Abstract
Objectives:
Mixed carcinoma of the endometrium (MCE) is an epithelial malignancy composed of two or more histological types. The recent WHO classification of MCE differs from its ovarian counterpart: dedifferentiated carcinomas are excluded from MCE, whereas ovarian dedifferentiated carcinomas are considered a special type of mixed carcinoma. However, the genetic basis of MCE remains poorly understood, as studies with comprehensive genomic analyses are scarce. This study aimed to identify genetic alterations involved in the development of MCE and endometrial dedifferentiated carcinomas (EDC) using next-generation sequencing (NGS) data and to determine whether MCE and EDC show distinct pathogenetic features supporting their classification as separate entities.
Methods:
We retrospectively reviewed 1,578 endometrial carcinomas treated at or referred to our institute between January 2017 and December 2025. Mixed carcinomas were identified by morphological examination and immunohistochemistry. NGS using the Oncomine Comprehensive Assay V3 was performed separately on each component of six MCEs. Although EDCs are not classified as mixed carcinomas according to WHO criteria, nine such cases were also analysed for comparison.
Results:
Excluding carcinosarcomas, MCEs accounted for 0.63% (10/1,578) of endometrial carcinomas, while EDCs represented 0.57% (9/1,578). With one exception, all MCEs and EDCs shared at least one identical pathogenic mutation between their components, supporting a clonal relationship. Among MCEs, 60% (6/10) showed mismatch repair deficiency (dMMR) in at least one component, and one case harboured a POLE mutation. Similarly, 89% (8/9) of EDCs were dMMR in at least one component, while one case was POLE-mutated. Both tumour groups frequently harboured alterations in the SWI/SNF and PI3K/AKT pathways, particularly involving ARID1A, PIK3CA, and PTEN. ARID1A mutations were especially frequent in EDCs (7/9, 78%).
Conclusion:
The components of almost all MCE and EDC were clonally related, suggesting development through transdifferentiation or divergent differentiation rather than independent collision. Despite their classification as separate entities in the current WHO system, both tumour types appear to arise predominantly in the setting of dMMR and genomic hypermutability. These findings provide further insight into the molecular pathogenesis of mixed carcinomas of the female genital tract and may contribute to future refinements in their classification.
Introduction
Mixed carcinomas of the female genital tract are defined as carcinomas comprising two or more distinct histological subtypes. Mixed carcinomas of the endometrium (MCE) are specifically recognised in the World Health Organization (WHO) Classification of Female Genital Tumours 5th edition, where they are defined as carcinomas containing at least one component of either serous or clear cell carcinoma. The diagnosis requires the identification of at least two spatially distinct histological types of endometrial carcinoma by histology and immunohistochemistry []. MCEs are rare, reported to represent less than 3% of endometrial malignancies. The most common mixed histological types are endometrioid/serous, serous/clear cell, and endometrioid/clear cell carcinoma []. It is recommended that the proportion of each component be reported, as even a small proportion of a high-grade component can result in unfavourable prognosis and influences therapeutical decisions [–].
Although the diagnosis is primarily based on morphological features, additional immunohistochemical studies may support the identification of distinct components. This diagnostic category should not, however, be applied to morphological variations of endometrioid, clear cell carcinoma, serous carcinoma, to carcinomas with ambiguous morphology, endometrial dedifferentiated carcinomas (EDCs), and carcinosarcomas [, ]. Although the WHO classification excludes EDCs from the category of MCE, the most common type of endometrial carcinoma with multiple components is EDC (well-differentiated carcinoma with an undifferentiated carcinoma component). This exclusion of EDCs from the category of MCE differs from that of ovarian carcinomas, where EDCs are included within the mixed carcinoma category in the current WHO classification [].
Based on genetic, epigenetic, and proteomic analyses, The Cancer Genome Atlas (TCGA) have proposed a molecular classification scheme for endometrial carcinomas []. According to this classification, endometrial carcinomas can be divided into four molecular groups: POLE-mutated (ultramutated), mismatch repair deficient (dMMR, hypermutated), TP53-mutated, and no specific molecular profile (NSMP) groups. However, there is limited data on the development, molecular background, and classification of mixed carcinomas.
Conceptually, mixed carcinomas may arise through three different mechanisms. These tumours can theoretically develop through divergent differentiation from a common progenitor, via transdifferentiation, or as collision tumours. It has been suggested that most MCEs arise through divergent differentiation and transdifferentiation, while true collision tumours are rare [, , , ]. However, to the best of our knowledge, only a limited number of small gene panel studies and case reports investigating the molecular characteristics of MCEs have been published.
The aim of this study was therefore two-fold: first, to identify common factors involved in the development of MCE, and second, to investigate whether pathogenetic differences exist between mixed and EDCs that would substantiate their classification as separate diagnostic categories within the endometrium. We therefore reviewed all MCE and EDC cases reported at our institute between January 2017 and December 2025 and separately analysed the tumour components and morphologically distinct areas using a next-generation sequencing (NGS) panel.
