted their effect only via local recurrence which had a stronger negative influence on survival than any of the other variables. A linear survival function based on the prognostic variables was related to the probability of survival and the patients could be stratified into subsets with significantly different survival.(ABSTRACT TRUNCATED AT 400 WORDS)