Prognosis Stratification Tools in Early-Stage Endometrial Cancer: Could We Improve Their Accuracy?

Interpretable prognostic modeling of endometrial cancer

Endometrial carcinoma (EC) is one of the most common gynecological cancers in the world. In this work we apply Cox proportional hazards (CPH) and optimal survival tree (OST) algorithms to the retrospective prognostic modeling of disease-specific survival in 842 EC patients. We demonstrate that linear CPH models are preferred for the EC risk assessment based on clinical features alone, while interpretable, non-linear OST models are favored when patient profiles can be supplemented with additional biomarker data. We show how visually interpretable tree models can help generate and explore novel research hypotheses by studying the OST decision path structure, in which L1 cell adhesion molecule expression and estrogen receptor status are correctly indicated as important risk factors in the p53 abnormal EC subgroup. To aid further clinical adoption of advanced machine learning techniques, we stress the importance of quantifying model discrimination and calibration performance in the development of explainable clinical prediction models.

Morules but Not Squamous Differentiation are a Reliable Indicator of CTNNB1 (β-catenin) Mutations in Endometrial Carcinoma and Precancers

Although collectively regarded as "squamous differentiation (SD)" in endometrial endometrioid carcinoma (EEC) and atypical hyperplasia/endometrioid intraepithelial neoplasia (AH/EIN), morules (often referred to as "squamous morules") and true SD may represent two distinct phenomena. Here, we explored the distinction between morules versus SD and investigated the association of morules and SD with CTNNB1 mutations. A total of 270 cases of EEC and AH/EIN were studied, including EEC with (n=36) or without (n=36) morules and AH/EIN with (n=80) or without (n=118) morules. Cases were analyzed by immunohistochemistry and selected cases (n=20) by targeted next-generation sequencing panel. Near-perfect agreement was found between morules and glandular β-catenin nuclear staining in AH/EIN and EEC. A strong positive association was found between morules and glandular β-catenin nuclear staining (P<0.0001, Φ=0.59 in AH/EIN; P<0.0001, Φ=0.85 in EEC). There was no association between (1) morules and glandular PAX2 or PTEN aberrant expression or (2) SD and aberrant expression of β-catenin, PAX2 or PTEN (Φ=0.09, β-catenin; Φ=0.16, PAX2; Φ=0.13, PTEN). CTNNB1 mutations were identified in all 20 selected morule-containing cases (100%). Next-generation sequencing was performed on 2 (preprogestin and postprogestin treatment) biopsies from 1 patient, revealing identical mutational profile in morules and glands. In conclusion, (1) SD and morules are distinct biological phenomena; (2) the presence of morules, but not SD, is a reliable indicator of CTNNB1 mutations in EEC and AH/EIN. Our findings demonstrate that SD and morules are distinct biological phenomena. Since morules but not SD are associated with β-catenin mutations, the distinction is clinically relevant and should be included in diagnostic reports.