Predictive biomarkers in endometrial carcinomas: a review of their relevance in daily anatomic pathology

Abstract

Endometrial carcinoma (EC) is the most common gynecologic malignancy in high-income countries, with its incidence and mortality rising globally, particularly in countries undergoing rapid socioeconomic transitions. Over the past decade, the management of EC has shifted towards molecular classification and biomarker-driven therapies. This shift began in 2013 with the discovery of the 4 prognostic and molecular EC subgroups by The Cancer Genome Atlas (TCGA). Following the TCGA discovery, the Proactive Molecular Risk Classifiers for Endometrial Cancer (ProMisE) and the TransPORTEC initiative have provided 4 pragmatic molecular classifiers by combining next-generation sequencing (NGS) and surrogate immunohistochemical markers to TCGA’s categorization: Mismatch Repair Deficient (MMRd); p53-abnormal (p53abn); No Specific Molecular Profile (NSMP); and POLE-mutant (POLEmut). These subgroups not only provide insights into the biological behavior of EC but also have strong clinical relevance and prognostic implications. In line with these advancements, the World Health Organization (WHO) endorsed molecular classification in 2020, advocating for its integration into EC pathology reports. In 2023, the Federation of International Gynecology and Obstetrics (FIGO) proposed a new staging system that integrates histological parameters and molecular profiles into routine pathology practice. The use of complete molecular classification surrogates in all EC cases, whenever resources permit, is intended to promote prognostic risk-group stratification, data collection, Lynch Syndrome (LS) screening, and potentially influence adjuvant and systemic treatment decisions, as well as predict the efficacy of Immune Checkpoint Inhibitors (ICI). Consequently, the evolving landscape of predictive biomarkers in EC has gained importance in daily oncology practice, profoundly changing the role of pathologists. Pathologists are now crucial in interpreting molecular information and participating as members of multidisciplinary teams in therapeutic decisions. This review article aims to emphasize the importance of molecular classification in EC and encourage pathologists to become familiar with the use of predictive biomarkers in their daily anatomical pathology practice.

Introduction

Endometrial carcinomas (ECs) rank as the sixth most common neoplasm in women worldwide and the second most commonly diagnosed cancer of the female genital organs, with 420,242 new cases reported in 2022 (Bray et al. 2024). Both incidence and mortality rates are rising globally, particularly in countries undergoing rapid socioeconomic transitions (Lortet-Tieulent et al. 2018). Increased risk factors include exposure to higher total concentrations of estrogens, such as earlier age at menarche, later age at menopause, nulliparity, obesity, tamoxifen use, polycystic ovary syndrome, and estrogen-producing ovarian tumors (Ali 2014).

The traditional categorization of EC, based on the Bokhman classification (endometrioid-type and serous-type), has been central to risk stratification and guiding clinical decisions for decades (Alexa et al. 2021). Bokhman’s dual-stratified pathway divided ECs into type I and type II tumors based on pathological, clinical, and epidemiological features (Bokhman 1983). Type I tumors, such as endometrioid histotypes, account for over 70% of EC cases and are associated with high estrogen levels and favorable outcomes. In contrast, type II tumors make up approximately 10% of EC cases, including serous and clear cell types, which are not associated with high estrogen levels and have a worse prognosis, with recurrence rates exceeding 50% even in early stages (Inoue et al. 2021).

In 2013, the results from the TCGA project were published, marking a paradigm shift by describing four molecular subgroups of ECs with distinct clinical, pathological, and molecular features (Cancer Genome Atlas Research Network et al. 2013). The TCGA performed an integrated genomic, transcriptomic, and proteomic characterization of ECs and elegantly demonstrated the molecular diversity of EC based on mutational burden and somatic copy number variations. Subsequently, several large studies have confirmed the strong clinical and prognostic relevance of these molecular subgroups, leading to their incorporation into the management of EC (Vermij et al. 2020).

Following the TCGA discovery, the ProMisE and the TransPORTEC consortium—two research groups—independently developed a pragmatic molecular classification by combining next-generation sequencing (NGS) for detection of pathogenic POLE mutations and surrogate immunohistochemical markers to TCGA’s categorization: Mismatch Repair Deficient (MMRd); p53-abnormal (p53abn); No Specific Molecular Profile (NSMP); and POLE-mutant (POLEmut). In contrast to TCGA methods, which rely on freshly frozen tissue and require costly and complex methodologies, these research groups replicated the four TCGA molecular subgroups using lower-cost, clinically available diagnostic testing methods applied to formalin-fixed, paraffin-embedded (FFPE) tissues (Kommoss et al. 2018; Stelloo et al. 2016).

