Introduction
The role of endometrial cancer (EC) classification has evolved over decades, from simple systems based solely on anatomical criteria to more complex models incorporating molecular and pathological features. The 2023 update of the International Federation of Gynecology and Obstetrics (FIGO) classification introduced key changes aimed at aligning staging systems with the needs of modern medicine [1]. However, questions about its practical application and compatibility with clinical requirements remain open.
FIGO classification systems are fundamental tools for assessing the stage of EC, enabling the comparability of clinical data and treatment outcomes worldwide. Since the introduction of the first classification, these systems have evolved to better address clinical needs and integrate advancements in medical knowledge.
The early FIGO systems were based solely on assessing the anatomical spread of the tumor. The groundbreaking 1988 classification introduced parameters such as the depth of myometrial invasion and cervical involvement, setting a new standard for assessing disease progression [2]. The subsequent 2009 update marked another significant advance, introducing more precise criteria for lymph node assessment and a detailed division into substages [3].
However, the 2009 classification remained primarily relied on anatomical parameters, overlooking crucial discoveries in the molecular biology of endometrial cancer, which limited its use in personalized treatment strategies.
In response to the demands of modern oncology, FIGO 2023 integrates classical anatomical features with key molecular and pathological factors [1]. The goal of these changes was to create a system that not only predicts the disease course more accurately but also aids in therapeutic decision-making based on tumor biology. This update has the potential to revolutionize the approach to EC diagnosis and treatment, although its implementation presents several challenges.
Key differences between FIGO 2009 and FIGO 2023
The FIGO 2023 classification for EC introduces several significant changes compared to the 2009 version, reflecting advancements in oncology diagnostics and treatment. The most notable differences include:
1. Inclusion of molecular subtypes.
For the first time, the classification incorporates four key molecular subtypes of endometrial cancer: POLEmut, mismatch repair deficient (MMRd), p53abn, and no specific molecular profile (NSMP). This stratification allows for more precise risk assessment and improved alignment of treatment strategies with tumor biology.
2. Expansion of substages.
The new classification introduces a more detailed division into substages within each stage of disease progression. This approach enables a more nuanced assessment of tumor advancement and facilitates therapeutic decision-making.
3. Significance and differentiation of lymphovascular space invasion (LVSI).
The inclusion of LVSI as a critical prognostic parameter, with a distinction between substantial and focal invasion, allows for more refined risk stratification of patients.
4. Revised definitions for stages III and IV.
Stages III and IV have been more precisely defined, incorporating the location and characteristics of metastases. This enhances the ability to tailor therapy for advanced-stage disease.
5. Introduction of sentinel lymph node (SLN) biopsy and ultrastaging.
The FIGO 2023 classification acknowledges SLN biopsy and ultrastaging as standard methods for lymph node evaluation. These techniques improve diagnostic accuracy, particularly in detecting micrometastases.
6. Division into aggressive and non-aggressive tumors.
The classification emphasizes distinctions between tumors with favorable prognoses, such as POLEmut, and aggressive tumors, such as p53abn. This differentiation supports the adaptation of treatment intensity based on the individual patient’s risk profile.
Division into aggressive and non-aggressive tumors
One of the key principles of the FIGO 2023 classification is the clear distinction between aggressive and non-aggressive tumors. This dichotomy reflects differences in tumor biology, prognosis, and recommended therapeutic strategies.
Non-aggressive tumors
Non-aggressive tumors include low-grade tumors (G1 and G2), which are characterized by a favorable prognosis. These tumors typically exhibit limited invasiveness and a low risk of metastasis. For this group, FIGO 2023 incorporates additional prognostic factors, such as:
depth of myometrial invasion,
cervical involvement,
presence of LVSI [1].
These parameters play a crucial role in risk assessment and enable individualized treatment, including potential de-escalation of therapy to avoid overtreatment.
Aggressive tumors
Aggressive tumors include high-grade tumors (G3) and those with unfavorable histological phenotypes, such as serous, clear cell, and mixed carcinomas. These tumors are characterized by:
a higher risk of recurrence and metastasis,
deeper tissue invasion,
more aggressive clinical behavior [4].
For aggressive tumors, the FIGO 2023 classification primarily relies on the histological type and molecular subtype (e.g., p53abn), which aligns with their biological specificity. However, this approach excludes additional prognostic factors such as LVSI or the depth of invasion. The omission of these parameters in aggressive tumors may limit the ability to perform more precise risk stratification.
Discrepancies in assessment and diagnostic challenges
The division into aggressive and non-aggressive tumors is complicated by significant variability in histopathological assessment. For example, in the PORTEC-3 trial, 43% of cases required pathological review, which altered patient risk stratification, and in 8% of cases, eligibility for the trial was completely changed [5].