Materials and methods
Case selection
Between January 2017 and December 2025, 1,578 endometrial carcinomas were treated or referred for consultation to the National Institute of Oncology, Budapest, Hungary. Cases were classified as MCEs when at least two spatially distinct histological types were identified on morphology and immunohistochemistry, according to the current WHO criteria. We also identified EDCs, defined as tumours comprising a differentiated component (typically of FIGO grade 1 or 2 endometrioid carcinoma) and an undifferentiated component. Mixed tumours containing a pure mesenchymal component (carcinosarcoma) were excluded.
The different components were manually dissected, and DNA was extracted separately from each component of six MCEs and nine EDCs. Molecular analysis of each component was performed on the resection specimen, if available. In a few cases, certain components were only identifiable in the preoperative sample. In these cases, molecular analysis of these components was performed on the preoperative specimen. NGS was performed on each component.
Immunohistochemistry (IHC)
For each case, a basic panel including antibodies for oestrogen receptor (ER), progesterone receptor (PR), mismatch repair proteins (PMS2, MLH1, MSH6, and MSH2), and p53 was applied. In addition, for identification of a clear cell carcinoma component, NapsinA, AMACR (p504S), and HNF1-beta immunohistochemistry were performed. In those cases where a serous carcinoma component was present, p16 was ordered, and in EDCs, PAX8 and CK8-18 IHCs were added to the panel. Antibody clones and dilutions are listed in Supplementary Table S1. Each IHC reaction was performed according to the manufacturers’ instructions and was evaluated in accordance with the College of American Pathologists’ “Template for Reporting Results of Biomarker Testing of Specimens from Patients with Carcinoma of Gynecologic Origin” [].
Nucleic acid extraction
Prior to macrodissection, the area of interest was identified and marked by a pathologist on haematoxylin and eosin-stained slides. Macrodissection of the components was performed using the marked areas as a guide, processing one to three 5 µm sections from the formalin-fixed, paraffin-embedded tissue blocks. DNA was extracted using the ReliaPrep DNA Clean-Up and Concentration System Kit (Promega, Madison, WI, United States) according to the manufacturer’s protocol. As this system contains only purifying reagents, the deparaffinization and lysis steps were performed using the corresponding reagents according to the Maxwell RSC DNA formalin-fixed, paraffin-embedded protocol (Promega). The concentration of the DNA samples was measured using a Qubit 4 fluorometer with a Qubit™ 1X dsDNA HS (High Sensitivity) Assay Kit (Thermo Fisher Scientific, Waltham, MA, United States) and stored at 4 °C until further use.
Mutation detection methods
Libraries were prepared using the Ion Chef™ System with Ion 540™ Chips (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s instructions with a DNA input of approximately 8.5 ng. Sequencing was performed using an Ion S5™ Plus Sequencer (Thermo Fisher Scientific, Waltham, MA, USA). Ion Reporter™ Software (v. 5.18) (Thermo Fisher Scientific, Waltham, MA, USA) was used for data analysis. Oncomine Comprehensive Assay v3 DNA GX was used as the analysis workflow for all the samples. Only those detected mutations classified as pathogenic or likely pathogenic were included for the purpose of this study.
Results
Based on morphology and IHC, the overall frequency of MCEs was 0.63% (10 out of 1578 cases), while 0.57% (9 out of 1578 cases) for EDCs. Six patients with MCE (Cases M1-M6) and all 9 patients with EDCs were further analysed by NGS. In the remaining four patients with MCE, further NGS analysis could not be performed, as no material was available for nucleic acid extraction (Tables 1, 2; Supplementary Table S2). Among the MCEs, clear cell carcinoma admixed with low-grade endometrioid carcinoma was the most common histological type (4 cases). Additional combinations were serous carcinoma admixed with low-grade endometrioid carcinoma, and clear cell carcinoma with high-grade endometrioid carcinoma (2 cases each) (Figure 1), serous carcinoma admixed with clear cell carcinoma, and serous carcinoma with high-grade endometrioid carcinoma (1 case each) (Figure 2).
