New FIGO 2023 Staging System of Endometrial Cancer: An Updated Review on a Current Hot Topic.

The introduction of molecular classification results is the main and most discussed change in the new staging system. The molecular classification of endometrial cancer and its clinical relevance is rapidly evolving, which could lead to some contradictions in the future 6 10 16 . Moreover, there is a lack of recent prospective studies addressing the integration of molecular classification into FIGO 2023 EC staging. The methodology for determining the molecular group is not specified in the 2023 FIGO update, which may cause differences in the interpretation of results between laboratories. However, the WHO has specified and approved the methodology for performing immunohistochemistry in detail 17 18 . Based on current evidence, a combination of immunohistochemistry and separate POLE testing appears to be the preferred approach 3 4 15 19 . Nevertheless, these auxiliary tools can exhibit a variability in the possible workflow and interpretation of results. Previous papers comparing immunohistochemistry (IHC) and polymerase chain reaction (PCR) in endometrial tumours have shown discordance rates as high as 5 to 10%. However, recent studies have indicated that next-generation sequencing (NGS) can help resolve discrepant results when selecting the appropriate tumour tissue for both IHC and molecular is ensured 20 . The introduction of new sequencing panels for NGS to extend the testing of POLE mutations has on the one hand also provided information on new pathogenic or likely pathogenic POLE mutations, but on the other hand identified new variants of unknown significance. There are currently twelve known mutations of the POLE gene that are proven to be pathogenic 19 . This spectrum will undoubtedly continue to expand. The distinction between pathogenic and non-pathogenic POLE mut may not be clear-cut. Due to the high mutational load, non-pathogenic POLE mutations are mostly found in the group of hypermutated MMRd tumours as a secondary genetic event and the biological behaviour is consistent with the characteristics of this group. Since a positive POLE mutation result often changes the stage of the disease and thus the indication for adjuvant therapy, knowledge of the correct pathogenic POLE mutations is essential for the classification of endometrial cancer as a “ POLE mut” 16 . Immunohistochemical detection of abnormal p53 expression is very sensitive with respect to detection of TP53 gene mutation status; only a minority of cases with mutations show normal expression type on immunohistochemical level. A study on ovarian carcinomas showed that with proper validation, immunohistochemical testing has a concordance with TP53 gene mutation status of 100% specificity and 95.9% sensitivity 21 . However, unlike ovarian cancer, endometrial carcinoma can have abnormal, mostly clonal expression patterns, which are particularly common in patients with MMRd or POLE mutations 21 . Patients with TP53 -mutated cancers are the only ones among 4 molecular groups to benefit from chemoradiotherapy 22 . Misclassification of the tumour as TP53 abn has significant negative consequences for the patient. Some experts believe that p53 immunostaining is not a perfect surrogate of TP53 mutations 23 . A study on 132 patients with endometrial cancer observed sensitivity, specificity, positive (PPV), and negative predictive values (NPV) of IHC for MSI status in 89.3%, 87.3%, 78.1%, and 94.1%, and for p53 in 92.3%, 77.1%, 60.0% and 96.4% of cases, respectively. The authors concluded that the moderate concordance (Cohen’s kappa coefficient of 0.59) between IHC and NGS for p53 status implies that they cannot be used interchangeably 24 . Singh et al. found the concordance between p53 IHC and TP53 mutation in 92.3% (155/168) of cases overall, and in 95.1% (117/123) after excluding MMRd and POLEmut endometrial cancers 25 . Matsumoto et al. demonstrated the concordance of p53 IHC and NGS for TP53 mutation status in 100% (43/43) of high-grade endometrial cancers 26 . A meta-analysis assessing the diagnostic accuracy of p53 immunohistochemistry as surrogate for TP53 sequencing included 13 studies with 727 endometrial cancers. Immunohistochemical criteria used to define aberrant p53 expression were “overexpression” and “overexpression or complete absence”. Both “overexpression” and “overexpression or complete absence” showed high diagnostic accuracy (area under the curve, AUC = 0.9088 and 0.9030, respectively). The subgroup with “overexpression” and NGS showed the best results, with very high diagnostic accuracy (AUC = 0.9927). The conclusions of this meta-analysis confirmed that IHC is a highly accurate surrogate of TP53 