An audit conducted by Spoor and Cross revealed that in 12% of cases initially classified as low-grade tumors (G1–G2), the classification was revised to high-grade (G3), while in 6% of cases, the classification was downgraded from high-grade to low-grade [6].
Particular challenges arise in the classification of tumors with intermediate features, such as G2. In the GOG-99 study, G2 tumors were grouped with G3 due to their more aggressive clinical behavior compared to G1 tumors [7]. This grouping highlights the difficulties in definitively determining their clinical behavior and prognosis, which can impact treatment choices.
For intermediate-risk groups, which include G2 tumors, even small differences in interpretation can influence treatment decisions. For instance, the presence of LVSI or deeper myometrial invasion may lead to treatment escalation, while their absence might justify a more conservative approach. Making optimal therapeutic decisions requires consistent and precise diagnostic criteria, which remains a significant challenge.
Lymphovascular space invasion
One of the most significant additions to the FIGO 2023 classification is the recognition of LVSI as an independent prognostic factor [8]. This decision reflects the growing importance of this parameter in risk assessment and treatment planning for EC. Lymphovascular space invasion, defined as the presence of tumor cells in lymphatic or blood vessels surrounding the tumor, is associated with an increased risk of metastasis and disease recurrence [9].
Despite widespread acknowledgment of LVSI’s prognostic value, its assessment remains challenging [10]. The division of LVSI into substantial (≥ 5 vessel involvements per high-power field) and focal (< 5 vessel involvements per high-power field) represents a step towards more detailed risk stratification [11]. However, the lack of uniform diagnostic criteria complicates the interpretation of results. For instance, the World Health Organization (WHO) defines substantial LVSI as involving at least five vessels, whereas the National Comprehensive Cancer Network adopts a threshold of four vessels [11]. Additionally, retrospective studies often apply varying definitions, limiting the comparability of results [12].
In a study by Turashvili et al. the intraclass correlation coefficient for assessing LVSI presence was only 0.6, and for substantial LVSI, it was 0.56 [13]. This low level of interobserver reproducibility among pathologists can lead to errors in assigning patients to appropriate risk groups, potentially resulting in overtreatment or undertreatment.
On the other hand, an analysis conducted on a cohort of 335 EC patients revealed that the presence of substantial LVSI did not significantly affect local recurrence-free survival or distant metastasis-free survival compared to patients with focal LVSI or no LVSI in cases without lymph node metastases [14]. These findings suggest that LVSI’s role as a prognostic factor warrants further investigation and greater precision in its assessment.
Certain difficulties in evaluating LVSI stem from potential diagnostic artifacts. In specimens obtained during laparoscopic hysterectomy, especially when uterine manipulators are used, so-called pseudo-vascular invasion may occur because of mechanical disruption of tumor tissue. These artifacts can lead to misinterpretation of results and inappropriate treatment [15].
These examples suggest that the inclusion of LVSI as an independent prognostic factor in the FIGO 2023 classification, while significant, may represent an overly ambitious extension. The challenges in achieving consistent evaluation of this parameter, the lack of standardized diagnostic criteria, and limited data regarding its impact on treatment outcomes indicate that LVSI’s role requires further validation.
Conducting studies aimed at standardizing diagnostic procedures and better understanding the significance of LVSI in the context of other prognostic factors is essential.
Redefinition of stages III and IV – ultrastaging and diagnostic challenges
The introduction of a detailed subdivision for stages III and IV in the FIGO 2023 classification allows for a more precise representation of disease progression and supports therapeutic decision-making. These changes account for the location of metastases and their molecular characteristics, representing a step towards more personalized treatment for endometrial cancer.
The role of ultrastaging and lymph node assessment
Accurate classification requires advanced diagnostic methods such as lymph node ultrastaging, which enables the detection of micrometastases and minimal residual disease [16]. This procedure plays a crucial role in precisely assessing disease stage, but its effectiveness depends on resource availability and the expertise of diagnostic teams.
The lack of standardized ultrastaging protocols across different centers primarily concerns pathological assessment methods, affecting consistency and comparability of results. Variability in the number of sections analyzed, staining techniques, and result interpretation can lead to discrepancies in staging classification. An additional challenge, often overlooked in clinical practice, is the inadequate evaluation of para-aortic lymph nodes, which are not routinely sampled unless clinically suspicious. This approach may lead to an underestimation of disease stage, with significant prognostic and therapeutic implications [16, 17].
POLEmut subtype – exceptionally favorable prognosis, but questions still remain
Mutations in the POLE (DNA polymerase epsilon) gene are ultramutational changes that occur in approximately 10% of EC cases, primarily in endometrioid tumors. A hallmark of the POLEmut subtype is its exceptionally favorable prognosis [18].