TABLE 1
| Case number | Histology | ER (%) | PR (%) | p53 | MMR proteins | NapsinA | AMACR | HNFb | p16 | NGS | TCGA MPGs | Final TCGA MPG |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Mixed endometrial carcinomas | ||||||||||||
| Case M1 | Clear cell carcinoma | 0 | 0 | Wild type | Retained | Focal | Focal | NP | NP | Yes | POLE mutated | POLE mutated |
| Low grade endometrioid carcinoma | 50 | 10 | Wild type | Retained | NP | NP | NP | NP | POLE mutated | POLE mutated | ||
| Case M2 | Clear cell carcinoma | 0 | 0 | Wild type | Retained | Positive | Positive | Positive | Positive | Yes | NSMP | NSMP |
| Low grade endometrioid carcinoma | 90 | 90 | Wild type | PMS2 and MLH1 loss | Negative | Negative | Negative | Negative | MMR deficient | MMR deficient | ||
| Case M3 | Clear cell carcinoma | 0 | 0 | Wild type | MSH6 loss | Positive | NP | Positive | NP | Yes | MMR deficient | MMR deficient |
| Low grade endometrioid carcinoma | 100 | 70 | Wild type | PMS2 and MLH1 loss | NP | NP | NP | NP | MMR deficient | MMR deficient | ||
| Case M4 | Clear cell carcinoma | 70 | 0 | Wild type | MSH6, MSH2 loss | Positive | NP | Positive | NP | Yes | MMR deficient | MMR deficient |
| Low grade endometrioid carcinoma | 10 | 15 | Wild tpye | MSH6, MSH2 loss | Negative | NP | NP | Focal | MMR deficient | MMR deficient | ||
| Case M5 | Serous carcinoma | 0 | 0 | Aberrant (overexpression) | PMS2 and MLH1 loss | NP | NP | NP | Diffuse | Yes | MMR deficient/p53-aberrant | MMR deficient |
| High grade endometrioid carcinoma | 100 | 90 | Wild type | PMS2 and MLH1 loss | NP | NP | NP | Focal | MMR deficient | MMR deficient | ||
| Case M6 | Clear cell carcinoma | 90 | 50 | Wild type | PMS2 loss | Positive | NP | NP | NP | Yes | MMR deficient | MMR deficient |
| High grade endometrioid carcinoma | 90 | 50 | Wild type | PMS2 and MLH1 loss | NP | NP | NP | NP | MMR deficient | MMR deficient | ||
| Case M7 | Serous carcinoma | 100 | 100 | Aberrant (overexpression) | Retained | NP | NP | NP | NP | No | NA | NA |
| Low grade endometrioid carcinoma | 100 | 100 | Wild type | Retained | NP | NP | NP | NP | NA | NA | ||
| Case M8 | Serous carcinoma | 70 | 60 | Aberrant (overexpression) | Retained | NP | NP | NP | NP | No | NA | NA |
| Low grade endometrioid carcinoma | 90 | 80 | Wild type | Retained | NP | NP | NP | NP | NA | NA | ||
| Case M9 | Serous carcinoma | 80 | 0 | Aberrant (overexpression) | Retained | Negative | NP | NP | NP | No | NA | NA |
| Clear cell carcinoma | 0 | 0 | Wild type | Retained | Positive | NP | NP | NP | NA | NA | ||
| Case M 10 | Clear cell carcinoma | 0 | 0 | Wild type | PMS2 and MLH1 loss | Negative | Negative | NP | NP | No | NA | NA |
| High grade endometrioid carcinoma | 13 | 2 | Wild type | PMS2 and MLH1 loss | NP | NP | NP | NP | NA | NA | ||
| Dedifferentiated carcinomas | ||||||||||||
| | Histology | ER (%) | PR (%) | p53 | MMR | CK8-18 | PAX8 | NP | NP | NGS | TCGA MPGs | Final TCGA MPG |
| Case D1 | Undifferentiated carcinoma | 0 | 0 | Wild type | PMS2 and MLH1 loss | Focal | Negative | NP | NP | Yes | MMR deficient | MMR deficient |
| Low grade endometrioid carcinoma | 95 | 50 | Wild type | PMS2 and MLH1 loss | Positive | Positive | NP | NP | MMR deficient | MMR deficient | ||
| Case D2 | Undifferentiated carcinoma | 0 | 0 | Wild type | PMS2 and MLH1 loss | Focal | Negative | NP | NP | Yes | MMR deficient | MMR deficient |
| Low grade endometrioid carcinoma | 100 | 100 | Wild type | Retained | NP | Positive | NP | NP | NSMP | NSMP | ||
| Case D3 | Undifferentiated carcinoma | 0 | 0 | Wild type | PMS2 and MLH1 loss | Focal | Negative | NP | NP | Yes | MMR deficient | MMR deficient |
| Low grade endometrioid carcinoma | 100 | 90 | Wild type | PMS2 and MLH1 loss | NP | NP | NP | NP | MMR deficient | MMR deficient | ||
| Case D4 | Undifferentiated carcinoma | 0 | 0 | Wild type | MSH6, MSH2, PMS2, and MLH1 loss | Focal | Negative | NP | NP | Yes | MMR deficient | MMR deficient |
| Low grade endometrioid carcinoma | 100 | 70 | Wild type | PMS2 and MLH1 loss | Positive | Positive | NP | NP | MMR deficient | MMR deficient | ||
| Case D5 | Undifferentiated carcinoma | 0 | 0 | Subclonal aberrant (overexpression) | PMS2 and MLH1 loss | NP | NP | NP | NP | Yes | MMR deficient/p53-aberrant | MMR deficient |
| Low grade endometrioid carcinoma | 90 | 90 | Wild type | PMS2 and MLH1 loss | NP | NP | NP | NP | MMR deficient | MMR deficient | ||
| Case D6 | Undifferentiated carcinoma | 0 | 0 | Wild type | PMS2 and MLH1 loss | Focal | Negative | NP | NP | Yes | MMR deficient | MMR deficient |
| Low grade endometrioid carcinoma | 100 | 90 | Wild type | PMS2 and MLH1 loss | NP | NP | NP | NP | MMR deficient | MMR deficient | ||
| Case D7 | Undifferentiated carcinoma | 0 | 0 | Wild type | PMS2, MLH1, and MSH6 loss | Focal | Negative | NP | NP | Yes | MMR deficient | MMR deficient |
| Clear cell carcinoma | 0 | 0 | Wild type | MSH6 loss | Positive | Positive | NP | NP | MMR deficient | MMR deficient | ||
| Case D8 | Undifferentiated carcinoma | 0 | 0 | Wild type | Retained | Positive | Negative | NP | NP | Yes | POLE mutated | POLE mutated |
| Low grade endometrioid carcinoma | 100 | 100 | Wild type | Retained | Positive | Positive | NP | NP | POLE mutated | POLE mutated | ||
| Case D9 | Undifferentiated carcinoma | 0 | 0 | Wild type | PMS2 and MLH1 loss | Positive | Negative | NP | NP | Yes | MMR deficient | MMR deficient |
| High grade endometrioid carcinoma | 90 | 90 | Wild type | PMS2 and MLH1 loss | Positive | Positive | NP | NP | MMR deficient | MMR deficient | ||
Histological, immunophenotypical characteristics, and TCGA group classification of each component in mixed and dedifferentiated endometrial carcinomas.
