One DNA Test Could Transform How Doctors Classify Endometrial Cancer

Endometrial carcinoma is the most common gynecological cancer in developed countries and the second most common malignant tumor of the female reproductive system in China, yet the way doctors classify it has been undergoing a quiet revolution. Since 2013, when The Cancer Genome Atlas project revealed that endometrial tumors fall into distinct molecular groups with dramatically different prognoses, pathology laboratories worldwide have been racing to implement molecular classification. Now, a new study from Fujian Medical University in China suggests that a single next-generation sequencing test can do the job of multiple platforms, with higher sensitivity than the traditional approach and a bonus: it simultaneously uncovers mutations that could guide targeted therapies for advanced disease.

The study, published in Pathology and Oncology Research, evaluated a targeted DNA sequencing panel covering 116 cancer-related genes against the established Sanger sequencing plus immunohistochemistry (IHC) method for classifying endometrial carcinoma. The research team, led by Lingfeng Chen, Zhijie You, and Jie Lin at Shengli Clinical Medical College of Fujian Medical University, analyzed 194 tumor samples from patients who underwent surgical resection between January 2022 and December 2024. Their goal was ambitious: to prove that one clinically validated sequencing test could replace the multi-platform workflow that has made molecular classification inconsistent and costly in clinical practice.

The molecular classification system at the heart of this work traces back to the TCGA framework and has been refined into the ProMisE classifier, which sorts endometrial carcinomas into four subtypes: POLE ultramutated, mismatch repair deficient, p53 abnormal, and non-specific molecular profile (NSMP). Each subtype carries distinct prognostic implications. POLE-mutated tumors, for instance, tend to have favorable outcomes despite their high mutation burden, while p53-abnormal tumors are typically aggressive. The 2020 WHO classification of female genital tumors incorporated these molecular criteria, and the 2023 FIGO staging recommendations now advocate comprehensive molecular testing for all endometrial carcinoma patients.

But implementing this classification has been anything but straightforward. The traditional approach requires combining Sanger sequencing of the POLE gene with IHC staining for mismatch repair proteins and p53, a multi-platform process that is complex and expensive. Approximately 3 percent of cases exhibit overlapping molecular characteristics, complicating accurate subtype determination. The NSMP subtype, sometimes called the junk category because it encompasses tumors lacking POLE mutations, MMR deficiencies, and p53 abnormalities, accounts for between 30 and 47 percent of cases in Western populations and even higher proportions in Chinese populations. This heterogeneous group includes aggressive variants such as dedifferentiated endometrial carcinoma with SMARCA4 loss, which have poor prognoses despite the absence of TP53 mutations.

The results of the new study are striking. The concordance between NGS-based and Sanger plus IHC classifications was 93.8 percent, with a kappa coefficient of 0.908, indicating almost perfect agreement. When the researchers excluded seven discordant POLE mutations, concordance improved to 97.4 percent with a kappa of 0.962. The seven excluded cases were all POLE mutations that Sanger sequencing had missed because of its lower sensitivity for variants with mutation allele frequencies below 10 percent. NGS identified 30 POLE mutations compared to only 23 detected by Sanger sequencing, demonstrating the superior sensitivity of the sequencing panel.

The technical explanation for these discrepancies reveals the fundamental limitations of Sanger sequencing. The seven missed POLE mutations had allele frequencies ranging from 3.2 to 7.2 percent, far below the roughly 20 percent detection threshold of Sanger sequencing under routine conditions. These cases had low tumor content, around 20 percent, and when the researchers performed tumor region delineation and enrichment based on hematoxylin and eosin staining before DNA extraction, Sanger sequencing confirmed the POLE mutations in all seven cases. This finding underscores a critical quality control issue: when paraffin tumor samples are not quality-controlled for tumor cell content, Sanger sequencing may miss POLE mutations, incorrectly classifying POLE-mutated tumors as p53-abnormal or mismatch repair deficient, leading to erroneous prognostic predictions and overtreatment.

The study also examined the concordance between NGS-based microsatellite instability (MSI) analysis and IHC-based mismatch repair status, finding high agreement at 97.9 percent. However, NGS-based TP53 mutation detection showed only moderate agreement with p53 IHC results, with a kappa of 0.688. Most discrepancies occurred within the POLE-mutated and MMR-deficient subtypes, where TP53 mutations are often passenger mutations that do not influence classification. When these subtypes were excluded, the agreement improved dramatically to 97.5 percent, underscoring the importance of sequential molecular classification steps in improving diagnostic precision.

Beyond classification, the NGS panel revealed a wealth of mutations associated with targeted therapy trials. Among 131 samples analyzed with the comprehensive panel, mutations in PTEN were found in 76.3 percent of cases, PIK3CA in 50.4 percent, and ARID1A in 35.9 percent. Additional mutations with frequencies exceeding 10 percent included KRAS at 25.2 percent and CTNNB1 at 17.6 percent. These findings are clinically significant because PI3K/AKT pathway mutations are actionable targets for inhibitors such as alpelisib and capivasertib, while HER2 amplifications identified in p53-abnormal subtypes can inform the use of trastuzumab, which has been incorporated into clinical guidelines for HER2-positive advanced or recurrent serous endometrial carcinoma. BRCA1/2 mutations open avenues for PARP inhibitor therapy.

The distribution of gene mutations varied significantly across tumor grades and stages. PTEN mutations decreased from 86.8 percent in grade 1 tumors to 59.3 percent in grade 3, while TP53 mutations increased markedly from 2.6 percent in grade 1 to 59.3 percent in grade 3. HER2 mutations rose from 0 percent in grade 1 to 11.4 percent in grade 3. In advanced-stage disease, HER2 mutations were significantly more frequent at 13.0 percent compared to 1.0 percent in early stages. These patterns reinforce the molecular heterogeneity of endometrial carcinoma and highlight how comprehensive profiling can refine prognostic assessment beyond what histopathology alone can achieve.

The implications of this study extend far beyond the laboratory. The researchers conclude that NGS-based molecular classification aligns well with the established Sanger plus IHC method while offering higher sensitivity, particularly for detecting low-frequency POLE mutations that are crucial for accurate subtype classification. The single-test approach streamlines the molecular classification process, reduces the complexity and cost associated with multi-platform testing, and simultaneously identifies mutations that may inform prognosis and therapeutic strategies for patients with advanced endometrial carcinoma. While the retrospective nature, limited cohort size, and single-center design of the study warrant further prospective validation, the findings support the integration of NGS-based molecular classification into clinical practice to improve diagnostic precision and treatment outcomes for endometrial carcinoma patients worldwide.

Subject of Research: Molecular classification of endometrial carcinoma using a targeted NGS panel

Article Title: Molecular classification of endometrial carcinoma: clinical utility of an NGS panel with targeted detection of 116 cancer-related genes

Article References: Molecular classification of endometrial carcinoma: clinical utility of an NGS panel with targeted detection of 116 cancer-related genes. (n.d.). https://doi.org/10.3389/pore.2026.1612465

Image Credits: AI Generated

DOI: 10.3389/pore.2026.1612465

Keywords: endometrial carcinoma, next-generation sequencing, molecular classification, POLE mutations, precision oncology, cancer genetics, Molecular, classification, endometrial, carcinoma, clinical, utility