PEC-PRO: A new prognostic score from a series of 93 patients with localized perivascular epithelioid cell neoplasms (PEComas) treated with curative intent.

11571 Background: Perivascular Epithelioid Cell Neoplasms (PEComas) encompass a heterogeneous family of mesenchymal tumors. The current understanding of their natural history is limited. Previously described clinicopathological factors aimed to define benign or malignant variants, but there is a lack of prognostic factors associated with recurrence of surgically resected tumors, preventing the development of a prognostic score to better optimize patient’s management. Methods: This is a retrospective analysis of clinicopathological features from patients diagnosed with a localized PEComa, within all centers from the French Sarcoma network and one center in Belgium. We analyzed 12 clinicopathological factors in a Cox proportional hazard framework to derive a multivariate prognostic risk model for progression-free survival (PFS). We built the PEComa PROgnostic score (PEC-PRO) ranging from 0 to 5, based on the coefficients of the multivariate model. Three different prognostic groups were identified: low risk (score = 0), intermediate risk (score = 1) and high risk (score ≥2). Results: Ninety-three patients were analyzed with a median follow-up of 46 months (range, 3-253). At diagnosis, the median age was 54 years (range, 13-84), with female predominance (72%). Most common primary locations were uterus (n = 15;16%) and kidney (n = 15;16%). Median tumor size was 6.2 cm (range, 0.8-30). Among patients with reported surgical margins, 64 (73%) and 23 (27%) had R0 and R1-2 margins, respectively. The median PFS was 26 months (IC95, 2.9-124.4), with 1- and 5-year overall survival (OS) rates of 95.7% and 69.9%, respectively, while the median OS was not reached. Using univariate analyses, male gender, primary tumor size > 5 cm, high nuclear grade and cellularity, high mitotic rate > 1/50 HPF, necrosis, vascular invasion, nodal invasion, and R1-2 margins were associated with a shorter PFS. Among those, male gender (HR = 2.88; IC95 1.12-7.411, p = 0.03), vascular invasion (HR = 3.14; IC95 1.10-8.96, p = 0.034), necrosis (HR = 3.93; IC95 1.35-11.47, p = 0.015), and R1-2 margins (HR 4.47; IC95 1.60-12.46, p = 0.007) remained associated with PFS in the multivariate analysis and were included in the multivariate model. Median PFS in patients with high PEC-PRO score was 16 months as compared to 104 months and not reached for patients with intermediate and low PEC-PRO scores, respectively (p < 0.001). We also confirmed the prognostic relevance of the PEC-PRO score in terms of OS. Conclusions: Using a weighted combination of clinicopathological features, the PEC-PRO score reliably predicts the post-operative recurrence risk in patients with localized PEComas. It has the potential to better improve follow-up strategies and personalize adjuvant treatments. The findings of this retrospective analysis require validation in a prospective trial.