NP: not performed; NA: not applicable; MPG: molecular profile group. Where there are multiple TCGA, molecular profile groups (double classifier) present within the tumo rs, these are listed in the TCGS MPGs coloumn, while the final classification, based on the TCGA guidelines and PROMISE algorithm is listed in the Final TCGA MPG column [].
TABLE 2
| Case number | Histology | 1st component | 2nd component | |
|---|---|---|---|---|
| Mixed endometrial carcinomas | ||||
| Case M1 | CCC/LGEC | POLE, ARID1A, PTEN, PIK3A, XPO1, IDH1 | ||
| Case M2 | CCC/LGEC | | KRAS, PIK3CA, PTEN, NFE2L2 | |
| Case M3 | CCC/LGEC | PIK3CA | ||
| ARID1A, FGXW7 | ||||
| MSH6, KRAS, ATM, SMAD4 | PTEN, FGFR2, CCND1, ERBB3 | |||
| Case M4 | CCC/LGEC | MSH6, PIK3CA, PPP2R1A, ARID1A | ||
| Case M5 | SC/HGEC | PALB2, PTEN, MAPK | ||
| TP53, RB1 | ATM | |||
| Case M6 | CCC/HGEC | PTEN | ||
| PIK3CA | ||||
| NRAS, ATM | | |||
| Dedifferentiated carcinomas | ||||
| Case D1 | UC/LGEC | KRAS, ARID1A, PTEN, CCND1 | ||
| PIK3R1 | ||||
| | RAD50, MED12 | |||
| Case D2 | UC/LGEC | AKT1, PIK3CA, CTNNB1, MAP2K2 | ||
| PIK3CA, CCND1 | ||||
| Case D3 | UC/LGEC | PIK3CA, KRAS, ARID1A | ||
| Case D4 | UC/LGEC | FGR2, ARID1A, PTEN | ||
| CTNNB1 | | |||
| Case D5 | UC/LGEC | FGFR2, PTEN, PIK3CA | ||
| TP53 | PMS2 | |||
| Case D6 | UC/LGEC | PTEN, ARID1A | ||
| PIK3CA | ||||
| SMARCA4, PIK3R1 | | |||
| Case D7 | UC/CCC | MSH6 | ||
| ATM, PTEN, ARID1A, NF1 | ||||
| ERBB2, APC, KMT2D, PIK3R1, SPOP | PIK3CA, IDH1, PPP2R1A, TP53, FANCI, SMAD4 | |||
| Case D8 | UC/LGEC | POLE, BRCA2, RAD50, ARID1A, NF1, FBXW7 | ||
| SETD2, BRCA2, RAD50, ARID1A | ||||
| BRCA1, NF2, RB1, PIK3R1, CREBBP, ATRX, KIT | PIK3CA, ATM, SLX4 | |||
| Case D9 | UC/HGEC | KRAS, MAX, PTEN | ||
| PIK3CA | ||||
| MYC, CTNNB1, BRAF, ATM, RNF43, ARID1A, CCND1, CDKN2A, CCND2 | | |||
Pathogenic mutations detected in each component in mixed and dedifferentiated endometrial carcinoma.
Abbreviations: CCC, clear cell carcinoma; LGEC, low grade endometrioid carcinoma; SC, serous carcinoma; HGEC, high grade serous carcinoma; UC, undifferentiated carcinoma. Blue: the same pathogenic mutations are present in both components. Orange: both components harboured pathogenic mutations in the same genes, although the specific mutations differed. Grey: mutations present only in one component.
FIGURE 1
FIGURE 2
Regarding the genotype of the tumour components, all MCEs and EDCs, apart from one case (Case M2), had at least one identical pathogenic mutation shared in both components. In Case M2, no pathogenic mutation was detected in the clear cell carcinoma component. This might be due to low tumour cell content or a result of an analytical error. In two MCE and one EDC patients, all components had the same mutational profile with the NGS panel used (Table 2).
Out of the 10 MCEs, 5 showed dMMR phenotype in both components, while there was a single case with dMMR in only one component on IHC. These dMMR cases were predominantly tumours with clear cell carcinoma and endometrioid carcinoma components (5 cases), with only one case (Case M5) comprising serous and endometrioid carcinoma components. Pathogenic POLE mutation was observed in one case, in both clear cell and low-grade endometrioid carcinoma components of the tumour (Case M1). Aberrant p53 expression (diffuse nuclear overexpression) was detected in the serous component of four cases (Cases M5, M7, M8, and M9) (4/10). NGS analyis was performed only on Case M5, where a pathogenic TP53 mutation was confirmed in the serous component.
