Pleomorphic leiomyosarcoma (PLMS) was recently described as a morphologic variant of leiomyosarcoma; however, its diagnostic criteria, as shown by morphologic features and biologic behavior, remain controversial. We describe 28 cases of pleomorphic sarcoma with pleomorphic areas in more than two thirds of the tumor and an ordinary leiomyosarcomatous fascicular area covering less than one third as PLMS. PLMS comprised 8.6% of all the leiomyosarcomas (322 cases) registered in our institute. Patients ranged in age from 31 to 89 years (average, 57.9 years). Seventeen patients (60.7%) were male and 11 were female. Tumor location was as follows: the extremities in 17 cases, the retroperitoneum or abdominal cavity in 7 cases, the chest/abdominal wall in 3 cases, and the scalp in 1 case. Histologically, all cases showed at least small foci of fascicles consisting of smooth muscle tumor cells, in addition to pleomorphic areas mimicking storiform-pleomorphic malignant fibrous histiocytoma. The border between pleomorphic and leiomyosarcomatous fascicular areas was sharp in 3 cases, gradual in 2 cases, and blending in 23 cases. Sixteen cases (57.1%) showed a typical storiform pattern, 6 cases revealed extensive stromal hyalinization, 6 cases showed a chronic inflammatory infiltrate, 2 cases had the foci of foamy xanthomatous cells, and 7 cases contained myxoid malignant fibrous histiocytoma-like areas covering less than 50% of the tumor. The tumors had a tendency to be of a morphologically higher grade (10 tumors were French Federation of Cancer Centers grade 2, 18 were grade 3). Five of 28 cases (18%) showed rhabdoid features. Immunohistochemically, all of the 28 tumors examined showed a positive reactivity for at least one smooth muscle marker (desmin, muscle-specific actin, and alpha-smooth muscle actin) in the leiomyosarcomatous fascicular areas. In the pleomorphic areas the expression of smooth muscle markers (desmin 10 of 28, muscle-specific actin 13 of 28, and alpha-smooth muscle actin 14 of 28) was significantly reduced, compared with that in leiomyosarcomatous fascicular area (desmin 18 of 28, muscle-specific actin 26 of 28, and alpha-smooth muscle actin 24 of 28). No significant difference was observed between the MIB-1 labeling index in the leiomyosarcomatous fascicular areas (26.10 on average) and that in the pleomorphic areas (26.17 on average). However, the MIB-1 labeling index in PLMS was significantly higher than that in ordinary leiomyosarcoma (n = 20, 12.86 on average) or storiform-pleomorphic malignant fibrous histiocytoma (n = 16, 16.63 on average). In 23 patients follow-up data were available with a duration of 1-239 months. Eleven patients developed metastases, and lung accounted for the most common site of metastasis (9 cases). Fifteen of 23 patients (65.2%) died of disease. Our results indicate that PLMS should be differentiated from ordinary leiomyosarcoma because of its high proliferative activities and rather aggressive biologic behavior.