Bridging morphology and genomics: p53, mismatch repair proteins, and β-catenin as immunohistochemistry surrogates for risk stratification in endometrial carcinoma

ABSTRACT Background: Endometrial carcinoma (EC) is the most frequent gynecologic cancer in developed countries, with considerable biological heterogeneity. The Cancer Genome Atlas molecular classification provides important prognostic information, but genomic profiling is not always practical. Immunohistochemistry (IHC)-based surrogates using p53, mismatch repair (MMR) proteins, and β-catenin offer a practical alternative. Objective: This study aims to evaluate the prognostic value of an expanded IHC-based molecular panel (p53, MMR, β-catenin) for risk stratification, with a focus on identifying high-risk patients within the heterogeneous “no specific molecular profile” (NSMP) subgroup. Methods: This retrospective cohort study included 150 EC patients treated between 2018 and 2022. IHC classified cases as p53-abnormal (p53abn), MMR-deficient (MMRd), or p53-wild type (NSMP). Nuclear β-catenin (surrogate for CTNNB1 mutation) was assessed. Survival analysis used Kaplan-Meier and Cox regression. Results: Nuclear β-catenin positivity was found in 22.0% (33/150) of all cases and in 24.8% (26/105) of NSMP tumors. It was associated with lymphovascular space invasion (LVSI; 69.7% vs. 24.0%; P < 0.001) and nodal metastasis (24.2% vs. 8.5%; P = 0.02). Within the NSMP subgroup, β-catenin-positive patients had significantly worse 3-year disease-free survival (68% vs. 89%; P = 0.003). In multivariable analysis, β-catenin positivity was an independent predictor of recurrence (hazard ratio = 2.2 [95% CI, 1.1–4.5]; P = 0.02), alongside LVSI and non-endometrioid histology. Conclusion: An IHC panel incorporating p53, MMR, and β-catenin effectively stratifies EC risk. Nuclear β-catenin identifies a higher-risk subset within the prognostically heterogeneous NSMP group, supporting its integration into routine prognostic protocols to guide adjuvant therapy and surveillance, particularly in resource-limited settings.