sequencing 27 . In addition, 3 to 6% of all endometrial carcinomas have more than one molecular characteristic (referred to as multiple classifier) 2 3 4 15 28 29 . The most common is the “double classifier” (e.g., POLEmut and p53abn or MMRd and p53abn), but there are also “triple classifiers” (POLEmut and MMRd and p53abn simultaneously). In these cases, p53abn is a secondary manifestation of the ultramutated state in POLE mutation or within the microsatellite instability in MMRd. The prognosis of these tumours is then based on the POLE mutation or MMRd and the appropriate stage is assigned accordingly 29 . The current consensus is that in tumours with a pathogenic mutation of the POLE gene, this mutation is superior to all other abnormalities, and such tumours should be classified as POLEmut 4 . However, data are limited for tumours that exhibit both the common molecular characteristics of POLEmut and MMRd and in these cases more detailed investigation should be considered to exclude Lynch syndrome 15 . Although molecular classification is not a mandatory part of the new FIGO staging, its inclusion appears problematic for several reasons, including the varying availability across different countries. In addition, there are other immunohistochemical markers (mainly estrogen [ER] and progesterone receptor [PR] expression) that are more affordable and significantly improve prognosis 30 , but have not been included in the staging or in the ESGO/ESTRO/ESP guidelines at all. This was confirmed, for example, in a prospective study of 132 patients 31 . In the preoperative stratification of low- or high-risk patients, the authors evaluated the importance of the presence of certain immunohistochemical biomarkers (L1CAM, ER, PR and p53) in addition to standard criteria (age, stage, histological type, grade, lymphovascular invasion). The results of immunohistochemical markers significantly improved the sensitivity for determining the high-risk group (48.4% vs. 75.8%, p < 0.001) with a statistically insignificant decrease in specificity to 80% (p = 0.238). The positive predictive value (PPV) was similar for the two methods, whereas the negative predictive value (NPV) (i.e., the probability of extremely low risk in negative test cases) improved statistically significantly (66.0% vs. 78.9%, p < 0.001) 31 . Another study of 763 EC patients showed that abnormal expression of p53, L1CAM, ER or PR was significantly associated with a higher risk. Moreover, ER-/PR-negative status and p53abn were independently associated with reduced disease-specific survival (HR 2.74; 95% CI 1.48–5.07; p = 0.001 and HR 1.88; 95% CI 1.00–3.51; p = 0.048, respectively) 32 . Another limitation of the FIGO 2023 staging system of endometrial cancer is the inclusion of variables whose definitions are still evolving, as well as variables that are subject to considerable interobserver variability in their assessment. The presence of substantial LVSI plays a significant role in tumour prognosis, and therefore this parameter has been implemented in staging 6 33 34 35 . The problem remains that several different definitions are used to quantify LVSI. FIGO 2023 as well as the WHO Classification of Female Genital Tumours 17 18 36 and ESGO/ESTRO/ESP guidelines 10 define substantial (or extensive) LVSI as involvement of ≥ 5 lymphovascular structures (focal LVSI < 5 vessels involved, negative LVSI is without vessel involvement). However, other organisations use different definitions. The NCCN (National Comprehensive Cancer Network) defines substantial LVSI as involvement of ≥ 4 lymphovascular structures on a single hematoxylin-eosin-stained section 37 and the ICCR (International Collaboration on Cancer Reporting) 38 , ISGyP (The International Society of Gynaecological Pathologists) 39 , CAP (College of American Pathologists) 40 , DGGG (Deutsche Gesellschaft für Gynäkologie und Geburtshilfe, German Society for Gynecology and Obstetrics) and DKG (Deutsche Krebsgesellschaft, German Cancer Society) 41 , only ≥ 3 vessels are sufficient to define massive LVSI involvement. Additionally, most definitions (including FIGO 2023) do not specify whether the extent of LVSI is based on the maximum involvement in a single tissue section or the cumulative involvement across all examined sections of the affected tissue. This ambiguity may lead to the assignment of different stages in oncologically identical patients with different outcomes of identical therapy 16 . Another example is the renewed requirement for a histological distinction between tumours confined to the endometrium and those with minimal invasion of the myometrium. This distinction is particularly problematic in non-aggressive (low grade endometrioid) carcinomas. Accurate histological