TCGA molecular classification was only possible on Cases M1-M6, where NGS analysis was performed, as the remaining cases had no POLE mutational data available. Accordingly, both components of one mixed carcinoma (Case M1) were classified as POLE-mutated. Both components of four tumours and one component of one tumour were classified as dMMR The TP53-mutated serous carcinoma component of case 5 also showed MMR deficiency, therefore this tumour was classified as dMMR according to the TCGA guidelines and PROMISE algorithm for double classifier tumours []. There was only one tumour where there was a difference in the TCGA classification of the two components (Case M2): here, the clear cell carcinoma component was classified as NSMP while the low-grade endometrioid carcinoma as dMMR. No mixed carcinoma was classified as TP53-mutated, although no molecular profiling could be performed on Cases M7, M8 and M9, with aberrant p53 expression.
Two of the 5 tumours with dMMR components, where NGS data were available, showed pathogenic mutation in the MMR genes, suggesting that in the other cases loss of protein expression may have occurred due to epigenetic changes, although no methylation analysis for performed to confirm this assumption (Table 2). There were two cases where there was dMMR in both components, but with different patterns. In one case (Case M3), while the low-grade endometrioid carcinoma component showed loss of MLH1 and PMS2 expression, the clear cell carcinoma was MSH6 deficient with a corresponding MSH6 mutation. In another case (Case M6), one component (high-grade endometrial carcinoma) showed loss of both PMS2 and MLH1 expressions, while the second component (clear cell carcinoma) was only PMS2-deficient on IHC.
In the group of EDCs, all cases underwent NGS analysis and could be classified to TCGA groups (Table 2). The most common differentiated histological type was low-grade endometrioid carcinoma (7 out of 9 cases) (Figure 3), followed by high-grade endometrioid, and clear cell carcinoma (1 case each). We identified one case with pathogenic POLE mutation in both components (Case D8), classified as POLE-mutated (1/9). Seven cases had MMR deficiency in both components, classified as dMMR, while one case was dMMR in the undifferentiated component only (Case D2). In this case (Case D2), the low-grade endometrioid component was NSMP. There were two cases with pathogenic TP53 mutations (Cases D5 and D7), although no aberrant p53 expression was seen in Case D7. These cases were also dMMR on immunohistochemistry, therefore classified as dMMR according to the TCGA guidelines and PROMISE algorithm for double classifier tumours []. Out of the 8 cases with dMMR components, only two showed pathogenic MMR gene mutations. Interestingly, in Case D5, while both components showed PMS2 loss on immunohistochemistry, pathogenic PMS2 mutation was detected only in the low-grade endometrioid carcinoma component. The remaining dMMR cases, where pathoge nic MMR mutation was detected, it is presumed that the loss of MMR protein expression was secondary to promoter methylation, although no methylation studies were performed to substantiate this.
FIGURE 3
As an increased frequency of protein expression loss of the members of the SWI/SNF pathway was reported in EDCs [, ], we specifically looked at the frequency of the mutations of those genes in both tumour groups. In MCEs, pathogenic mutation of ARID1A was detected in both components in three cases (3/6). Among the dedifferentiated tumours, six cases had pathogenic mutations of the complex in both components (ARID1A), while in one case there was pathogenic mutation (ARID1A, Case D9) in the undifferentiated component only (7/9). In one case with identical ARID1A mutations in both components, an additional, but different ARID1A mutation was detected in both components (Case D8), while in Case D6, the dedifferentiated component showed an additional pathogenic SMARCA4 mutation.
We also looked at the mutational frequency of genes involved in the PI3K/AKT pathway. All MCEs had either PIK3CA or PTEN mutations, including three cases with clear cell and three cases with serous carcinoma components. (Cases M3, M4, M6 and Cases M1, M5, M6). This was similar in EDCs, where all cases carried mutations in the PI3K/AKT pathway genes, with PIK3CA mutation present in the undifferentiated carcinoma components in 5 cases (Cases D2, D3, D5, D6, and D9) and PTEN mutation in 6 cases (Cases D1, D4, D5, D6, D7, and D9).
In terms of genes involved in homologous recombination pathway (HRR), there was PALB2 mutation detected in both components in one MCE case (Case M5) with ATM mutation in the high-grade endometrioid carcinoma component. In two additional cases, ATM mutation was detected in one (clear cell carcinoma) component (Cases M3, M6). In EDCs, we found BRCA2 mutations in both components (Case D8) with an ATM mutation in the low-grade endometrioid carcinoma component. This case, however, also carried a pathogenic POLE mutation, therefore these HRR gene mutations were likely passenger rather than driver mutations. In Case D7, both components had ATM gene mutation, and in Case D9, there was ATM mutation in the undifferentiated component only.
