Clinicopathological significance of immunohistochemical expression of p53 and PTEN in endometrial carcinoma

1. Introduction

Endometrial carcinoma is a heterogeneous group of neoplasms whose initiation and progression are driven by the accumulation of genetic and epigenetic alterations that affect signalling pathways, proto-oncogene activation, and inactivation of tumour suppressor genes [1, 2, 3, 4]. Although surgical, radiation, and pharmacological treatments are well-established, the identification and validation of biomarkers remain key to improving our understanding of molecular pathogenesis, enabling more accurate prognostic assessment, and facilitating the development of molecularly targeted therapies, especially in cases of recurrent disease and unfavourable clinical outcomes [1, 2, 3, 4]. Over the past decade, molecular classification has become the gold standard for stratifying EC, offering a biologically meaningful framework beyond traditional histopathologic parameters. The Cancer Genome Atlas (TCGA) project identified four molecular subtypes with distinct prognostic implications: DNA Polymerase Epsilon (POLE)-mutated, Mismatch Repair (MMR)-deficient, no specific molecular profile, and p53-abnormal [5]. These groups differ markedly in molecular alterations and clinical outcomes, with p53-abnormal tumours showing the poorest prognosis and frequent serous-like morphology. The Proactive Molecular Risk Classifier for Endometrial Cancer (ProMisE) provides a clinically applicable surrogate of the TCGA taxonomy using immunohistochemistry for p53 and MMR proteins, combined with targeted POLE sequencing [6]. Within this framework, aberrant expression of p53 and PTEN reflects critical oncogenic events and remains central to prognostic stratification in EC [7].

The proapoptotic gene p53 is a tumour suppressor gene and expresses a nuclear 53-kDa phosphoprotein called p53. The p53 protein is a transcription factor that induces cell cycle arrest at the G1 checkpoint, promotes DNA repair when damaged, and initiates apoptosis when this repair fails [8, 9]. Dysfunction of p53 promotes uncontrolled cellular proliferation and tumorigenesis. Immunohistochemically, p53 expression stratifies EC into a wild-type and a mutant staining pattern (p53abnormal EC) (p53abn). Although p53abn accounts for about 15% of all ECs, this subgroup is associated with a disproportionately high mortality rate (50–70%) and poor clinical outcomes, particularly in poorly differentiated (G3) and serous carcinomas [10]. Increasing evidence suggests that the addition of systemic chemotherapy to radiotherapy improves oncological outcomes in p53-abnormal EC. Furthermore, targeted therapeutic strategies are currently being explored [10]. PTEN (phosphatase and tensin homolog) is a tumour suppressor gene that suppresses cell proliferation and differentiation and participates in the insulin signalling pathway. This gene plays a key role in the phosphatidylinositol 3-kinase/protein kinase B (PI3K-AKT) pathway by dephosphorylating phosphatidylinositol (3,4,5)-trisphosphate (PIP3) at the cell membrane [11, 12, 13, 14]. Under normal physiological conditions, PTEN exerts inhibitory effects on intracellular signalling pathways, cellular proliferation, cell migration, and cell adhesion, while also promoting apoptosis in damaged cells [15, 16]. The absence of functional PTEN protein leads to unregulated activation of the PI3K signalling pathway [17]. One of the main factors in the pathogenesis of many tumours, particularly ECs, is the loss of PTEN function through somatic mutations, transcriptional/post-transcriptional mechanisms, and microRNA dysregulation [18].

Despite the well-established individual roles of p53 and PTEN, their combined IHC assessment and its association with key clinicopathological characteristics in surgically staged cohorts remain insufficiently explored in the literature. This study investigates the associations between p53 and PTEN expression, both individually and in combination, and clinicopathological parameters in EC. We hypothesise that p53-abnormal expression and low or negative PTEN expression are associated with adverse clinicopathological features.

2. Materials and methods

3. Results

The analysis included 78 consecutive patients with histologically verified EC (Table 1).

