Pleomorphic liposarcoma is an uncommon form of liposarcoma that only recently has been properly characterized. A series of 57 cases is presented. Patient age at presentation ranged from 27 to 95 years (median, 54 years), and there was a slight male predilection (male/female ratio = 1.2:1). Tumors most frequently involved the lower limb (47% of cases) or upper limb (18%). Other anatomic sites, including trunk (14%), retroperitoneum (7%), head and neck (5%), abdomen/pelvis (5%), and spermatic cord (4%), were less frequently involved. Tumor size ranged from 1.5 to 21 cm (median, 8 cm), with deep (subfascial) locations (39 cases) being more frequent than subcutaneous (11 cases) or dermal sites (5 cases). All lesions showed features of pleomorphic sarcoma and at least focally contained typical multivacuolated lipoblasts. Although there was considerable overlap, tumors fell into three broad categories: high-grade pleomorphic/spindle cell sarcoma with scattered lipoblasts or sheets of lipoblasts (60%), high-grade pleomorphic sarcoma with epithelioid areas and scattered lipoblasts (28%), and intermediate- to high-grade sarcoma predominantly resembling myxofibrosarcoma except for the presence of lipoblasts (12%). Immunohistochemistry revealed focal staining for smooth muscle actin in 13 of 29 cases (45%), S-100 protein positivity in lipoblasts in 15 of 45 cases (33%), focal staining for keratin in 6 of 28 cases (21%), including 5 of 13 (38%) with epithelioid morphology, and focal staining for desmin in 4 of 30 cases (13%). Follow-up data, available in 50 patients (88%) (median, 33 months), showed local recurrence in 34% of patients, systemic metastases in 32%, and tumor-related death in 32%. Only 2 of the 16 superficial (dermal or subcutaneous) lesions metastasized. Five-year overall, local recurrence-free, metastasis-free, and disease-free survivals were 63%, 58%, 58%, and 39%, respectively. By univariate analysis, central (nonextremity) location, deep situation, tumor size > or =10 cm, mitotic rate > or =10 per 10 HPF, necrosis, and epithelioid morphology were associated with a worse prognosis. However, by multivariate analysis, only age > or =60 years, central location, tumor size, and mitotic rate remained independent predictors for an adverse outcome. By multivariate analysis, wide local excision or amputation and postoperative radiotherapy protected against local recurrence.