Improved risk assessment of endometrial cancer by combined analysis of MSI, PI3K-AKT, Wnt/β-catenin and P53 pathway activation.

OBJECTIVE To investigate if analysis of genetic alterations in the main pathways involved in endometrioid type carcinogenesis (PI3K-AKT, Wnt/β-catenin, P53-activation and MSI) improves the current risk assessment based on clinicopathological factors. METHODS Formalin fixed paraffin embedded (FFPE) primary tumor samples of 65 patients with FIGO-stage I endometrioid type endometrial cancer (EEC) were selected from the randomized PORTEC-2 trial. Tumors were stained by immunohistochemistry for P53, PTEN and β-catenin. Tumor DNA was isolated for sequence analysis of TP53 (exons 4 to 8), hotspot mutation analysis of KRAS (exon 1) and PI3K (exon 9 and 20) and microsatellite-instability (MSI) analysis including MLH1 promotor-methylation status. Univariate and multivariate analyses for disease-free survival (DFS) using Cox regression models were performed. RESULTS P53 status (HR 6.7, 95%CI 1.75-26.0, p=0.006) and MSI were the strongest single genetic prognostic factors for decreased DFS, while high PI3K-AKT pathway activation showed a trend and β-catenin was not prognostic. The combination of multiple activated pathways was the most powerful prognostic factor for decreased DFS (HR 5.0; 95%CI 1.59-15.6 p=0.006). Multiple pathway activation, found in 8% of patients, was strongly associated with aggressive clinical course. In contrast, 40% of patients had no alterations in the investigated pathways and had a very low risk of disease progression. CONCLUSIONS Activation of multiple oncogenic pathways in EEC was the most powerful prognostic factor for decreased DFS, resulting in an individual risk assessment superior to the current approach based on clinicopathological factors.