Lymph node assessment was performed in 74 patients (94.8%). Pelvic lymphadenectomy was carried out in 48 patients, including 2 cases in which para-aortic lymphadenectomy was additionally performed. Selective pelvic lymphadenectomy was performed in 23 patients, and only 3 patients underwent pelvic lymph node biopsy. Four patients did not undergo lymph node dissection. Regional lymph node metastases were identified in 12 patients (16%): 10 had pelvic nodal metastases and 2 had combined pelvic and para-aortic metastases.

Histologic grade (G1–G3) was not assigned in 7 patients (5 serous carcinomas, 1 pure clear cell carcinoma, and 1 carcinosarcoma), as these histotypes are generally regarded as high-grade by definition. The second clear cell carcinoma was of mixed type with a defined endometrioid component, and the reported grade (G3) referred to the epithelial (endometrioid) component.

One carcinosarcoma case was included in the cohort. The patient was 67 years old at the time of surgery and had FIGO stage IB disease. She underwent total hysterectomy with pelvic lymphadenectomy (19 lymph nodes were removed), followed by adjuvant chemoradiotherapy, and was alive at the end of follow-up. Immunohistochemically, the tumor demonstrated a p53 wild-type expression pattern and high PTEN expression. As carcinosarcoma is considered a high-grade tumor by definition, no conventional histologic grading (G1–G3) was assigned. Accordingly, this case was included in the high-grade (Type II) category within the dualistic classification used in the study. Given that only a single carcinosarcoma case was included in the cohort, no separate statistical analysis was performed for this subtype.

Of all the analysed cases, 58 were classified as Type I EC (G1–G2 endometrioid carcinomas), whereas 20 were classified as Type II EC, including serous, clear cell, carcinosarcoma, and high-grade (G3) endometrioid carcinomas. The surgical interventions were performed in 2019 (n = 35; 44.9%) and 2020 (n = 43; 55.1%). Throughout the follow-up (until 31 December 2024), 17 (21.8%) patients died from the disease, 14 (17.9%) died from other causes, and 47 (60.3%) survived with no evidence of recurrence.

Immunohistochemical assessment of p53 and PTEN was performed retrospectively for the study and did not influence therapeutic decisions.

In terms of IHC expression categories, 62 (79.5%) tumours were classified as p53 wild-type (p53wt), and 16 (20.5%) showed aberrant p53 expression (p53mut). Based on the IRS, 30 (38.5%) cases showed negative expression for PTEN, 17 (21.8%) low expression, 20 (25.6%) moderate expression, and 11 (14.1%) high expression. For analytical purposes, cases were grouped into two categories: PTEN negative + low (n = 47; 60.3%) versus PTEN moderate + high (n = 31; 39.7%) (Table 1). Representative histologic and immunohistochemical findings are illustrated in Fig. 1.

Fig. 1.Representative histologic and immunohistochemical findings in endometrial carcinoma. (A) Hematoxylin-eosin staining showing endometrioid carcinoma G2 ×10. (B) Aberrant p53 expression (null pattern) G2 ×10. (C) Loss of PTEN expression in tumor cells G2 (score 0) ×10. (D) Strong PTEN expression showing intense cytoplasmic staining in tumor cells G2 (score 12) ×10. Panels A–C are from the same tumour, whereas panel D represents a different tumour. All tumours correspond to endometrioid carcinoma, grade 2.

A comparison of the clinical and pathological characteristics between tumors with p53 wild-type expression (p53wt) and aberrant p53 expression (p53mut) revealed a more adverse clinicopathological profile in the p53mut group: the frequency of LVSI was higher (43.7% vs. 17.7%; p = 0.028), deep myometrial invasion (>50%) was more common (75.0% vs. 46.8%; p = 0.044), and the proportion of poorly differentiated tumors (G3) was significantly higher in the p53mut group (46.2% vs. 12.07%) both in the three-grade comparison (G1 vs. G2 vs. G3; p = 0.014) and in the binary comparison (G1 + G2 vs. G3; p = 0.004) (Table 2).