assessment is often difficult because of the presence of adenomyosis and the irregular interface between the endometrium and myometrium. Consequently, the description of tumour invasion into the myometrium can vary considerably from one institution to another 16 23 39 . Another issue is the histopathological determination of uterine serosa tumour involvement, which is listed as “uterine subserosa” for the definition of Stage IIIA2. ISGyP includes submesothelial fibrous tissue as part of the definition of serous invasion 39 . The definition of subserous (as opposed to serous) invasion lacks clearly defined histopathological criteria and is not included in any professional guidelines or other source documents and cannot be clearly interpreted 16 23 . Already in 2021, ISGyP conducted a questionnaire survey addressing many of the questions that were later integrated into the 2023 FIGO staging system among ISGyP and International Gynecologic Cancer Society members and received responses from 172 pathologists and 135 clinical oncogynaecologists 42 . Relatively few questions elicited a consensus response, defined as at least 75% agreement. Consensus agreement was reached on differentiating nodal involvement into micro- and macrometastases, defining criteria for categorizing cases of simultaneous uterine body and ovarian cancer LR endometrioid cancer, and the importance of LVSI expression as an independent risk factor. However, only half of pathologists (48%) and two thirds of oncogynaecologists (61%) agreed with the implementation of LVSI results in the staging system, and there was similar support for the inclusion of molecular classification (48% of pathologists and 63% of oncogynaecologists) or histological type of tumour (52% of pathologists and 65% of oncogynaecologists) in staging 42 . Interestingly, while two thirds of clinical oncologists agreed with these changes, approximately half of pathologists were in favour while the other half were not. This reflects current opinions on the FIGO 2023 staging of endometrial cancer 42 . While pathologists tend to criticise it 16 23 , clinicians have embraced the changes, although they raise some questions 43 . The parameters needed to determine the specific stage, with indication of appropriate therapy, largely depend on histopathological parameters. Some experts believe that the lack of representation of the pathology community was a major shortcoming in the design of the FIGO 2023 staging 16 . Many of the implemented changes and recommendations are presented without any citations to the original peer-reviewed literature. An example is the inclusion of grade 3 endometrioid carcinoma in the group of “aggressive” histological types, alongside other non-endometrioid carcinomas, although it is a markedly molecularly and clinically heterogeneous group. This stratification is not substantially supported by the cited literature, and the 2021 ESGO/ESTRO/ESP guidelines also separate grade 3 endometrioid carcinoma from other aggressive non-endometrioid carcinomas ( Table 2 ) 10 16 36 .

The general definition of Stage IV is the same (spread to the bladder mucosa and/or intestinal mucosa and/or distant metastasis). Local invasion of bladder mucosa and/or intestinal/ bowel mucosa remains Stage IVA. However, an additional substage was added for patients with extrapelvic peritoneal metastasis (Stage IVB), patients with involvement of the pelvic peritoneum belong to Stage IIIB2. Stage IVC means the presence of distant metastases. Comment: Isolated peritoneal carcinomatosis (new Stage IIIB2) is rare (approximately 2% of all patients with endometrial carcinomas) and these patients should be distinguished from those with distant metastases, because, unlike in Stage IV of endometrial cancer, they are usually indicated for primary surgical treatment 6 10 13 14 . The level of lymph nodes involvement between Stage III and IV remains the same. Distant metastases mean involvement of intra-abdominal nodes above the renal vessels and/or any extra- or intra-abdominal lymph nodes.

The new staging system for endometrial cancer FIGO 2023 differs significantly from previous versions, as it includes additional histopathological parameters and molecular classification of the tumour in the definition of individual stages, alongside the traditional anatomical extent of the tumour. This change has further strengthened the role of the pathologist in staging. The inclusion of these parameters has significantly refined the prognosis estimation of each stage, enabling the indication of targeted therapy. On the other hand, the FIGO 2023 staging system may appear overly complex, and its implementation in daily practice requires the full involvement of all stakeholders.