Retrospective studies have shown that patients with POLE mutations exhibit the lowest risk of recurrence and mortality compared to other molecular subtypes. For instance, retrospective analyses report a hazard ratio (HR) for overall survival (OS) of 0.36, indicating a significant reduction in mortality risk [4]. This unique feature allows for treatment de-escalation, reducing the intensity of adjuvant therapy. By minimizing side effects while maintaining high treatment efficacy, de-escalation represents a significant advancement toward personalized therapy.
Despite its favorable prognosis, the rarity of POLEmut and the limited number of cases described in the scientific literature result in uncertainty regarding optimal management strategies. Moreover, the follow-up periods in many studies are too short to conclusively evaluate the long-term effects of treatment reduction [19, 20].
There is a lack of randomized clinical trials to definitively confirm whether the exceptionally favorable prognosis in POLEmut cases is solely attributable to tumor biology or influenced by adjuvant treatments commonly administered to patients in retrospective cohorts. For example, a study by McAlpine et al., involving 395 patients with POLE mutations, reported favorable outcomes but advised caution when deciding to forgo/give up adjuvant therapy, even in cases with good prognosis [18]. Ongoing research, such as the RAINBO trial (refining adjuvant treatment in EC based on molecular features), specifically focuses on tailoring adjuvant treatment strategies for molecular subtypes, including POLEmut tumors. This study aims to clarify whether treatment de-escalation is a safe approach for these patients and may provide essential evidence to guide future clinical practice [21].
Significance and challenges of the p53abn subtype
P53abn tumors are characterized by high aggressiveness, resulting in an increased risk of recurrence and lymph node metastasis, regardless of the disease stage. Studies suggest that identifying p53 status enables more accurate prognostication and tailoring of the therapy to the tumor’s specific characteristics [22].
Patients with p53abn derive significant benefits from intensified treatment, particularly combined chemoradiotherapy (CHRT). According to available data, this approach improves five-year recurrence-free survival by 22% and OS by 23% compared to less intensive therapeutic strategies [23].
Additionally, some reports suggest that this subtype may exhibit greater sensitivity to immune checkpoint inhibitors, highlighting its potential responsiveness to immunotherapy [24].
Accurate assessment of TP53 status and p53 protein expression
The accurate evaluation of TP53 gene status and p53 protein expression presents several challenges. Immunohistochemical (IHC) staining for p53, widely used in clinical practice, is not always a reliable surrogate for detecting TP53 mutations. Tumor heterogeneity can result in abnormal p53 protein expression being present only in certain tumor regions, increasing the risk of diagnostic errors. Such inaccuracies may lead to inappropriate treatment, impacting therapeutic outcomes [25].
To improve the reliability of p53 status assessment, WHO guidelines define three primary IHC staining patterns: overexpression, complete absence, and wild-type expression. These standardized cutoffs help classify tumors according to their molecular profile. However, differences in sample preparation and staining protocols can lead to interobserver variability, emphasizing the need for centralized evaluation of specimens obtained from dilatation and curettage or surgical resection to ensure consistent and accurate interpretation [26].
Although IHC remains the most accessible method for assessing p53 status, it has limitations. Next-generation sequencing provides a more detailed analysis of TP53 mutations, but some detected mutations may have limited clinical significance. This highlights the need for further research to refine diagnostic criteria and standardize the interpretation of TP53 mutations in clinical decision-making.
Misclassification of p53 status can have serious clinical consequences. Over-intensification of treatment, such as administering CHRT to patients with superficial invasion without clear indications, may cause unnecessary side effects. Conversely, underestimating p53abn may lead to insufficient treatment, increasing the risk of recurrence, especially in highly aggressive tumors. However, current literature lacks definitive evidence to fully support this hypothesis, underscoring the necessity for further prospective studies.
Another issue related to the impact of TP53 alterations on clinical practice is the regional variability in assessing p53 status. A study conducted in southeastern Poland revealed significant differences in the frequency of p53 mutation detection among oncological centers, ranging 9–26% [27]. This variability may stem from differences in the availability of molecular technologies, lack of standardized diagnostic protocols, and variations in pathological assessment criteria.
These findings underscore the urgent need for harmonizing diagnostic methods to minimize inter-center discrepancies and improve the reliability of molecular classification. As a highly aggressive subtype of endometrial cancer, p53abn requires special attention in clinical practice. Further research and harmonization of diagnostic standards are essential to ensure accurate diagnosis and optimize therapeutic strategies.
Mismatch repair deficient subtype – mixed outcomes and therapeutic controversies
The mismatch repair deficient subtype, associated with microsatellite instability, occurs in approximately 20–30% of EC cases. These tumors exhibit unique biological properties that are both prognostically and predictively significant. A particularly notable feature of this subtype is its heightened sensitivity to immune checkpoint inhibitors, presenting new therapeutic opportunities [28].
