Abstract
Although the majority of low-grade, early-stage endometrial cancer patients have good survival with surgery alone, patients who recur tend to do poorly. Identification of patients at high risk of recurrence who would benefit from adjuvant treatment or more extensive surgical staging would help optimize individualized care of endometrial cancer patients. CTNNB1 (encodes β-catenin) mutations identify a subset of low-grade, early-stage endometrial cancer patients at high risk of recurrence. Mutation of CTNNB1 exon 3 is classically associated with translocation of the β-catenin protein from the membrane to the nucleus and activation of Wnt/β-catenin signaling. Given the clinical utility of identifying endometrial carcinomas with CTNNB1 mutation, the purpose of this study was to determine if immunohistochemistry could act as a surrogate for CTNNB1 gene sequencing. Next-generation sequencing was performed on 345 endometrial carcinomas. Immunohistochemical localization of β-catenin was determined for 53/63 CTNNB1 exon 3 mutant tumors for which tissue was available and a subset of wild-type tumors. Nuclear localization of β-catenin had 100% specificity in distinguishing CTNNB1 mutant from wild type, but sensitivity was lower (84.9%). Nearly half of CTNNB1 mutant cases had only 5-10% of tumor cells with β-catenin nuclear localization. The concordance between pathologists blinded to mutation status in assessing nuclear localization was 100%. The extent of β-catenin nuclear localization was not associated with specific CTNNB1 gene mutation, tumor grade, presence of non-endometrioid component, or specific concurrent gene mutations in the tumor. For comparison, nuclear localization of β-catenin was more diffuse in desmoid fibromatosis, a tumor also associated with CTNNB1 mutation. Thus, nuclear localization of β-catenin assessed by immunohistochemistry does not detect all endometrial cancers with CTNNB1 gene mutation. The extent of nuclear localization may be tumor type dependent. For endometrial cancer, immunohistochemistry could be an initial screen, with CTNNB1 sequencing employed when nuclear localization of β-catenin is absent.