The development of these pragmatic classifiers with strong prognostic and clinical significance prompted a revision of the 2020 WHO Classification of Female Genital Tumours, 5th edition (Léon-Castillo 2023; Concin et al. 2021). The 2020 WHO Classification endorsed incorporating well-established molecular parameters into endometrioid endometrial carcinomas (EECs) as relevant prognostic information, encouraging their adoption through biomarkers such as targeted POLE sequencing, MSH6, PMS2, and p53 immunohistochemistry (IHC) (WHO Classification of Tumours Series, 5th ed.; vol. 4). The availability of surrogate markers for all molecular subtypes, which can be easily applied to FFPE tissues, facilitates the identification of subgroups analogous to those previously described by TCGA (Léon-Castillo, 2023; Concin et al., 2021).

Following the 2020 WHO classification, the Federation of International Gynecology and Obstetrics (FIGO) proposed a new staging system in 2023 (Berek et al. 2023). The updated system focuses on diagnostic parameters such as histological type and grade, lymphovascular space invasion (LVSI), and molecular alterations, to better reflect the improved understanding of the complex nature of the various types of endometrial carcinoma and their underlying biological behavior (Berek et al., 2023).

The rapid advancement in molecular biological technologies has enhanced our understanding of the multistage process of endometrial carcinogenesis, extending beyond the knowledge of mutations in single oncogenes and tumor suppressor genes (Banno et al. 2014). Over the last two decades, research has increasingly focused on epigenetic mechanisms, such as DNA methylation, histone modification, and non-coding RNAs (Banno et al., 2014). Understanding the genetic and epigenetic events involved in endometrial carcinogenesis has become integral to advances in biomarker-driven cancer therapies (Eskander et al. 2018).

Tumor biomarkers, which are substances produced by tumors or by the body’s response to tumors during tumorigenesis and progression, have demonstrated critical and promising value in screening and early diagnosis, prognosis prediction, recurrence detection, and monitoring therapeutic efficacy (Zhou et al. 2024). Over the past decades, continuous efforts have been made to discover novel and cost-effective biomarkers for application in clinical cancer management. Biomarker-based anticancer targeted therapies have advanced significantly, promoting personalized medicine and improving outcomes for cancer patients (Zhou et al., 2024).

The aim of this review is to provide a better understanding of new approaches to EC management, reinforce the importance of molecular classification, and encourage pathologists to become familiar with the use of key predictive biomarkers in daily anatomical pathology practice.

An overview of molecular classification in clinical decisions

Since the publication of the EC TCGA in 2013, significant attention has been given to incorporating molecular classification into treatment decision algorithms (Cancer Genome Atlas Research Network et al. 2013; Concin et al. 2021). Following the ProMisE approach, which is more pragmatic, many retrospective and prospective cohorts have confirmed the prognostic value of molecular classification (Talhouk et al. 2015). Although there are no randomized clinical trials, some studies have demonstrated that molecular classification can indeed predict the benefits of escalating or de-escalating adjuvant treatment. In the PORTEC-3 trial, external beam radiotherapy was compared to a combination of chemoradiation followed by four cycles of carboplatin and paclitaxel in a population of high intermediate/high-risk patients (De Boer et al. 2018). The study achieved its primary endpoint, showing an increase in progression-free and overall survival with the experimental arm. A special retrospective subgroup analysis was performed after applying molecular classification to the entire patient cohort. In p53 aberrant tumors, a clear benefit was observed with the addition of chemotherapy to radiotherapy. Conversely, patients with POLE mutations had excellent overall survival regardless of the treatment received (León-Castillo et al. 2020). Based on these findings, in 2020, the European Societies of Gynecological Oncology (ESGO), Radiotherapy (ESTRO), and Pathology (ESP) recommended, for the first time, the inclusion of molecular classification to guide treatment decisions when the molecular classification is known (Concin et al., 2021). This recommendation has been endorsed by many international guidelines (Oaknin et al. 2022). Another important retrospective analysis was conducted in the PORTEC-1 and PORTEC-2 studies (Creutzberg et al. 2023). After applying molecular classification, p53 aberrant tumors benefited from external beam radiation, NSMP tumors from brachytherapy, and POLE tumors did not appear to benefit from any form of adjuvant radiotherapy. The prospective RAINBO study will further elucidate how molecular classification can be incorporated. In this study, patients with POLE mutations will be observed, p53 aberrant patients will be randomized to receive olaparib or placebo maintenance after chemotherapy, non-specific molecular profile patients will receive hormonal therapy, and mismatch repair patients will receive immunotherapy (RAINBO Research Consortium 2022).

