Phức tạp phân tử của ung thư nội mạc tử cung: Cách các mẫu sinh hiệu thay đổi qua các giai đoạn khối u, mô học và kiểu phân tử

Background: The Molecular Landscape of Endometrial Cancer

Endometrial cancer ranks among the most commonly diagnosed gynecologic malignancies worldwide, with incidence rates steadily increasing over the past several decades. While traditional classification systems have relied heavily on histological features—distinguishing between endometrioid, serous, clear cell, and other subtypes—advances in genomic technologies have revealed that endometrial cancer represents a remarkably heterogeneous disease at the molecular level. This molecular diversity has profound implications for prognosis, treatment selection, and patient outcomes.

The Cancer Genome Atlas (TCGA) project fundamentally transformed our understanding of endometrial cancer biology by identifying four distinct molecular subtypes: POLE ultramutated, microsatellite instability-high (MSI-H), copy-number low (endometrioid), and copy-number high (serous-like). Each subtype carries unique mutation signatures, recurrence risks, and therapeutic vulnerabilities. However, many clinicians question how these molecular classifications, derived from controlled research cohorts, translate to the complex, heterogeneous populations encountered in routine clinical practice.

Real-world evidence offers a critical complement to randomized controlled trials by capturing the full spectrum of patient characteristics, comorbidities, and treatment patterns that influence clinical outcomes. The study by Paranjpe and colleagues addresses this gap by examining molecular biomarker prevalence across a large, contemporary cohort of endometrial cancer patients diagnosed in routine clinical settings.

Study Design and Patient Population

This retrospective cohort study utilized de-identified clinical and tumor sequencing data from the Oncology Research Information Exchange Network (ORIEN), a collaborative network of academic cancer centers across the United States. The study included patients diagnosed with endometrial cancer between 2006 and 2020 who underwent next-generation sequencing (NGS) testing for at least one biomarker.

The biomarkers evaluated in this study encompassed the most clinically relevant molecular alterations in endometrial cancer: POLE mutations, microsatellite instability (MSI) status, TP53 mutations, PTEN mutations, PIK3CA mutations, ARID1A mutations, tumor mutational burden (TMB), ESR1 mutations, ERBB2 (HER2) mutations, and ERBB2 amplification. These biomarkers were selected based on their established roles in endometrial carcinogenesis, prognostic significance, and potential to guide therapeutic decision-making.

The primary objective was to describe the prevalence and distribution of these molecular biomarkers across different tumor stages, histological subtypes, and TCGA-defined molecular classifications. This approach enabled a comprehensive characterization of molecular heterogeneity in a real-world context, complementing the findings from the original TCGA project.

Patient Characteristics and Histological Distribution

The study cohort included 671 patients with a mean age of 62.4 years. This age distribution aligns with the known epidemiology of endometrial cancer, which predominantly affects postmenopausal women. The histological composition of the cohort reflected the overall distribution of endometrial cancer subtypes in the general population, with the majority of patients (76%) presenting with endometrioid histology.

Endometrioid adenocarcinoma, the most common histological subtype, typically arises from the glandular epithelium of the endometrium and is often associated with unopposed estrogen exposure, obesity, and metabolic dysfunction. These tumors generally carry a more favorable prognosis compared to non-endometrioid subtypes, which include serous carcinoma, clear cell carcinoma, and undifferentiated carcinoma. Non-endometrioid subtypes, while less common, are associated with more aggressive behavior, higher rates of recurrence, and poorer survival outcomes.

Biomarker Prevalence and Patterns

The analysis revealed substantial variation in biomarker prevalence across the study cohort. Among the evaluated biomarkers, PTEN mutations emerged as the most prevalent alteration, detected in 65.9% of patients. This finding aligns with the established role of PTEN loss as a hallmark of endometrioid endometrial cancer, where it contributes to constitutive activation of the PI3K/AKT signaling pathway and drives tumor growth.

ARID1A mutations were identified in 52.0% of patients, making it the second most prevalent biomarker. ARID1A encodes a component of the SWI/SNF chromatin remodeling complex, and its loss has been implicated in the pathogenesis of both endometrial and ovarian clear cell carcinomas. PIK3CA mutations, which occur in the catalytic subunit of PI3K, were found in 39.2% of patients, consistent with the high frequency of PI3K pathway alterations reported in previous genomic studies of endometrial cancer.

