Uterine perivascular epithelioid cell tumors (PEComa) and the accuracy of proposed classification systems in predicting the malignant versus non-malignant behavior

Objective: To compare the accuracy of available classification systems (Folpe, modified Folpe, Bennet, and Schoolmester) in predicting the behavior of uterine Perivascular Epithelioid Cell tumors (PEComas). Methods: We reviewed the pathology registry to identify all uterine PEComas treated at our center. We conducted a systematic literature review searching electronic databases from inception to November 2023. We included all references reporting at least one case of uterine PEComa; cases associated with tuberous sclerosis complex were excluded. Patient-level data were extracted by identified records. Survival analysis was used to assess the accuracy of all proposed classification systems to classify uterine PEComas as malignant versus non-malignant. Results: Six uterine PEComas were treated at our center. The literature search identified 101 uterine PEComas from 32 studies. Eighty-five out of 107 PEComas (28 studies and our series) reported enough follow-up data and details to apply all four classifications. The modified Folpe classification demonstrated the highest hazard ratio (HR) for relapse (HR:8.63; 95% confidence interval [CI] 2.06-36.1) and death due to PEComa (HR:6.8, 95%CI:0.89-51.6) for malignant versus non-malignant PEComas. Changing the cut-off of PEComa size to ≥8 cm and mitotic figures per 50 high power fields to ≥5, the HR for recurrence lowered (HR:6.26; 95% CI 2.20-17.80), but HR for death increased (HR:10.3; 95% CI 1.35-77.80). Conclusions: The modified Folpe classification was the most accurate in predicting the PEComa behavior. Changing the cut-off of PEComa size and number of mitotic figures may improve the accuracy in predicting death due to disease.