In the palliative setting, understanding molecular classification has also enhanced patient management. For patients who progress on platinum-based therapies, the standard of care includes the combination of pembrolizumab and lenvatinib for those with mismatch repair proficient tumors (MMRp), and anti-PD-L1 monotherapy (pembrolizumab or dostarlimab) for mismatch repair deficient tumors (MMRd), based on the results of the Keynote 775, Keynote 158, and GARNET trials, respectively (Makker et al. 2023; O’Malley et al. 2022; Oaknin et al. 2022). With the success of immunotherapy in the recurrent setting, four recent randomized clinical trials have also demonstrated the benefit of adding anti-PD-1 (dostarlimab or pembrolizumab) and anti-PD-L1 (durvalumab or atezolizumab) therapies in terms of progression-free survival (Mirza et al. 2023; Eskander et al. 2023; Colombo et al. 2024; Westin et al. 2024). The benefit was observed in the overall population, but the reduction in progression or death was particularly significant in the MMRd cohort.

An overview of molecular classification in pathological practice

The decision to apply molecular classification to all ECs, especially those that are high-grade or high-risk, depends on resource availability and the decisions made by the multidisciplinary team at each center (Concin et al. 2021). A complete molecular classification surrogate (POLEmut, MMRd, NSMP, p53abn) aids in prognostic risk-group stratification, data collection, LS screening, and serves as potential factors influencing adjuvant and systemic treatment decisions, as well as for predictive value regarding ICI (Addante et al. 2024; Chacon et al. 2024).

In early-stage ECs (stages I and II), the presence of POLE mutations (POLEmut) or p53 mutations (p53mut) can influence the FIGO stage based on surgical, anatomical, and histological findings. In contrast, advanced stages III and IV of ECs are not typically modified by additional molecular analyses. POLEmut ECs are rare at advanced stages (Yasuda 2024). Irrespective of the data supporting the usefulness of identifying POLE mutations in ECs, this sequencing technique is not widely available, and its relatively high cost may constrain its routine use in pathology (Yasuda, 2024). Due to the limitations of POLE mutation analysis, it is suggested that morphological screening of particularly ambiguous histological features and rare bizarre nuclei can be useful for selectively enriching ECs for POLE mutation analysis (Keyhanian et al. 2024). Recently, it has been shown that the presence of morules or nuclear expression of β-catenin can effectively rule out the POLE mutation, which may help in deciding when to withhold further POLE mutation testing based on morphological aspects (Fan et al. 2024).

WHO Recommendations

The WHO encourages the use of complete molecular classification surrogate for all ECs, whenever resources allow (WHO Classification of Tumours Series, 5th ed.; vol. 4, 2020). Molecular testing for EC, including mutational analysis for POLE and surrogate IHC for MMR proteins and p53, is particularly recommended for high-grade endometrioid endometrial cancers (EECs) and intermediate-high risk ECs (Addante et al. 2024).

British Association of Gynecological Pathologists (BAGP) Recommendations

The BAGP provides an algorithm for implementing molecular classification that closely aligns with the daily practices of gynecological oncologists at their hospital. According to this algorithm, all EC biopsies, regardless of histotype, are tested by IHC to determine mismatch repair (MMR) and p53 status. In contrast to WHO recommendations, POLE next-generation sequencing (NGS) testing is reserved for cases where it is essential for patient care, such as: (1) when POLE testing is indicated due to abnormal MMR and/or p53 IHC results; (2) in Stage I-II non-endometrioid EC or any grade endometrioid EC at stage IA with lymphovascular space invasion (LVSI) or stage IB/stage II regardless of LVSI; (3) when POLE testing is not indicated in Stage III-IV EC unless directed by the multidisciplinary team or based on patient choice (British Association of Gynaecological Pathologists 2022).

