Highlight
In a TCGA-based analysis of 209 endometrial cancers, composite expression of chemoresistance-associated genes strongly stratified 5-year progression-free survival, with the high-score group showing markedly worse outcomes.
High CCNA2 and E2F1 expression was tightly linked to transcriptional activation of DNA damage repair, cell-cycle, and antiapoptotic programs, providing biologic plausibility for platinum insensitivity.
The combined CCNA2 + E2F1 signal separated prognosis across histologic subtypes, TCGA molecular groups, and both clinicopathologic low- and high-risk disease, suggesting utility beyond conventional risk assessment.
The study supports a clinically relevant concept: pretreatment transcriptome profiling may identify patients unlikely to benefit from platinum-based chemotherapy and may direct them toward alternative targeted strategies.
Background and unmet clinical need
Endometrial cancer is the most common gynecologic malignancy in many high-income countries, and although many patients present with apparently localized disease, recurrence risk varies substantially. Current treatment decisions are driven by stage, histology, grade, depth of myometrial invasion, lymphovascular space invasion, and increasingly by molecular classification. Yet these factors do not directly identify whether an individual tumor is sensitive or insensitive to platinum-based chemotherapy (PbCT), the backbone of systemic treatment for advanced and recurrent disease.
This is a major clinical gap. Patients who receive ineffective platinum therapy are exposed to neuropathy, myelosuppression, fatigue, renal toxicity, and delays in transition to potentially more active therapies. In endometrial cancer, a pretreatment biomarker that predicts platinum insensitivity could improve counseling, trial selection, and therapeutic sequencing. The 2026 study by Gonzalez-Bosquet and colleagues addresses this problem by testing whether tumor transcriptome profiling can uncover gene-expression patterns linked to chemoresistance and poor prognosis.
The work is also timely in the context of contemporary molecular stratification. The Cancer Genome Atlas (TCGA) defined four broad endometrial cancer molecular groups: POLE-ultramutated, microsatellite instability/hypermutated, copy-number low, and copy-number high/serous-like. These categories have reshaped prognostic thinking, but they still do not fully resolve chemotherapy responsiveness for individual patients. A transcriptome-based chemoresistance signature could therefore complement rather than replace current classification systems.
Proposed section structure for interpreting this study
A clinically useful reading of this paper is best organized into the following sections: clinical context and unmet need; study design and cohort selection; biomarker construction and analytic strategy; key efficacy and biologic findings; clinical interpretation and integration with current practice; limitations and next research steps; and practical summary for oncology care teams.
Study design and methods
Data source and population
The investigators analyzed 209 endometrial cancer cases from The Cancer Genome Atlas with robust mRNA expression data. Cases with POLE variants were excluded from the chemoresistance-focused analysis, a reasonable choice given the unusually favorable prognosis and biologically distinct treatment behavior of POLE-mutated tumors.
Risk grouping
Patients were categorized as clinicopathologic high risk (CPHR) if they had stage III or IV disease, or stage I or II disease with myometrial invasion of at least 66%. The remaining cases were categorized as clinicopathologic low risk (CPLR). This framework allowed the authors to ask whether transcriptomic markers add discriminatory value beyond standard clinicopathologic features.
Transcriptomic analysis
The team evaluated expression of genes previously associated with chemoresistance. Expression data were normalized and log2 transformed. Pearson correlation was used to examine the relationship between gene-expression patterns and 5-year progression-free survival (PFS). Survival differences were tested with log-rank methods, and continuous variables were compared by t-test.
Endpoints
The principal clinical endpoint was 5-year PFS. The central translational objective was to identify a transcriptomic profile that could plausibly indicate pretreatment platinum insensitivity while also nominating alternative therapeutic targets.
Key findings
Composite chemoresistance score stratified prognosis
The study’s first major result was the strength of the composite chemoresistance-associated gene score. Five-year PFS was 84.4% in tumors with low scores, 74.7% in those with intermediate scores, and only 23.0% in those with high scores. The gradient is clinically meaningful, not merely statistically interesting. A decline from 84.4% to 23.0% implies a biologically distinct subset with markedly worse disease control despite standard treatment paradigms.
Although the abstract does not provide hazard ratios or confidence intervals, the magnitude of the separation suggests that the signature captures aggressive biology beyond traditional staging alone. For a practicing clinician, this type of signal raises the possibility that some patients currently viewed as appropriate for platinum-based treatment may actually harbor tumors with a transcriptional program favoring treatment escape.
CCNA2 and E2F1 emerged as central biomarkers
Among the genes examined, the transcription factor axis involving CCNA2 and E2F1 stood out. High expression of CCNA2 and E2F1 significantly correlated with increased expression of genes involved in DNA damage repair, cell-cycle progression, and antiapoptosis, with most correlation coefficients exceeding 0.80. That degree of correlation is unusually strong in tumor transcriptomic analyses and supports the idea that these markers are not isolated observations but part of a broader coordinated resistance network.