Discussion
In this study, we examined MCEs and EDCs in order to assess their genetic composition and investigate their potential pathogenesis. MCEs and EDCs were rare in our cohort, representing less than 1% of the examined tumours (0.63% and 0.57%, respectively). Apart from one MCE, where no mutations were detected in the clear cell component, all MCEs or EDCs examined with NGS had at least one common mutation identified in all components. This supports the hypothesis that both tumour groups either arise through divergent differentiation from a common progenitor or through transdifferentiation []. The complete lack of pathogenic mutations in a component of one case was considered more likely to represent an analytical error or could be secondary to low tumour cell content, rather than truly representative of a genetically different phenotype (i.e., collision tumour).
In non-mixed endometrial carcinomas, dMMR tumours were reported to represent about one third of the cases [, ]. In our cohort, there was a considerably higher frequency of dMMR in both MCEs and EDCs (60% and 83%, respectively). Both groups also contained POLE-mutated phenotypes, while TP53-mutated and NSMP tumours were rare. High frequency of dMMR has been reported in EDC [, , , ], while there are reports of frequent dMMR in mixed clear cell/endometrioid carcinomas, but not in mixed serous/endometrioid carcinomas [, ]. Our results in MCEs may therefore be biased by the high proportion of clear cell/endometrioid tumours in our cohort (6 cases), out of which 5 were dMMR in at least one component, while the sixth case was POLE-mutated one. There was a relatively low number of serous/endometrial carcinomas in our set (3 cases), out of which 1 was dMMR. The remaining case, a mixed serous/clear cell carcinoma did not show MMR deficiency or POLE mutation. Nevertheless, the high proportion of tumours with DNA repair deficiency (mainly dMMR but also with POLE mutation) suggest that genetic heterogeneity is likely driven by genomic hypermutability in both MEC (at least those with clear cell carcinoma component) and EDC.
A recent Japanese study suggested a potential role of HRR gene mutation in mixed endometrial tumours, with a frequency of 30%, and a potential therapeutic role in PARP inhibitor therapy []. The high proportion of MMR (and POLE) mutation and genomic hypermutability among both MECs and EDCs, however, may introduce potential difficulties in any further analysis of the genotype of these tumours. As these tumours are characterised by high tumour mutational burden due to an erroneous DNA repair mechanism, it is difficult to assess the significance of a pathogenic mutation with regards to their driver or passenger status in tumour pathogenesis. This was also taken into consideration in the TCGA molecular profiling, where POLE-mutated and dMMR phenotypes take precedence over TP53 mutation in endometrial carcinomas. Nevertheless, we found a similar frequency (30%) of HRR gene mutation in our cohort. There were HRR gene mutations detected in 2 (out of 6) MCEs and 3 (out of 9) EDCs - one of the latter being also POLE-mutated. The clinical and therapeutic significance of these findings however, are uncertain, as to the best of our knowledge there are no reports of high incidence of HHR deficiency/high genomic instability score in endometrial carcinoma. Furthermore, trials of additional PARP inhibitor therapy in endometrial carcinomas appear more empirical rather than based on the tumours’ HRR status [].
Loss of ARID1A (part of the SWI/SNF complex) expression is considered a poor prognostic factor in endometrial carcinoma, and mutation of other genes of the SWI/SNF complex have been reported at a higher rate in dedifferentiated and undifferentiated endometrial carcinomas [, ]. The SWI/SNF complex plays a role in chromatin remodelling, and disruption of this pathway affects homeostasis and cellular development []. Alterations of this pathway is associated with a poorer prognosis in endometrial carcinoma, with rapid disease progression, and therapy resistance even when EDCs are considered separately [, ]. Mutation of the SWI/SNF complex (predominantly ARID1A) occurs in up to 40% in non-mixed endometrioid carcinoma, and 25% in clear cell carcinoma of the uterus []. However, this genetic change is rare in serous endometrial carcinoma, with mutations in both ARID1A and TP53 considered mutually exclusive []. This reported high frequency of SWI/SNF mutations in both undifferentiated and dedifferentiated endometrial carcinomas may suggest that in EDCs, it is the undifferentiated component that carries the mutation. In line with previous reports, there was a high frequency of mutation (78%, 7 out 9) of the SWI/SNF complex in EDCs (mainly ARID1A mutation, with SMARCA4 mutation in one case) in our cohort. However, out of these 7 cases, ARID1A mutation was present not only in the undifferentiated, but also in the other, differentiated carcinomatous component. This suggests that ARID1A mutation is a relatively early event in tumorigenesis, occurring before the transdifferentiation or divergent differentiation of the tumours, rather than a new event in the undifferentiated component. In MCEs, only ARID1A mutation was detected, but again, it was present in both components of 50% in the MCEs (3 out of 6, all mixed clear cell/endometrioid).
Similar to the SWI/SNF pathway, mutations and disruption of the PI3K/PTEN pathway, mainly PTEN and PIK3CA mutations, are predominantly involved in the pathogenesis of endometrioid endometrial carcinomas (up to 50%), with frequent co-occurence of PIK3CA and PTEN mutations []. Activation of this pathway affects gene expressions that regulate apoptosis and promote cell proliferation. In pure clear cell or serous carcinoma, PIK3CA mutation is less frequent (<30%), while PTEN mutation is exceedingly rare (0%–15%) [, ]. In our cohort, all MCEs and EDCs showed mutation in PIK3CA and/or PTEN genes in at least one component. In addition, in contrast to pure serous and clear carcinomas, it was frequent in serous and clear cell carcinomatous components of MCEs (4/6). This again suggests that the pathogenesis of these tumours differs from pure serous and clear cell carcinomas, supporting the hypothesis of transdifferentiation or divergent differentiation from an endometrioid carcinomatous component, with PIK3CA and PTEN mutation being an early event in carcinogenesis.