Table 1 Comparison of Molecular Classification Algorithms (WHO vs. BAGP).

Table 1 Algorithm for molecular classification in ECs (WHO and BAGP)

Main features of the molecular subtypes

POLEmut group

Frequency and clinicopathologic features

POLE mutations are significantly more frequent in high-grade ECs (12.1%) compared to low-grade ECs (6.2%) (Travaglino et al. 2020). Morphological heterogeneity and marked atypia are common features of POLEmut ECs, which may also present with giant anaplastic cells (Soslow et al. 2019). A relatively high frequency of POLE mutations has been observed in undifferentiated/dedifferentiated carcinoma (UDC/DDC) (12.4%), whereas lower frequencies are found in clear cell carcinoma (CCC) (3.8%) and carcinosarcoma (CS) (5.3%) (Travaglino et al., 2020). Notably, POLE mutations have been identified in a significant proportion of mixed endometrioid-serous carcinomas in young women (16%). These tumors are thought to originate as EECs that secondarily develop a serous morphology with or without p53 mutations. Both high-grade features and p53 mutations result from a high mutational load but have no additional clinical significance (Casanova et al. 2024). A small subset of tumors (3–5%) exhibits more than one molecular feature (POLEmut and p53mut or MMRd and p53mut), and these are referred to as "multiple classifiers" (León-Castillo et al. 2020). Figure 1 illustrates a case with histological features indicative of a high probability of POLEmut EC.

Fig. 1

A A case of mixed EC (endometrioid-serous type) at stage 1 (HE, 4X). B Image showing the high-grade atypia and lymphocytes (TILs). (HE, 10X) (C) Giant anaplastic cells (HE, 40X). D p53 aberrant staining in the serous component (*) and p53 wild type (p53wt) in the endometrioid component (IHC, 10X). E p53wt (IHC, 40X). F p53abn (IHC, 40X)

The mutational load itself induces a strong immune response due to exposure to multiple neoantigens, which is reflected in the lymphocytic infiltration observed in most (79%) POLEmut ECs and may contribute to their favorable prognosis (Van Gool et al. 2015). Given the prominent lymphocytic infiltrates found in most POLEmut ECs, it is reasonable to hypothesize that these tumors may benefit from immunotherapy (Talhouk et al. 2019).

Molecular features: POLE exonuclease domain mutational status as a biomarker

POLE mutations tend to cluster in the exonuclease domain (POLE-exo), especially in the conserved residues 286 to 411 (Billingsley et al. 2016; Bellone et al. 2017; Barbari et al., 2017). Somatic pathogenic variants in the exonuclease domain (ED) of POLE have been identified in 2–8% of colorectal cancer (CRC), 7–15% of endometrial tumors, and more rarely in other tumor types (Barbari et al., 2017).

POLE is critical for DNA proofreading and replication. Located on chromosome 12, the POLE gene encodes the catalytic subunit of DNA polymerase epsilon, which has both DNA polymerase activity (polymerase domain) essential for DNA elongation and 3’-5’ exonuclease proofreading activity (exonuclease domain) responsible for correcting mismatched bases (Yao et al. 2024). Mutations in POLE are associated with hypermutated tumors and a favorable response to ICI therapy. The detection of POLE mutations in tumors has significant implications for treatment decisions. Cancer patients with somatic pathogenic variants in the POLE-ED typically have an excellent prognosis and a robust response to immune checkpoint inhibition, likely due to the strong immune response elicited by the large number of neoantigens produced as a result of hypermutation (Talhouk et al. 2019).

The occurrence of two proofreading-inactivating events in POLE is extremely rare, suggesting that POLE may not act as a classical tumor suppressor gene (Heitzer et al. 2014). Pathogenic variants of POLE (only the POLE hotspot mutations were defined as POLE mutated (Léon-Castillo 2023) can have different functional effects – whose functional studies were carried out mostly e.g. by yeasts. Different POLE mutation sites can lead to different degrees of tumor mutation burden (TMB).