TP53 mutations were detected in 25.3% of the overall cohort, but their distribution varied dramatically when stratified by histology and stage. In contrast, ERBB2 amplification—a potentially actionable biomarker with therapeutic implications—was relatively uncommon overall (4.8%) but showed marked enrichment in specific patient subgroups.

Molecular Heterogeneity by Histology

Stratification by histological subtype revealed striking differences in biomarker patterns. Among patients with non-endometrioid histology, TP53 mutations were present in 64.2% of cases—a substantially higher prevalence than the 16.8% observed in endometrioid tumors. This finding reflects the distinct molecular pathogenesis of serous and other high-grade non-endometrioid carcinomas, which are characterized by frequent TP53 mutations and chromosomal instability.

Similarly, ERBB2 amplification showed a pronounced predilection for non-endometrioid tumors, with a prevalence of 13.2% compared to just 2.5% in endometrioid cases. This differential distribution has important therapeutic implications, as ERBB2 amplification identifies a subset of patients who may benefit from HER2-targeted therapies such as trastuzumab or the antibody-drug conjugate trastuzumab deruxtecan.

Conversely, POLE mutations, PTEN alterations, and high tumor mutational burden (TMB-high) were more commonly associated with endometrioid histology. These biomarkers define the POLE ultramutated and MSI-H molecular subtypes, which exhibit distinctive immunological profiles and may be particularly susceptible to immune checkpoint inhibitor therapy.

Molecular Patterns Across Tumor Stage

Stage-dependent analysis demonstrated that advanced-stage disease (stages III-IV) was characterized by a significantly higher prevalence of TP53 mutations (41.4%) and ERBB2 amplification (11.2%) compared to early-stage disease. This stage-dependent accumulation of TP53 mutations likely reflects the progressive accumulation of genomic alterations during tumor progression and the selection pressure for more aggressive, genomically unstable clones.

In contrast, POLE mutations, PTEN alterations, and high TMB showed a predilection for early-stage disease. These findings suggest that the molecular subtypes associated with these biomarkers may follow a distinct clinical trajectory, with some tumors characterized by high initial mutational burden but relatively indolent behavior.

The correlation between molecular subtype and stage at diagnosis has important implications for screening and early detection strategies. Patients with POLE-mutant or MSI-H endometrial cancers may benefit from earlier clinical recognition, as these subtypes may have distinct clinical presentations or progression patterns.

Biomarker Distribution Across TCGA Molecular Subtypes

Integration with TCGA-defined molecular subtypes revealed distinct biomarker co-occurrence patterns. Within the POLE-positive subgroup, tumors exhibited high prevalence of PTEN mutations, ARID1A alterations, and elevated TMB. This constellation of findings defines the POLE ultramutated subtype, characterized by exceptionally high mutation rates and robust infiltration by tumor-infiltrating lymphocytes.

The MSI-H subgroup demonstrated similar patterns, with high prevalence of PTEN mutations, ARID1A alterations, and TMB-high status. MSI-H endometrial cancers harbor defects in DNA mismatch repair, leading to accumulation of frameshift mutations and neoantigen generation that renders them susceptible to immune checkpoint inhibitor therapy.

In stark contrast, TP53-mutated tumors—corresponding to the TCGA copy-number high subtype—showed the highest rates of ERBB2 amplification. This molecular profile defines a distinct disease entity with characteristic biological behaviors and therapeutic vulnerabilities. The co-occurrence of TP53 mutation and ERBB2 amplification identifies a particularly aggressive subgroup that may benefit from combined targeting of both pathways.

Clinical Implications and Therapeutic Opportunities

The molecular heterogeneity documented in this real-world cohort has direct implications for clinical management and therapeutic decision-making. The differential distribution of actionable biomarkers across histological subtypes and molecular classifications supports a personalized, biomarker-driven approach to endometrial cancer treatment.

For patients with MSI-H or POLE-mutant tumors, immune checkpoint inhibitors such as pembrolizumab have emerged as effective treatment options, with the FDA granting approval for pembrolizumab in tumors with high MSI or high TMB. The high prevalence of PTEN and ARID1A mutations in these subtypes may also inform combinations targeting the PI3K pathway or exploiting synthetic lethal relationships with ARID1A loss.

For patients with TP53-mutant, ERBB2-amplified tumors, HER2-targeted therapies represent a promising treatment strategy. While HER2 positivity has been extensively studied in breast and gastric cancers, its role in endometrial cancer is increasingly recognized. The 13.2% prevalence of ERBB2 amplification in non-endometrioid tumors represents a substantial patient population that may benefit from targeted approaches.