Biologically, this makes sense. E2F1 is a major regulator of cell-cycle entry and DNA replication programs, and CCNA2 is a key cyclin involved in S phase and G2/M transition. Tumors with sustained activation of these pathways may tolerate platinum-induced DNA damage more effectively, continue proliferating under genotoxic stress, and evade apoptosis. In practical terms, the biomarker pair appears to summarize a high-proliferation, high-repair, high-survival state that is unfavorable for platinum response.
CCNA2 + E2F1 expression strongly separated progression-free survival
Independent of most traditional risk factors, low versus high CCNA2 + E2F1 expression stratified patients by 5-year PFS at 89% versus 50%, respectively, with all reported comparisons achieving P values below .001. This is one of the most clinically compelling findings in the paper because it suggests that the transcriptomic marker is not simply recapitulating established risk variables but adding information on top of them.
The breadth of the effect is equally important. The low versus high CCNA2 + E2F1 signal also separated outcomes in endometrioid and serous histologies, across TCGA subgroups, and within clinicopathologic risk strata. This cross-context consistency improves confidence that the finding is not restricted to a single subtype.
Utility in both low- and high-risk clinicopathologic groups
In CPLR disease, low versus high CCNA2 + E2F1 expression corresponded to 5-year PFS of 89% versus 59%. In CPHR disease, the corresponding 5-year PFS was 87% versus 39%. These results may be among the most practice-relevant in the study. Conventional low-risk classification often reassures patients and clinicians, yet a 59% 5-year PFS in the transcriptomically high-expression subset suggests that some apparently favorable cases still harbor biologically aggressive disease. Conversely, the 87% PFS in transcriptomically low-expression CPHR cases raises the possibility that not all clinicopathologic high-risk tumors behave equally poorly.
This kind of bidirectional refinement is precisely what precision oncology aims to achieve: identifying unexpectedly high-risk patients who need treatment intensification or trial enrollment, while also recognizing patients whose tumors may still be platinum-sensitive despite adverse anatomic features.
Clinical interpretation
Why this matters for platinum-based chemotherapy
The authors conclude that CPHR cases, and CPLR cases that recur, with low CCNA2 + E2F1 expression likely have platinum sensitivity, whereas analogous cases with high expression likely have platinum insensitivity. This is an important translational hypothesis. Today, many patients are treated empirically with carboplatin-based regimens because reliable predictive biomarkers are lacking. A pretreatment transcriptomic tool could potentially spare selected patients from ineffective platinum exposure and accelerate the use of other approaches.
Still, the wording should be interpreted carefully. The study correlates gene expression with prognosis and inferred chemoresistance biology; it does not directly prove lack of platinum benefit in a prospective treatment-comparison framework. Therefore, the findings are best viewed as hypothesis-generating and clinically promising rather than immediately practice-changing.
Potential treatment implications
The mechanistic associations identified in this study suggest several therapeutic directions. Tumors with high CCNA2 and E2F1 expression and accompanying activation of DNA repair and cell-cycle pathways may be candidates for trials of cell-cycle inhibitors, DNA damage response inhibitors, or rational combinations with immunotherapy or antibody-drug conjugates, depending on molecular context. The antiapoptotic signature also raises interest in agents targeting survival pathways, although the paper does not itself test any such interventions.
From a multidisciplinary standpoint, this could eventually influence recurrent-disease planning, adjuvant trial design, and molecular tumor board discussions. If validated, the assay might help answer a common question in advanced endometrial cancer: should the next systemic step remain platinum-centered, or should an alternative strategy be prioritized?
Relationship to current evidence and guidelines
Modern endometrial cancer management increasingly incorporates molecular data. ESGO/ESTRO/ESP guidance and other international recommendations support molecular classification using POLE mutation status, mismatch repair deficiency, p53 abnormality, and no specific molecular profile groupings to improve prognostication and adjuvant treatment decisions. The present study extends that framework by focusing not primarily on prognosis alone, but on putative chemotherapy sensitivity.
The idea is consistent with broader oncology experience: proliferative and DNA repair-enriched transcriptomic states can influence response to cytotoxic therapy, but the exact direction of effect is tumor-type dependent. In endometrial cancer, platinum resistance remains clinically recognized but poorly defined at the molecular level. This paper therefore contributes a plausible biomarker framework centered on CCNA2 and E2F1.
At the same time, current guidelines do not recommend routine transcriptome profiling specifically to determine platinum sensitivity in endometrial cancer. Standard-of-care biomarker workup remains focused on mismatch repair status, p53, POLE, HER2 in selected serous or p53-abnormal tumors, and other context-specific markers. Adoption of CCNA2 and E2F1 into clinical care would require external validation, assay standardization, predefined cutoffs, and prospective demonstration that biomarker-guided treatment actually improves outcomes.
Strengths of the study
This analysis has several notable strengths. First, it uses TCGA, a deeply characterized and publicly accessible dataset with integrated clinicopathologic and molecular information. Second, the investigators identify not just a prognostic signal but a biologically coherent network linking CCNA2 and E2F1 to DNA repair, cell cycle, and antiapoptotic programs. Third, the biomarker retained stratification value across multiple clinically relevant subgroups, including histology, molecular subtype, and clinicopathologic risk category. Finally, the conclusions are framed in a translationally meaningful way, moving beyond descriptive genomics toward treatment relevance.
