A fully next-generation sequencing-based method of classifying molecular sub-types of endometrial cancer retains prognostic value and expands biomarker targets

Objective: The molecular classification of endometrial cancer into POLE–ultra-mutated, mismatch repair-deficient, p53-mutated, and no specific molecular profile sub-types has significant prognostic value and is recommended in the evaluation of all patients with endometrial cancer. Nonetheless, there has been inconsistent clinical implementation. One possible barrier is the current practice of using several different assays, each with its own result, that subsequently need to be integrated. We developed a single, fully next-generation sequencing (NGS)-based assay that sub-types endometrial samples and evaluated it against an existing algorithm. Methods: Molecular sub-typing was retrospectively performed on 142 formalin-fixed, paraffin-embedded endometrial cancer samples using a clinically validated NGS panel targeting 133 genes and additional loci for micro-satellite instability and tumor mutational burden. In parallel, these same samples were classified by the World Health Organization algorithm using mismatch repair and p53 immunohistochemistry, and POLE sequencing. Concordance between algorithms was assessed, and the prognostic value of each was evaluated. We further explored racial disparities in outcomes and assessed other prognostic and potentially therapeutic biomarkers. Results: The sequencing-based method was highly concordant with the World Health Organization algorithm (136/142 patients, Cohen's κ = 0.94) and retained prognostic value, with a significant difference in overall survival and progression-free survival by sub-type, and similar stratification to that initially identified in The Cancer Genome Atlas analysis. In addition, 11 cases had sequence variants in the previously established prognostic biomarker CTNNB1, and 3 cases had potentially targetable sequence variants in ERBB2. Endometrial cancer outcomes in Black individuals were worse, in part owing to the increased incidence of sub-types with a poor prognosis. Conclusions: A fully sequencing-based assay streamlines molecular classification of endometrial cancer and retains the prognostic value of other validated methods, which may aid clinical implementation. The additional genomic information obtained with an NGS panel, beyond the classification markers, can broaden potentially applicable therapies.