Among low-grade tumours (G1–2), seven cases (9.0% of the total cohort) demonstrated aberrant p53 expression. The median age in this subgroup was 66 years (range 61–83). The disease stage was heterogeneous, including early-stage disease (IA–IB) in five patients, locally advanced disease (IIIB) in one patient, and advanced metastatic disease (IVB) in one patient. LVSI was present in 3 of 7 cases. Regarding adjuvant treatment, two patients received combined chemoradiotherapy, three received external beam radiotherapy alone, and two received brachytherapy. One patient died one month after surgery due to an intercurrent disease, before an exhaustive oncologic assessment could be performed. During follow-up, two patients died of endometrial carcinoma at 30 and 55 months, respectively; two patients died of causes unrelated to the malignancy (at 45 months and one month postoperatively, respectively), and three patients remain alive without evidence of disease. Despite their favourable histologic grade, this subgroup demonstrated heterogeneous but overall, clinically aggressive behaviour, including advanced stage at diagnosis, frequent LVSI, and disease-related mortality.

The assessment of PTEN in relation to LVSI showed that the PTEN negative + low group was associated with a higher frequency of LVSI compared with the PTEN moderate + high group (31.9% vs. 9.7%). This combined PTEN negative + low expression group was significantly associated only with LVSI (p = 0.023), whereas no statistically significant associations were observed with histologic grade, myometrial invasion, or histologic subtype (Type I vs. Type II) (Table 2). The combined analysis of p53 and PTEN demonstrated statistically significant differences in grade distribution between the molecular subgroups (Table 3). A higher proportion of G3 tumors was observed in the p53 abnormal + PTEN negative/low group compared to the p53 wild-type + PTEN moderate/high group (37.5% vs. 4.2%; p = 0.038 and p = 0.011). However, given that p53 status alone was significantly associated with grade (Table 2), whereas PTEN expression was not, this enrichment of G3 cases is most likely attributable to p53 abnormality rather than to an independent effect of PTEN expression.

Of the 24 cases with p53wt + PTEN moderate/high, only one had a G3 tumour. The frequency of LVSI was also higher in the p53mut + PTEN negative/low group compared to the p53wt + PTEN moderate/high group (55.6% vs. 4.2%; p = 0.015) (Table 3). The survival rate analysis supports the prognostic relevance of p53: OS is lower in the p53mut group compared with the p53wt group (log-rank χ² = 6.706; p = 0.01), with a median OS of 55.23 months (95% confidence interval (CI): 49.236–61.223) for p53wt versus 34.30 months (95% CI: 20.396–48.204) for p53mut (Table 4; Fig. 2).

Fig. 2.Overall survival of patients with endometrial carcinoma according to p53 status (p53wt vs. p53mut). Kaplan-Meier analysis demonstrated significantly reduced OS in the p53mut group compared with the p53wt group (log-rank χ² = 6.706, p = 0.01). Censored observations are indicated by tick marks. The label “p53mut” used in the figure corresponds to aberrant p53 expression detected by immunohistochemistry (p53 abnormal pattern). Detailed survival estimates are presented in Table 4. p53wt: p53 Wild-type; p53mut: p53 Mutant.

The analysis of DSS showed that of the 17 patients who died from the disease, 9 (52.94%) had a p53 abnormality. The DSS was lower in the p53mut group compared with the p53wt group (log-rank χ² = 16.841; p = 0.001), with a mean DSS of 63.715 months (95% CI: 58.926–68.504) for p53wt versus 38.040 months (95% CI: 23.328–52.752) for p53mut (Table 4; Fig. 3).

Fig. 3.Disease-specific survival of patients with endometrial carcinoma according to p53 status (p53wt vs. p53mut). Kaplan-Meier analysis demonstrated significantly reduced DSS in the p53mut group compared with the p53wt group (log-rank χ² = 16.841, p = 0.001). The label “p53mut” used in the figure corresponds to aberrant p53 expression detected by immunohistochemistry (p53 abnormal pattern). Detailed survival estimates are presented in Table 4. p53wt: p53 Wild-type; p53mut: p53 Mutant.

A significant difference was also found in one-, three-, and five-year OS and DSS (Gehan-Breslow-Wilcoxon = 7.695; p = 0.006 and 13.451; p = 0.001, respectively), with lower values observed in the p53mut group (50%, 38%, and 29% for OS; 54%, 48% and 37% for DSS, respectively) (Table 5).

