A recent trend in the classification of uterine smooth muscle neoplasms (USMNs) into clinically benign and clinically malignant groups has been to move from exclusive reliance upon mitotic index (MI) to an approach that incorporates additional histopathologic characteristics. In furtherance of this goal, we assessed a variety of histopathologic features of 213 problematic smooth muscle neoplasms for which we had greater than or equal to 2 years of clinical followup data or for which there was an unfavorable outcome. One hundred and thirteen of these patients have had a minimum follow-up of 5 years, and 48 have been followed for greater than or equal to 10 years. Cases eliminated from the study group included USMNs with a significant myxoid or epithelioid component and cases of intravenous leiomyomatosis. USMNs, whether cellular or not, with no cytologic atypia and with a mitotic index (MI = number of mitotic figures [mf]/10 high-power fields [hp]l) of <5 mf/10 hpf (usual leiomyomas) were also excluded unless they had unusual features or were associated with an adverse clinical outcome. Fifty-six patients were initially treated by myomectomy or another form of local tumor removal; the remainder had a hysterectomy. From a wide variety of light microscopic features assessed, the important predictors that emerged, using a variety of data exploratory techniques, were MI, the degree of cytologic atypia, and the presence or absence of coagulative tumor cell necrosis (CTCN). Stratification of the USMNs with respect to these three features resulted in a five-group classification of USMNs with the following major characteristics. Group 1: Of the 89 USMNs with an MI in the range 5 less than or equal to MI<20 without CTCN and with no more than mild atypia, 88 were clinically benign. One patient with a tumor in this group died of metastatic disease 96 months after her uterine cervical primary neoplasm was removed. Combining our data with that in the literature, the failure rate in this group is similar to 1/200 (0.5%). This low failure rate war-rants the use of the label ''leiomyoma with increased mitotic index'' for USMNs with these histologic features. Two patients whose USMNs were characterized by mild atypia, no necrosis, and MI <5 developed identical-appearing pulmonary metastases and were judged in retrospect to have the syndrome ''benign metastasizing leiomyoma.'' Group 2: USMNs with no CTCN and diffuse moderate to severe atypia fell into two groups based on the MI, For those patients whose neoplasms had an MI greater than or equal to 10 mf/10 hpf, four of 10 failed. These tumors we label ''leiomyosarcoma.'' For the patients with tumors demonstrating diffuse moderate to severe atypia, no CTCN, and an MI<10 mf/10 hpf, one of 46 failed. The patient who failed is alive with disease at 60 months. This failure rate of one in 46 (2%) warrants the term ''atypical leiomyomas with a low risk of recurrence'' for these tumors. Group 3: Twenty-four of 33 patients whose tumors demonstrated both diffuse moderate to severe atypia and CTCN failed. The failures occurred in all atypia-MI subsets. These neoplasms should be labeled ''leiomyosarcomas.'' Group 4: Eight patients had a USMN with insignificant cytologic atypia but CTCN. Four of these patients experienced a clinical relapse; all except one of the latter tumors had an MI greater than or equal to 10 mf/10 hpf. These findings would suggest that this group be labeled ''leiomyosarcoma'' when the MI exceeds 10 mf/10 hpf. Group 5: Five patients had USMNs with no CTCN and focal or multifocal moderate to severe atypia, and all had favorable outcomes. Two of the tumors had a mitotic index between 10 and 20. This study highlights the importance of using features other than mitotic index to predict clinical outcome in USMNs. Coagulative tumor cell necrosis emerged as a particularly crucial feature. An accurate MI determination in this classification scheme is less important.