Mismatch repair deficient tumors demonstrate increased responsiveness to immunotherapy with checkpoint inhibitors. Recent studies suggest that combining immunotherapy with chemotherapy could become the standard treatment in both first-line and subsequent therapies. There is growing optimism about potentially replacing chemotherapy with immunotherapy alone in first-line treatment, which could reduce treatment toxicity [28, 29].
Despite promising outcomes, the introduction of immunotherapy for MMRd tumors faces several challenges:
These challenges underscore the need for harmonized diagnostic protocols and further research to optimize treatment strategies for patients with MMRd tumors.
Additional molecular features in no specific molecular profile
The no specific molecular profile subtype encompasses tumors that do not exhibit the defining features of other molecular groups – POLEmut, MMRd, and p53abn. It represents the largest category within the PROMISE classification [32], recommended by European scientific societies such as ESMO and ESGO [4]. However, the NSMP subtype is characterized by significant biological heterogeneity, which considerably complicates clear prognostic assessment and precise therapeutic decision-making.
Within the NSMP group, several molecular characteristics have been identified which could aid in further differentiation of this heterogeneous category:
1. CTNNB1 mutations.
Mutations in the CTNNB1 gene, which activate the Wnt signaling pathway, are associated with a higher risk of recurrence and worse disease-free survival (HR = 2.83) in the NSMP subtype. Including CTNNB1 status as an additional prognostic factor could significantly improve patient stratification and facilitate tailored therapies for this group [33].
2. High L1CAM expression.
L1CAM expression correlates with aggressive clinical features, including lymph node metastases and the presence of LVSI. Patients with NSMP tumors exhibiting high L1CAM expression have markedly worse prognoses, with significantly reduced OS (HR = 3.62) and recurrence-free survival (HR = 4.11). Incorporating L1CAM as a biomarker could enhance prognostic accuracy and support more precise therapeutic approaches [34].
3. Absence of estrogen receptor (ER) expression.
The absence of ER expression is a strong prognostic factor associated with higher recurrence risk and poorer treatment outcomes. In the NSMP subtype, ER positivity is linked to better prognosis, while its absence indicates a more aggressive tumor phenotype [35].
The no specific molecular profile subtype’s heterogeneity highlights the need for additional biomarkers, such as CTNNB1 mutations, L1CAM expression, and ER status, to improve prognostic accuracy and guide personalized treatment strategies. Incorporating these factors into clinical practice could help overcome the challenges posed by the diverse nature of NSMP tumors and provide more effective patient care.
Tumor heterogeneity challenges
Tumor heterogeneity, both at the genetic and IHC levels, remains one of the most significant challenges in the diagnosis and treatment of endometrial cancer. Different clonal populations of tumor cells within the same tumor can exhibit distinct molecular features, which impacts treatment response [30]. This phenomenon presents several challenges:
genetic variability – within a single tumor, cells may harbor different mutations, complicating molecular and prognostic assessments,
immunohistochemical heterogeneity – the expression of markers such as p53, L1CAM, or MMR can be inconsistent across the tumor, making result interpretation difficult and potentially leading to incorrect therapeutic decisions,
limitations of standard biopsies – fragmentary analysis of tissues obtained during biopsy may not capture the full molecular profile of the tumor. Consequently, this increases the risk of misclassification and suboptimal treatment decisions.
Tumor heterogeneity necessitates a more comprehensive diagnostic approach. Advanced technologies, such as spatial analysis and multigene profiling, may provide better insights into this complexity. Incorporating heterogeneity into classifications and treatment algorithms is essential for more accurately tailoring therapy to the tumor’s biology.
Conclusions
An effective classification system should support clinicians in treatment planning, provide reliable prognostic guidance, enable the evaluation of therapeutic outcomes, facilitate communication between centers, and promote advancements in cancer research. By introducing molecular subtypes of EC and enabling more precise differentiation, the FIGO 2023 system meets some of these objectives but also presents significant challenges.
The lack of consistent diagnostic standards, particularly in the assessment of LVSI and SLN ultrastaging, complicates the interpretation of results and affects therapeutic decisions. Additionally, the limited availability of molecular technologies and difficulties in their implementation constrain the full utilization of the system’s potential.
Tumor heterogeneity, both genetic and IHC, further complicates the diagnostic process, increasing the risk of incorrect therapeutic decisions based on incomplete data. Moreover, many recommendations are based on retrospective studies, underscoring the need for prospective analyses to validate their clinical value.
Despite these challenges, FIGO 2023 represents a significant step forward in the personalization of EC treatment. Key factors for its successful implementation include the standardization of diagnostic procedures, training of medical personnel, and technological advancements. Future research should focus on further differentiation within molecular groups, such as NSMP, and the identification of new biomarkers. Harmonizing standards and implementing consistent diagnostic protocols will better address clinical needs and contribute to improved treatment outcomes for patients.