Univariate and multivariate Cox proportional hazards regression analyses were performed to evaluate prognostic factors for OS and DSS, including p53 status and relevant clinicopathologic covariates (Tables 6,7). In the multivariate model for OS (Table 6), Type II EC (Hazard Ratio (HR) 2.544, p = 0.021), regional lymph node metastases (HR 3.697, p = 0.001), and p53 abnormal expression (IHC) (HR 2.703, p = 0.019) remained independent prognostic factors. In the multivariate model for DSS (Table 7), age >65 years (HR 10.975, p = 0.029), G3 histology (HR 6.757, p = 0.027), lymph node metastases (HR 8.843, p = 0.003), and p53 abnormal expression (IHC) (HR 4.651, p = 0.047) were independently associated with poorer DSS.

We included the subgroup of low-grade (G1–2) tumours with aberrant p53 expression in the multivariate Cox regression model, and aberrant p53 status remained independently associated with survival after adjustment for stage, grade, LVSI, lymph node status, and adjuvant treatment.

4. Discussion

In the present study, we investigated the relationship between clinicopathological characteristics and IHC expression of p53 and PTEN in a surgically staged cohort of patients with EC (2019–2020). The proportion of tumours showing aberrant p53 expression (p53mut) was 20.5%, which is consistent with previously reported frequencies of approximately 15–20% in EC [10, 24]. In our study, p53 abnormality was associated with LVSI, deep myometrial invasion, and a higher histologic grade, while PTEN negative/low expression was significantly associated only with LVSI. Although co-expression analysis revealed differences in grade distribution among subgroups, these findings appear to be largely attributable to p53 status, as PTEN alone showed no significant association with grade.

Our cohort-specific distribution of p53mut tumours by histotype and grade (serous: 3/5, 60%; G3: 5/13, 38.5%; G1–G2: 7/58, 12.07%; clear-cell: 1/2, 50%) partially differs from aggregated molecularly classified cohort studies, which report approximate percentages: serous ~93%, carcinosarcoma ~85%, clear cell ~38%, G3 endometrioid ~22%, G1–2 endometrioid ~5% [7, 10, 25, 26, 27]. Such a distribution probably reflects differences in cohort composition, subgroup sizes, and methodologies, although our overall frequency remains within the ranges reported in the literature [10, 24]. It is noteworthy that certain endometrial carcinomas may harbour TP53 mutations while retaining a normal, wild-type p53 immunostaining pattern. These non-disruptive mutations result in functionally abnormal, yet immunohistochemically undetectable, protein alterations. Because our study relied solely on IHC evaluation, such cases may have been misclassified as p53 wild-type despite an underlying TP53 mutation. This potential discrepancy could contribute to the biological and prognostic heterogeneity found within the p53-wild-type group. Nevertheless, immunohistochemistry remains a practical and widely accepted surrogate marker of TP53 mutation status in both diagnostic and research practice.

The presence of aberrant p53 expression in a subset of morphologically low-grade (G1–2) tumours represents a biologically discordant yet clinically highly relevant phenotype. In our cohort, these tumours demonstrated a heterogeneous stage distribution, frequent LVSI, and disease-related mortality, underscoring the fact that favourable histologic grade did not necessarily translate into indolent clinical behaviour. Disease-related deaths occurred despite the use of intensified adjuvant treatment, including combined chemoradiotherapy, highlighting the dominant role of tumour biology over conventional histopathologic risk factors. These findings support the concept that p53 status may override histologic grade in defining tumour aggressiveness and should be incorporated into risk stratification even in tumours classified as low-grade by conventional morphology. The adverse prognostic impact of aberrant p53 expression persisted in multivariate analysis even when low-grade tumours were included, further supporting the concept that molecular alterations may supersede histologic grade in defining tumour aggressiveness. In the endometrioid-only subgroup (n = 70; 12 disease-specific events), multivariable survival modelling would have been underpowered and prone to overfitting. Larger cohort studies are needed to evaluate histotype-restricted prognostic effects.

