Gynaecological perivascular epithelioid cell tumour (PEComa): comparative analysis of proposed algorithms for prediction of clinical outcome

Aims: Perivascular epithelioid cell tumours (PEComas) are rare mesenchymal tumours that coexpress smooth muscle and melanocytic markers. They have a predilection for gynaecological organs, where they present a unique diagnostic challenge, because of morphological and immunohistochemical overlap with more common smooth muscle and stromal tumours. Limited information regarding the natural history, owing to the rarity of this tumour, makes accurate risk stratification difficult. We aimed to review clinicopathological features of gynaecological PEComa and compare accuracy of five different classification systems for prediction of prognosis.

Methods and results: We have described the clinicopathological features of 13 new cases and tested five prognostic algorithms in a total of 67 cases of gynaecological PEComa. Receiver operating characteristic curves were constructed and areas under the curve (AUCs) were calculated to evaluate predictive accuracy. The modified gynaecological-specific algorithm showed high sensitivity and specificity and yielded the highest AUC (0.864). It's earlier version, the gynaecological-specific algorithm, suffered from lower specificity (AUC = 0.843). The post-hoc McNemar test confirmed significant differences between the performances of the modified gynaecological-specific algorithm and the gynaecological-specific algorithm (P = 0.008). The original Folpe algorithm for PEComas of all sites showed low specificity, had a lower AUC (0.591), and was inapplicable in 18% of cases. Its two later versions (the revised Folpe algorithm and the modified Folpe algorithm) also yielded lower AUCs (0.690 and 0.591, respectively).

Conclusion: We have shown that the modified gynaecological-specific algorithm predicts the clinical outcome of gynaecological PEComa with high accuracy, and have validated its use for prognostic stratification of gynaecological PEComa.