Limitations and cautionary points
Several limitations should temper interpretation. The study is retrospective and observational. TCGA was not designed as a prospective predictive biomarker trial for chemotherapy response, and treatment details, sequencing, and recurrence management may not be sufficiently granular to isolate true platinum-specific effects. The sample size, while respectable for transcriptomic work in endometrial cancer, is still modest once stratified into multiple subgroups.
The abstract also does not report multivariable hazard ratios, confidence intervals, calibration metrics, or model discrimination statistics such as C-index, which would help quantify incremental predictive value over existing clinical and molecular classifiers. In addition, transcriptomic signatures can be sensitive to platform effects, normalization strategy, tumor purity, and intratumoral heterogeneity. These factors matter greatly if one hopes to transfer a biomarker from a research dataset to a reproducible clinical assay.
Another issue is the distinction between prognosis and prediction. A marker associated with poor PFS is not automatically a marker of platinum insensitivity. Tumors with high CCNA2 and E2F1 may simply be more aggressive overall. To establish predictive utility, one would ideally compare outcomes with platinum-based versus non-platinum strategies according to biomarker status in prospective or carefully controlled retrospective cohorts.
Implications for future research
The next step should be external validation in independent endometrial cancer cohorts with detailed treatment annotation, including response rates, progression on platinum, and platinum-free interval. Prospective-retrospective analyses from randomized trials could be especially informative. A clinically deployable assay would also need analytically validated cutoffs for low and high CCNA2 + E2F1 expression.
Mechanistically, laboratory models should test whether CCNA2/E2F1-high endometrial cancers are functionally resistant to carboplatin and whether they are selectively vulnerable to inhibitors of CDK signaling, ATR/CHK1, WEE1, PARP in selected contexts, or apoptosis-regulating pathways. Finally, biomarker-enriched clinical trials are needed to determine whether transcriptome-guided treatment selection improves patient-centered outcomes compared with current empiric approaches.
Practical takeaway for clinicians
For gynecologic oncologists, medical oncologists, and molecular tumor boards, this study suggests that pretreatment transcriptome profiling may eventually help distinguish two clinically important populations in endometrial cancer: one with a relatively favorable, likely platinum-sensitive biology, and another with a high-risk transcriptional program centered on CCNA2 and E2F1 that may predict poor disease control and possible platinum insensitivity.
Today, the findings are best used as a framework for critical appraisal and for hypothesis generation in research settings, not as a standalone basis for withholding platinum therapy. However, they offer a persuasive argument that transcriptome-informed care could become the next layer of precision medicine in endometrial cancer, particularly for patients whose clinicopathologic features do not fully explain their outcomes.
Conclusion
Gonzalez-Bosquet and colleagues present a clinically relevant transcriptomic analysis showing that chemoresistance-associated gene expression, especially the CCNA2 + E2F1 axis, strongly correlates with 5-year progression-free survival in endometrial cancer. The study provides biologic plausibility for a platinum-insensitive phenotype defined by upregulated DNA repair, cell-cycle, and antiapoptotic pathways. If externally validated and prospectively tested, this approach could refine prognostication, identify patients unlikely to benefit from platinum-based chemotherapy, and open the door to more rational precision therapeutics.
Funding and ClinicalTrials.gov
Funding information was not provided in the abstract supplied here. No ClinicalTrials.gov registration number is applicable for this TCGA-based retrospective transcriptomic analysis.
Citation
Gonzalez-Bosquet J, Shahi M, Yadav S, Kanwar N, Tapia AL, McGree ME, Dowdy SC, Couch FJ, Halling KC, Weroha SJ, Bakkum-Gamez JN, Podratz KC. Clinical relevance of transcriptome profiling for assessing prognosis, platinum insensitivity, and precision therapeutics in endometrial cancer. Gynecologic Oncology. 2026-06-02;210:114-123. PMID: 42229042. URL: https://pubmed.ncbi.nlm.nih.gov/42229042/
Selected references
Kandoth C, Schultz N, Cherniack AD, et al. Integrated genomic characterization of endometrial carcinoma. Nature. 2013;497(7447):67-73.
Concin N, Matias-Guiu X, Vergote I, et al. ESGO/ESTRO/ESP guidelines for the management of patients with endometrial carcinoma. International Journal of Gynecological Cancer. 2021;31(1):12-39.
León-Castillo A, de Boer SM, Powell ME, et al. Molecular classification of the PORTEC-3 trial for high-risk endometrial cancer: impact on prognosis and benefit from adjuvant therapy. Journal of Clinical Oncology. 2020;38(29):3388-3397.
Oaknin A, Tinker AV, Gilbert L, et al. Clinical activity and safety of the anti-PD-1 monoclonal antibody dostarlimab for patients with recurrent or advanced mismatch repair-deficient endometrial cancer: a non-randomized phase 1 clinical trial. JAMA Oncology. 2020;6(11):1766-1772.