Regarding PTEN, our data (IRS: negative 38.5%, low: 21.8%, moderate: 25.6%, high: 14.1%; analytically combined as negative + low 60.3% versus moderate + high 39.7%) showed an association of negative/low expression with LVSI (31.9% versus 9.7%; p = 0.023) (Table 2). For cases with a negative score, PTEN alteration or functional loss is more likely, whereas cases with a high score are more likely to retain PTEN function. The remaining cases with low and moderate scores form a heterogeneous group: some cells harbour mutations to varying degrees, while others do not. For this reason, we grouped the PTEN scores into two pairs: negative + low and moderate + high. The literature data are heterogeneous, with some studies reporting no significant association between PTEN expression and standard prognostic factors [16]. Limitations of immunohistochemical assessment may partly explain this variability, as recent evidence suggests that certain dominant-negative PTEN alterations may retain protein expression despite functional inactivation, potentially masking associations with clinicopathological outcomes [28]. However, Stavropoulos et al. [29] reported a negative correlation between the percentage of PTEN-positive cells and the depth of myometrial invasion. Similarly, Daniilidou et al. [30] found that PTEN negativity correlates with G3. In contrast, in our series, 9/13 G3 cases were PTEN-negative and low, versus 4/13 moderate and high, without statistical significance. We also observed an age-related association: of the 9 patients with a p53mut + PTEN-negative/low profile, 8 were over 60. However, given the small number of cases, this observation should be interpreted cautiously and requires confirmation in larger cohorts.

Co-expression analysis reinforces the stratification value of the biomarkers studied. The combination of p53mut + PTEN negative/low is associated with higher frequencies of LVSI and G3. At the same time, we did not find any significance for the MI (Table 3). Similar observations of adverse outcomes with simultaneous PTEN and p53 pathway abnormalities in EC have been reported [31]. In addition, Terakawa et al. [32] reported that PTEN overexpression is a favourable prognostic marker for OS in advanced EC treated with adjuvant chemotherapy, possibly due to increased chemosensitivity. Based on our findings, we support the view that concomitant assessment of p53 and PTEN expression may contribute to the characterisation of tumour behaviour in EC [33, 34, 35, 36, 37, 38].

Our findings are consistent with the molecular framework defined by the TCGA and ProMisE classifiers, which recognise p53-abnormal tumours as a subgroup associated with aggressive morphology and poor prognosis [5, 6]. Large molecularly classified cohort studies have reported TP53 mutation rates approaching 90–95% in endometrial serous carcinoma. However, pooled analyses of whole-exome sequencing studies have demonstrated variability, with an average TP53 mutation rate of approximately 75% (range 60–88%) [39]. In addition, several immunohistochemical series have reported rates of aberrant p53 expression below 80%, indicating that variability exists across studies and may reflect differences in methodology, case mix, and the use of IHC versus molecular confirmation. In our cohort, p53 abnormal (IHC) expression was observed in 3 of 5 serous carcinoma cases (60%). Given the small number of serous cases included and the absence of molecular confirmation, this finding should be interpreted cautiously.

Similar to previous reports, our study confirmed the independent prognostic impact of aberrant p53 expression. As the role of p53 is already well established as a surrogate marker of the copy-number–high molecular subtype, attention has shifted to additional biomarkers. In this context, PTEN emerges as a potentially valuable complementary biomarker alongside p53. In our cohort, loss of PTEN expression was associated with adverse pathological features such as LVSI, and this loss was more frequently observed in high histologic grade tumours, suggesting a link between PTEN inactivation and more aggressive tumour morphology. This observation aligns with evidence that PTEN alterations represent one of the most frequent and earliest molecular events in endometrioid endometrial carcinogenesis and can be reliably identified by immunohistochemistry, which has been shown to perform robustly for assessing PTEN status [13, 40]. Furthermore, studies evaluating the combined expression of p53 and PTEN have shown that their concomitant immunohistochemical profiles correlate with histologic differentiation and other clinicopathologic features, supporting the notion that PTEN may further refine risk assessment when interpreted together with p53 status [29].

