Round cell liposarcoma, a high-grade sarcoma, is a poorly differentiated form of myxoid liposarcoma, which is low grade. It is not known, however, how much of a round cell component within an otherwise typical myxoid liposarcoma results in a neoplasm that behaves as a high-grade sarcoma. Twenty-nine cases of myxoid liposarcoma of the extremities with or without a component of round cell liposarcoma were studied to semiquantitate the amount of round cell component needed to adversely affect prognosis. An estimate of the percent of necrosis, round cell liposarcoma, myxoid liposarcoma, and transitional areas was obtained for each slide on all cases. Transitional areas were defined as those that displayed an increased cellularity compared with typical myxoid liposarcoma, but in which the cells remained spindled, did not have overlapping nuclear borders, and retained an easily discernible plexiform vascular pattern. The amount of necrosis was subtracted from the total material available for evaluation, and a composite estimate of the percent of round cell, myxoid, and transitional areas was obtained. Two tumors were located on the upper extremity, 27 on the lower extremity; tumor size ranged frm 3 to 30 cm (median, 14 cm). All 29 tumors had a myxoid component, with a range from 12 to 100% (median, 73%). The range of transitional component for all 29 tumors was 0 to 88% (median, 11%). Twenty-one tumors had transitional areas (range, 4-88%). The range of round cell component for all 29 tumors was 0 to 58% (median, 0%). Twelve tumors had round cell areas (range, 1-58%). Seventeen patients are either alive without disease, or died from unrelated causes at 24-202 months (median, 96 months). Twelve patients are either alive with metastases or died of disease at 10 to 180 months (median, 53 months). Patients with > 5% round cell component in their initial tumor had a statistically significant higher rate of metastasis or death due to disease than those with < or = 5% round cell liposarcoma (p = 0.05). In addition, patients with myxoid liposarcoma with transitional areas did not fare worse than those with myxoid liposarcoma alone. In conclusion, we found that a round cell component of > 5% portends a higher risk of metastasis or death from disease. Furthermore, transitional areas alone do not appear to alter the prognosis of myxoid liposarcoma. Thus, only those areas that are unequivocally round cell liposarcoma should be designated as high grade.