MCEs and EDCs are rare, and this is the cause of our relatively low case number, which is the main limitation of our study. Our results, however still suggest that the pathomechanisms of MCEs and EDCs are very similar in many aspects: they both appear to arise predominantly on a background of dMMR and genomic hypermutability, by means of transdifferentiation or divergent differentiation. Nevertheless EDCs show a higher frequency of SWI/SNF mutations and according to previous studies, still carry a poorer prognosis. [, , ]. Although our study lacks clinical follow-up data, due to this disparity in involvement of the SWI/SNF complex and the reported difference in prognosis observed in EDC and MCE, it may be worthwhile to maintain a separate WHO classification for endometrial carcinomas, despite their shared pathogenesis and clonal origin [].
Statements
Data availability statement
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors.
Ethics statement
The studies involving humans were approved by Medical Research Council - Scientific and Research Ethics Committee, Health and Scientific Committee of the Ministry of Human Resources of Hungary (approval number: BM-794-1-2024). The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required from the participants or the participants’ legal guardians/next of kin in accordance with the national legislation and institutional requirements.
Author contributions
JB: case review, writing – original draft, research and data collection; IV: conceptualization, writing – review and editing LS: research and data collection; AS: research and data collection; ET: resources, supervision; LK: supervision, writing – review and editing; ZM: conceptualization, writing – original draft. All authors contributed to the article and approved the submitted version.
Funding
The author(s) declared that financial support was received for this work and/or its publication. This research was funded by HUN-REN-ONKOL-TTK-HCEMM Oncogenomics Research Group.
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Supplementary material
The Supplementary Material for this article can be found online at: https://www.por-journal.com/articles/10.3389/pore.2026.1612616/full#supplementary-material
SUPPLEMENTARY TABLE S1Antibody clones and dilutions used for immunohistochemistry.
SUPPLEMENTARY TABLE S2Pathogenic point mutations detected in each component of mixed endometrial carcinomas and endometrial dedifferentiated carcinomas. Abbreviations: clear cell carcinoma (CCC), low grade endometrioid carcinoma (LGEC), SC (serous carcinoma), high grade serous carcinoma (HGEC), undifferentiated carcinoma (UC).
References
1.
B Board WC of TE. Female Genital Tumours. Available online at: https://publications.iarc.who.int/Book-And-Report-Series/Who-Classification-Of-Tumours/Female-Genital-Tumours-2020 (Accessed July 28, 2026).
2.
PappaCLe ThanhVSmythSLZouridisAKashifASadeghiNet alMixed endometrial epithelial carcinoma: epidemiology, treatment and survival rates—A 10-Year retrospective cohort study from a single institution. J Clin Med (2023) 12(19):6373. 10.3390/jcm12196373
3.
ConcinNMatias-GuiuXCibulaDColomboNCreutzbergCLLedermannJet alESGO–ESTRO–ESP guidelines for the management of patients with endometrial carcinoma: update 2025. Lancet Oncol (2025) 26(8):e423–35. 10.1016/S1470-2045(25)00167-6
4.
LiWLiLWuMLangJBiY. The prognosis of stage IA mixed endometrial carcinoma. Am J Clin Pathol (2019) 152(5):616–24. 10.1093/ajcp/aqz083
5.
RabbanJTGilksCBMalpicaAMatias-GuiuXMittalKMutterGLet alIssues in the differential diagnosis of uterine low-grade endometrioid carcinoma, including mixed endometrial carcinomas: recommendations from the international society of gynecological pathologists. Int J Gynecol Pathol Off J Int Soc Gynecol Pathol (2019) 38(Suppl. 1):S25–39. 10.1097/PGP.0000000000000512
6.
ArciuoloDTravaglinoARaffoneARaimondoDSantoroARussoDet alTCGA molecular prognostic groups of endometrial carcinoma: current knowledge and future perspectives. Int J Mol Sci (2022) 23(19):11684. 10.3390/ijms231911684
7.
KöbelMMengBHoangLNAlmadaniNLiXSoslowRAet alMolecular analysis of mixed endometrial carcinomas shows clonality in most cases. Am J Surg Pathol (2016) 40(2):166–80. 10.1097/PAS.0000000000000536
8.
MatraiCMotanaghSMirabelliSMaLHeBChapman-DavisEet alMolecular profiles of mixed endometrial carcinoma. Am J Surg Pathol (2020) 44(8):1104–11. 10.1097/PAS.0000000000001519
9.
GanesanRMillsABuzaN. Template for reporting results of biomarker testing of specimens from patients with carcinoma of gynecologic origin. Available online at: https://www.cap.org/wp-content/uploads/protocols/Gynecologic.Bmk_1.3.0.0.REL_CAPCP.pdf?download=true (Accessed September 30, 2026).