The last change to the endometrial cancer staging system was adopted by FIGO in 2008 and published in 2009. Two years later, similar changes were implemented in the TNM system 5 . The traditional concept of staging system was based on the description of the anatomical extent of the disease at the time of diagnosis based on clinical, radiological and possibly histopathological examination. Staging is the key and often the strongest prognostic factor for an individual patient. The new FIGO 2023 staging system aims to define individual stages with a more accurate prognosis by incorporating additional important prognostic factors that simple anatomical spread cannot capture. This approach, which includes histological-pathological and molecular data, has already been applied to staging systems of breast, head and neck, and prostate carcinomas 16 44 . The 2021 joint ESGO/ESTRO/ESP guidelines follow the same principle and categorise patients with endometrial cancer into several risk groups (low, intermediate, medium-high and high risk). This principle is also used in the joint guidelines for endometrial cancer of the German-speaking professional societies 41 45 . The definition of risk groups varies on whether the result of molecular classification of the tumour is known ( Table 2 ) 10 . Some experts find the FIGO 2023 staging system complicated and non-intuitive, with difficulty to facilitate correspondence with the 2009 FIGO staging 23 . This can make it significantly more difficult to adopt more widely in regions with different levels of healthcare. Incorporation of cancer molecular classification results into FIGO 2023 staging essentially precludes staging in under-resourced areas. This will fundamentally complicate the collection of health data for clinical, epidemiological and research purposes, including clinical trials. For existing clinical trials, it is not clear how the inclusion criteria could be adapted to fit the new FIGO staging system 16 . Also, the stage can change several times in one patient in a relatively short period of time. For example, the histology from a hysterectomy may show Stage IIC (aggressive histological type with myometrial invasion), but after review at the local cancer centre the resulting stage changes to, for example, IAm POLEmut (evidence of a pathogenic POLE mutation). It is still a matter of debate whether the stage of the disease should be an auxiliary component in the calculation of the patient’s risk stratification, or whether the stage, by incorporating other prognostic parameters beyond the anatomical extent of the tumour, should reflect the most accurate risk of a particular patient, i.e. whether the staging system should be the defining model itself, expressing the most accurate risk. The phase of refining the prognosis of a specific stage by incorporating various histopathological and molecular variables makes the determination of this stage significantly more complicated. Therefore, maintaining definitions of individual stages based on anatomical spread seems to be a better solution worldwide. Paradoxically, the sudden transition to the full FIGO 2023 staging system as proposed by FIGO may hinder progress towards a more accurate prognosis for individual stages 42 . The new FIGO 2023 endometrial cancer staging system has some undeniable advantages over the previous version. The new FIGO 2023 staging system reflects that endometrial cancer is not a single disease. The undisputed fact remains that the implementation of new histopathological parameters and molecular classification results for endometrial cancer will significantly improve the prognosis of a specific patient at a specific stage 2 4 6 15 . This was confirmed by an international retrospective study from three oncogynaecology centres in Austria and Italy comparing the FIGO 2009 and FIGO 2023 staging systems 9 . In this study, all applied statistical tests confirmed more accurate prediction of PFS and OS (overall survival) using the FIGO 2023 staging system compared to FIGO 2009 9 . A retrospective analysis of 519 patients with endometrial cancer showed a significantly higher 5-year PFS in Stage I according to the FIGO 2023 staging system compared to FIGO 2009 (93.0% vs. 87.4%). Two new molecularly defined sub-stages IAm POLEmut and IICm p53abn according to 2023 FIGO showed completely different oncological results 9 . Molecular classification can define patients with an excellent prognosis ( POLE mut) from a group with a poor prognosis ( p53 abn) 3 4 15 29 . This is a major advance in the diagnosis and treatment of endometrial cancer in the last 10 years 6 . MMRd testing plays an important role in screening for Lynch syndrome and is also a predictor of response to immunotherapy in advanced tumours using immune check point inhibitors (ICI). Two randomised phase III trials (ENGOT-en6/GOG-3031/RUBY and NRG-GY018/Keynote-868) have demonstrated a statistically significant PFS advantage with the addition of an immune checkpoint inhibitor (ICI) (dostarlimab or pembolizumab, respectively) to standard carboplatin/paclitaxel chemotherapy followed by ICI maintenance therapy in MMRd patients with a hazard ratio (HR) of 0.28 (95% confidence interval [CI] 0.16–0.5) and 0.30 (95% CI 0.19–0.48), respectively 46 47 . Similarly, phase III trial (AtTEnd) observed significantly increased PGS in patients with advanced or recurrent endometrial cancer (especially those with MMRd cancers) treated with ICI atezolizumab added to standard carboplatin/paclitaxel chemotherapy (HR 0.4; 95% CI 0.61–0.91; p = 0.022) compared to controls 48 . Recent trial (phase III DUO-E) showed statistically significant PFS benefit in ICI durvalumab (plus standard carboplatin/paclitaxel chemotherapy) (HR 0.71; 95% CI 0.57–0.89; p = 0.003) and durvalumab + olaparib (plus standard chemotherapy) arms (HR 0.55; 95% CI 0.43–0.69; p < 0.0001) versus standard chemotherapy arm among advanced or recurrent endometrial cancer patients. The PFS benefit was higher in MMRd subjects (HR 0.42; 95% CI 0.22–0.80 and HR 0.41; 95% CI 0.21–0.75, respectively) 49 . Molecular classification is particularly important in aggressive histological types. HG endometrioid carcinoma benefits most from molecular classification because it is a clinically, molecularly and prognostically very heterogeneous disease 36 . Without molecular classification, HG endometrioid carcinoma cannot be stratified into the appropriate risk group. Due to early symptomatology, almost two thirds of patients with endometrial cancer are diagnosed at early stages allowing for primary surgical treatment, and a relatively small proportion of patients are then indicated for adjuvant therapy, sometimes based solely on molecular classification results. Patients with POLE mut group cancers significantly benefit in a reduction of postoperative adjuvant radiotherapy in terms of elimination of adverse effects while maintaining the same prognosis 3 4 6 10 . In contrast, the detection of p53 abn has a significantly worse prognosis and extensive adjuvant therapy may be beneficial. However, according to our own experience in clinical practice, some clinical oncologists may have difficulty relying solely on the outcome of pathogenic POLE mut. They prefer to administer adjuvant therapy even in POLE mut carcinomas that are considered very aggressive by other criteria. The use of molecular classification is also likely to be used to determine the risk of disease for fertility preserving procedures in young women with endometrial cancer who meet the prescribed criteria. Furthermore, molecular classification is feasible in routine practice for all patients with endometrial cancer and does not prolong the time needed to decide on adjuvant therapy 50 . The largest study in Central Europe included a total of 270 molecularly classified endometrial cancers 51 . In total, 6.6% (18/270) of subjects had POLE mut, 31.5% (85/270) had MMRd, 11.1% (30/270) had TP53 mut, and 50.7% (137/270) had no specific molecular profile. Thirteen cases (4.8%) were classified as “multiple classifiers”. The NSMP group was often characterised by multiple genetic alterations, the most common being mutations in the PTEN (44%), PIK3CA (30%), ARID1A (21%) and KRAS (9%) genes 51 . NSMP endometrial carcinomas are a heterogenous group of tumours and comprise both aggressive and low-risk ECs; therefore NSMP does not change the stage. Within the NSMP cancer group, certain features significantly affect the patient’s prognosis. Grade 3 and/or ER-negative status were responsible for most of the disease-specific deaths at 5 years (HR 16.3; 95% CI: 8.4–31.7) compared with low-risk NSMP endometrial cancers (grade 1–2, ER-positive) 52 . A study on 648 cases revealed that only ER-positivity was independently associated with a reduced risk of recurrence (HR 0.33, 95% CI: 0.15–0.75) in high-risk NSMP endometrial cancers 53 . Treatment de-escalation could be considered in ER-positive NSMP ECs, which constitute the vast majority of NSMP