The clinical implications of our results are twofold. Among the investigated biomarkers, p53 stands out as an easily accessible prognostic marker. If p53 is not tested, patients with apparently low-risk (FIGO stage IA and IB) tumours and p53 abnormalities may not receive adequate adjuvant treatment and follow-up. Even with limited resources, routine IHC assessment can be useful in preoperative planning (by identifying the expected risk for LVSI and G3 tumours) and justify a more intensive adjuvant strategy in borderline cases. PTEN, both as an indicator of LVSI and, when assessed in combination with p53, as a marker of an enhanced risk profile, adds valuable information to standard morphology. This approach is particularly relevant given the limited implementation of molecular classification in Bulgaria. Testing for POLE/DNA Polymerase Delta 1 (POLD1) mutational status by Polymerase Chain Reaction (PCR) or POLE NGS, along with systematic immunohistochemical evaluation of mismatch repair (MMR) proteins and p53, is not yet available at all centres. Even a partial panel assessment (MMR + p53) would improve risk stratification, and p53 is key, given the disproportionately high risk in the p53 abnormal subgroup. Although the traditional histopathologic grouping was retained for analytic consistency, the molecular classification proposed by TCGA provides a more accurate biological framework and supersedes the former type I/II dichotomy.

Taken together, our findings reinforce the central prognostic role of aberrant p53 expression in endometrial carcinoma and support its value as a robust surrogate marker of aggressive tumour biology. Our data also suggest that PTEN loss, although not independently prognostic, is associated with adverse pathologic features and may provide additional biological stratification within the heterogeneous p53-wild-type subgroup. Despite the heterogeneity in adjuvant treatment modalities across our cohort, aberrant p53 expression remained an independent prognostic factor for survival in multivariate analysis, underscoring its prognostic relevance beyond treatment-related effects. In line with contemporary TCGA-based and Post Operative Radiation Therapy in Endometrial Carcinoma (PORTEC)-based integrated risk models, these findings support the concept that combining molecular markers with established clinicopathologic parameters yields a more refined, clinically meaningful risk stratification in EC.

The strengths of the study are the uniform surgically staged cohort, the standardised IHC protocol, and the long follow-up period (until 31 December 2024). Limitations include the retrospective design, the moderate subgroup sizes (especially the p53mut group), the use of IHC for assessment, and the lack of information on IHC expression of p53 and PTEN during patient treatment.

5. Conclusions

Aberrant p53 expression was significantly associated with adverse clinicopathological features and poorer survival outcomes. PTEN loss (negative/low IHC expression) was correlated with LVSI but did not demonstrate independent prognostic value in multivariate analysis. Although combined expression patterns revealed differences in risk distribution, these findings appeared to be largely attributable to p53 status. Further prospective studies are needed to clarify the potential clinical relevance of PTEN in endometrial carcinoma.

Availability of data and materials

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Author contributions

CT—conceptualisation, methodology, formal analysis and investigation, data collection, writing–original draft preparation, writing–review and editing, validation. SP—conceptualisation, methodology, formal analysis and investigation, writing–review and editing, validation. KP—methodology, formal analysis and investigation, data collection, validation. All authors contributed to critical editorial changes in the manuscript. All authors read and approved the final manuscript.

Ethics approval and consent to participate

This study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of Medical University–Pleven, Bulgaria (approval No. 733-KENID, dated 05 June 2023). Because this was a retrospective study, the requirement for individual informed consent was waived by the ethics committee, in accordance with institutional and ethical guidelines.

Acknowledgment

We thank Yanka Tzvetanova for English language editing and Mircho Vukov and Gena Grancharova for assistance with statistical analysis.

Funding

The support of the project BG16RFPR002-1.014-0002-С001, Centre of Competence in Personalised Medicine, 3D and Telemedicine, Robot-assisted and minimally Invasive Surgery, funded by the PRIDST 2021-2027, co-funded by the EU, is greatly acknowledged. Project No. 20/2023 was implemented with financial resources from the Medical University–Pleven, Bulgaria.

Conflict of interest

The authors declare no conflict of interest.