10.
León‐CastilloAGilvazquezENoutRSmitVTMcAlpineJNMcConechyMet alClinicopathological and molecular characterisation of ‘multiple‐classifier’ endometrial carcinomas. J Pathol (2020) 250(3):312–22. 10.1002/path.5373
11.
QiYYaoQXuYYuLTuXChengYet alDedifferentiated/undifferentiated endometrial carcinoma:SWI/SNF-deficiency and molecular heterogeneity indicating novel prognostic stratification in a single-center study of 71 cases. Hum Pathol (2026) 175:106155. 10.1016/j.humpath.2026.106155
12.
GiordanoGFerioliEGuareschiDTafuniA. Dedifferentiated endometrial carcinoma: a rare aggressive neoplasm-clinical, morphological and immunohistochemical features. Cancers (2023) 5(21):5155. 10.3390/cancers15215155
13.
DagherCSelenicaPMomeni-BoroujeniAEllensonLHMuellerJJAbu-RustumNRet alMolecular subtypes and genomic landscape of undifferentiated and dedifferentiated endometrial cancer. Int J Gynecol Cancer Off J Int Gynecol Cancer Soc (2025) 35(5):101815. 10.1016/j.ijgc.2025.101815
14.
TravaglinoARaffoneAGencarelliASaracinelliSRiccardiCMolloAet alClinico-pathological features associated with mismatch repair deficiency in endometrial undifferentiated/dedifferentiated carcinoma: a systematic review and meta-analysis. Gynecol Oncol (2021) 160(2):579–85. 10.1016/j.ygyno.2020.11.015
15.
CoenegrachtsLGarcia-DiosDADepreeuwJSantacanaMGatiusSZikanMet alMutation profile and clinical outcome of mixed endometrioid-serous endometrial carcinomas are different from that of pure endometrioid or serous carcinomas. Virchows Arch Int J Pathol (2015) 466(4):415–22. 10.1007/s00428-015-1728-5
16.
HaragaHNakayamaKRaziaSIshikawaMYamashitaHKannoKet alExploring the genetic and clinical landscape of dedifferentiated endometrioid carcinoma. Int J Mol Sci (2025) 26(9):4137. 10.3390/ijms26094137
17.
WestinSNMooreKChonHSLeeJYThomesPJSundborgMet alDurvalumab plus carboplatin/paclitaxel followed by maintenance durvalumab with or without olaparib as first-line treatment for advanced endometrial cancer: the phase III DUO-E trial. J Clin Oncol (2024) 42(3):283–99. 10.1200/JCO.23.02132
18.
ToumpekiCLiberisATsirkasITsirkaTKalagasidouSInagamovaLet alThe role of ARID1A in endometrial cancer and the molecular pathways associated with pathogenesis and cancer progression. Vivo (2019) 33(3):659–67. 10.21873/invivo.11524
19.
SachdevaMBlanc-DurandFTanD. Controversies in the management of clear cell carcinoma of the uterus and ovary. Int J Gynecol Cancer Off J Int Gynecol Cancer Soc (2025) 35(3):101681. 10.1016/j.ijgc.2025.101681
20.
CatasusLGallardoACuatrecasasMPratJ. PIK3CA mutations in the kinase domain (exon 20) of uterine endometrial adenocarcinomas are associated with adverse prognostic parameters. Mod Pathol Off J U S Can Acad Pathol Inc (2008) 21(2):131–9. 10.1038/modpathol.3800992
21.
NeroCCiccaroneFPietragallaAScambiaG. PTEN and gynecological cancers. Cancers (2019) 11(10):1458. 10.3390/cancers11101458
22.
HamiltonSNTinkerAVKwonJLimPKongISihraSet alTreatment and outcomes in undifferentiated and dedifferentiated endometrial carcinoma. J Gynecol Oncol (2022) 33(3):e25. 10.3802/jgo.2022.33.e25
23.
GohCFarahBLHoWYWongSLGohCHRChewSHet alDedifferentiated endometrioid adenocarcinoma of the uterus: a case series and review of literature. Gynecol Oncol Rep (2020) 32:100538. 10.1016/j.gore.2020.100538
Summary
Keywords
dedifferentiated endometrial carcinoma, divergent differentiation, mismatch repair deficiency, mixed endometrial carcinoma, next generation sequencing
Citation
Betenbuk J, Vereczkey I, Szalai L, Simon A, Tóth E, Kuthi L and Melegh Z (2026) Molecular characterisation of mixed and dedifferentiated endometrial carcinomas reveals shared pathogenesis and clonal origin. Pathol. Oncol. Res. 32:1612616. doi: 10.3389/pore.2026.1612616
Received
04 September 2026
Revised
03 October 2026
Accepted
27 October 2026
Published
19 November 2026
Volume
32 - 2026
Edited by
Anita Árokszállási, University of Debrecen, Hungary
Updates
Copyright
© 2026 Betenbuk, Vereczkey, Szalai, Simon, Tóth, Kuthi and Melegh.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Judit Betenbuk, betenbuk.judit@gmail.com; Levente Kuthi, kuthi.levente@oncol.hu
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.