The study also highlights the importance of comprehensive molecular profiling beyond single-biomarker testing. As the therapeutic landscape expands to include targeted agents, immunotherapy, and antibody-drug conjugates, the need for comprehensive NGS panels that capture the full spectrum of relevant biomarkers becomes increasingly critical.

Expert Commentary and Study Limitations

While this real-world study provides valuable insights into the molecular landscape of endometrial cancer, several limitations warrant acknowledgment. First, the retrospective design introduces potential selection bias, as patients who underwent NGS testing may differ from the broader endometrial cancer population. Practices regarding molecular testing have evolved over the study period (2006-2020), and testing rates may have varied across institutions.

Second, the study relied on clinical records for staging and histological classification, without centralized pathology review. While ORIEN data undergo quality assurance procedures, variations in diagnostic criteria and interobserver variability in histological grading could influence the observed biomarker distributions.

Third, the availability of specific biomarkers varied across the patient population, as not all patients were tested for all biomarkers included in the analysis. This incomplete testing could affect the estimated prevalence of certain alterations and limit comparisons across biomarkers.

Despite these limitations, the study’s strengths—including its large sample size, real-world setting, and comprehensive biomarker panel—provide important context for understanding molecular heterogeneity in clinical practice. The findings complement and extend the TCGA data by demonstrating how molecular patterns manifest in the diverse patient populations encountered in routine oncology care.

Future Directions and Unmet Research Needs

The results of this study underscore several important questions for future investigation. First, longitudinal studies are needed to understand how molecular profiles evolve during disease progression and treatment. The differential distribution of biomarkers by stage suggests that molecular evolution occurs during tumor progression, and understanding these dynamics could inform optimal timing of molecular testing.

Second, clinical trials specifically designed for molecularly selected endometrial cancer populations are needed to validate the efficacy of targeted therapies in this disease. While basket trials have explored targeted approaches across multiple tumor types, endometrial cancer-specific cohorts remain limited.

Third, the integration of molecular data with clinical outcomes, including progression-free survival, overall survival, and treatment response, will be essential to fully characterize the prognostic and predictive significance of the biomarkers evaluated in this study.

Finally, health equity considerations must be addressed to ensure that the benefits of precision oncology reach all patients with endometrial cancer. Disparities in access to molecular testing, based on race, socioeconomic status, geographic location, or insurance coverage, could perpetuate inequities in cancer care.

Conclusion

This real-world analysis of 671 patients with endometrial cancer demonstrates that substantial molecular heterogeneity exists across tumor stages, histological subtypes, and TCGA molecular classifications. PTEN, ARID1A, and PIK3CA mutations predominate in endometrioid tumors and early-stage disease, while TP53 mutations and ERBB2 amplification cluster in non-endometrioid histology and advanced-stage disease.

The distinct biomarker patterns associated with different molecular subtypes have important implications for prognostic stratification and therapeutic selection. As the arsenal of targeted therapies and immunotherapies continues to expand, comprehensive molecular profiling should become a standard component of endometrial cancer management. This approach will enable clinicians to match patients with the most appropriate treatment strategies based on the unique molecular characteristics of their tumors.

The findings from this study reinforce the importance of translating molecular insights from research cohorts to real-world clinical practice. By embracing precision oncology approaches, we can move closer to delivering truly personalized care for patients with endometrial cancer.

Funding and Disclosures

This study utilized data from the Oncology Research Information Exchange Network (ORIEN). The authors disclosed potential conflicts of interest, and funding sources are noted in the original publication.

References

1. Paranjpe R, Chen C, Sun Y, Hong JL, Prabhu VS, Meng R, Dunshee DR, Duska LR. Molecular heterogeneity of endometrial cancer in the real-world: Biomarker patterns by tumor stage, histology, and molecular subtype. Gynecologic oncology. 2026-04-01;208:69-76. PMID: 41930527.

2. The Cancer Genome Atlas Research Network. Integrated genomic characterization of endometrial carcinoma. Nature. 2013;497(7447):67-73.

3. Makker V, MacKay H, Ray-Coquard I, et al. Endometrial cancer. Nat Rev Dis Primers. 2024;10(1):45.

4. Oaknin A, Bosse TJ, Creutzberg CL, et al. Endometrial cancer: ESGO-ESMO-ESTRO consensus conference recommendations. Ann Oncol. 2022;33(8):769-786.