ECs 52 . Assessment of ER status in high-risk NSMP EC is feasible in clinical practice and can improve risk stratification and treatment. Multiple classifier endometrial cancers are at a higher risk of being classified into a different molecular group with a different prognosis estimate. In cases analysed solely by Sanger sequencing of the POLE gene, which has lower analytical sensitivity than NGS testing, pathogenic mutations in the POLE gene may go undetected, leading to misclassification of the EC as p53abn or MMRd. The clinical outcomes of patients with MMRd-p53abn and POLEmut-p53abn endometrial carcinomas, exhibiting a 5-year recurrence-free survival (RFS) of 92.2% and 94.1% for Stage I, respectively, were significantly different from those of single-classifier p53abn endometrial carcinomas, which had a Stage I RFS of 70.8% (p = 0.024 and p = 0.050, respectively) 29 . The authors of a prospective study with a follow-up duration of 24.7 months state that “multiple classifier” endometrial carcinomas have the potential to behave aggressively, and their categorization as POLEmut EC with treatment de-escalation may therefore not be safe 51 . However, the recent cohort study observed no recurrences in 15 POLEmut-p53abn cases, 2 POLEmut-MMRd and 3 POLEmut-MMRd-p53abn ECs during a median follow-up of 12.8 and 17.0 months, respectively. In contrast, recurrences were noted in 7.1% (2/28) of MMRd-p53abn cases at 5.0 and 6.9 months post-surgery, while MMRd and p53abn recurred in 4.0% (4/99) and 34.2% (25/73) cases, respectively, with a median time to recurrence of 8.8 and 8.4 months 54 . Collaborative prospective multi-institutional studies are needed to evaluate the prognostic significance of multiple classifiers, since current results are based on a limited number of patients. The experts also welcomed the creation of Stage IA3 for LR cancer with synchronous involvement of the endometrium and one ovary, as these cancers have a good prognosis if the defined criteria are met 7 34 35 36 . According to the WHO 2020 classification 18 and also according to the ESGO/ESTRO/ESP guidelines 10 , conservative management without adjuvant therapy is recommended for these patients 6 34 36 55 56 . On the other hand, in aggressive types of tumours (mainly serous carcinoma), there is a significant difference in patient prognosis between tumours limited to the polyp and those affecting the adjacent endometrium 57 . The creation of new Stages IC and IIC ( Table 1 ) is therefore justified from this perspective. Additionally, the extension of the definitions of Stages II, III and IV with more precise differentiation of different types of tumour spread outside the uterus, and thus more precise prognosis of individual sub-stages has been generally well received by the professional community. Another development well received by oncologists and pathologists was the differentiation of nodal involvement into micro- and macrometastases in accordance with the AJCC 16 40 . Various techniques of sentinel node detection with detailed histological examination using ultrastaging are now standard in Central Europe 58 59 60 .

Endometrial cancer is the sixth most common malignancy in women worldwide and the most common gynaecological cancer in Europe, with a continuous increase in incidence 1 . The lifetime risk of endometrial cancer is close to 3%; in patients with Lynch syndrome, it reaches 40–60% 2 3 4 . In June 2023, the FIGO (Fédération Internationale de Gynécologie et d’Obstétrique, The International Federation of Gynaecology and Obstetrics) Women’s Cancer Committee officially introduced a new updated staging system for endometrial cancer, replacing the last version from 2009 5 6 . The international team was chaired by Professor Nicole Concin from ESGO (European Society of Gynaecological Oncology), Professor Carien L Creutzberg from ESTRO (European Society for Radiotherapy and Oncology) and Professor Xavier Matias-Guiu from ESP (European Society of Pathology) 6 . The new staging system differs significantly from the previous versions, as it includes other histopathological parameters (histological type of tumour, the presence of substantial lymphovascular space involvement) and molecular classification of the tumour in the definition of individual stages, in addition to the traditional anatomical extent of the tumour. The FIGO 2009 and FIGO 2023 staging systems are shown in Table 1 . The staging system for carcinosarcoma remains identical to that for endometrial cancer.