Molecular characterization in the prediction of disease extent in endometrial carcinoma

Anne M Kolehmainen,Annukka M Pasanen,Riitta L Koivisto-Korander,Ralf C Bützow,Mikko J Loukovaara

DOI: https://doi.org/10.1016/j.ejogrb.2020.10.031

Abstract:Objective: Patients with endometrial carcinoma are usually triaged to staging lymphadenectomy selectively based on estimated risk of lymphatic spread. The risk is generally assessed by the presence of uterine risk factors, but their preoperative and intraoperative identification remain a challenge. The objective of this study was to assess the capability of molecular classification, described by The Cancer Genome Atlas (TCGA), to predict the stage of endometrial carcinoma. Study design: Sequencing of polymerase-ε (POLE) and immunohistochemistry of mismatch repair (MMR) proteins and p53 were performed to stratify endometrial carcinomas into subgroups of POLE exonuclease domain mutation (EDM), MMR deficiency, abnormal p53 (p53 abn) and 'no specific molecular profile' (NSMP). NSMP was the reference subgroup for comparisons. Associations of molecular subgroups and uterine risk factors with stage were examined in univariable and multivariable analyses. Results: Six hundred and four patients were included in the study. None of the POLE EDM tumours extended beyond the uterine cervix. In an unadjusted analysis, p53 abn was associated with increased risk for stage IIIC-IV disease [odds ratio (OR) 4.6, 95% confidence interval (CI) 2.3-9.2; p < 0.0005]. When controlling for uterine risk factors (histotype and grade, depth of myometrial invasion, tumour size, lymphovascular space invasion), p53 was not an independent predictor of advanced disease. In contrast, POLE EDM independently predicted local disease (OR 0.12, 95% CI 0.015-0.99; p = 0.049 for stage II-IV cancer). Of the molecular subgroups, p53 abn was most strongly associated with the presence of high-risk uterine factors (ORs between 2.2 and 19; p ≤ 0.010). Conclusion: Of the TCGA-based molecular subgroups, POLE EDM independently predicted early-stage endometrial carcinoma. Although p53 abn was not an independent predictor of advanced disease, its association with uterine risk factors could allow utilization of molecular data in deciding the type of staging surgery if knowledge of uterine factors is deficient.