Somatic POLE exonuclease domain mutations (EDMs) are found in sporadic ECs (6–10%), but rarely, germline POLE mutations can occur (0,25 – 4%). Both the germline and the somatic DNA polymerase EDMs can cause an ‘ultramutated’ type of cancer, sometimes leading to over a million base substitutions per tumor (Briggs et al. 2013). POLE-mut ECs are associated with striking CD8 + lymphocytic infiltrate, a gene signature of T-cell infiltration, and marked upregulation of cytotoxic T-cell effector markers. Because of their remarkable mutation burden, POLE proofreading-mutant cancers are predicted to display substantially more antigenic peptides than other tumors, providing a possible explanation about the strong association between cytotoxic lymphocyte infiltration and favorable outcome in multiple cancers (Inge et al. 2015; Howitt et al. 2015; Bourdais et al. 2017).

Germline heterozygous missense mutations affecting the POLE exonuclease domain are associated with Polymerase Proofreading-Associated Polyposis (PPAP), a syndrome that increases the risk for the development of multiple colorectal adenomas and colorectal cancer, which exhibits a high TMB and an immunotherapy response (Allgäuer et al., 2018; Endris et al., 2019). An even more immunogenic phenotype is observed when a heterozygous POLE-ED mutation is associated with a loss-of-function (LoF) genetic event in comparison with EC with single POLE-ED alterations. (Rosa et al. 2020).

Truncating mutations in POLE gene are unlikely to lead to PPAP phenotype, since a successful DNA synthesis must occur before the proofreading activity of Pol- (Heitzer et al. 2014). By itself, the germline frameshift mutation does not confer a genetic predisposition to EC and cannot lead to a mutator phenotype in the tumor, but might contribute to increase the mutational load because only proofreading-deficient Pol- will replicate DNA in these tumor cells. Normally, tumors that harbor the S459F mutation in POLE were found to be microsatellite stable (Shinbrot et al. 2014; Andrianova et al. 2017; Barbari et al., 2017).

Detection of a combination of POLE-ED and LoF POLE mutations could be considered as prognostic or therapeutic marker, even if a minority of tumors with POLE-ED shows LOH or other inactivating mutations that could act as ‘second hits’ (Heitzer et al. 2014), since the absence of a wild-type POLE allele can lead to a high TMB (Rosa et al. 2020). The complex relationship between POLE hotspot mutation and dMMR/MSI-H in EC was fully explored. EC with both POLE hotspot mutation and MSI-H is rare, occurring only in 4.3% of cases. However, the associated TMB is high (median TMB is 339.0 mut/MB), which is consistent with that of cases with both POLE hotspot mutation and MSS. The median TMB of tumors with POLE non-hotspot mutation and MSI is 207.1 mut/MB. The median TMB of EC with POLE non-exonuclease domain mutation and MSI is 48.5 mut/MB (Ma X et al. 2022).

Molecular testing methods: Next Generation Sequencing (NGS)

Although mutational signatures are preferably determined by genomic analysis (Jamieson et al. 2024), such as whole genome sequencing (WGS) and whole exome sequencing (WES)—see Supplementary Material 1 – some authors have identified some mutational signatures that can be extracted from sequencing data derived from a small gene panel (e.g. 63 genes) in tumors that are highly mutated (Van Hoeck et al. 2019).

MMRd group

Frequency and clinicopathologic features

The MMRd group represents approximately 30% of ECs and is characterized by distinct histopathological features, including: (1) origin in the lower uterine segment; (2) endometrioid differentiation; (3) severe nuclear atypia with an undifferentiated component; (4) high mitotic index; (5) high tumor-infiltrating lymphocytes (TILs) and/or peri-tumoral lymphocytes (≥ 40 TILs/10 HPFs); (6) high morphological heterogeneity; (7) substantial LVSI; (8) deeper myometrial invasion; and (9) synchronous ovarian cancer, particularly in clear cell or endometrioid variants (Addante et al. 2024; Yasuda 2024).

Regarding the prevalence of MMRd ECs across different histotypes, UDC/DDC is the most common MMRd subtype (44%), followed by neuroendocrine carcinoma (42.9%), high-grade EEC (39.7%), mixed forms (33.3%), low-grade EEC (24.7%), CCC (9.8%), and CS (7.3%). Only sporadic cases of serous carcinoma and mesonephric-like carcinoma have been reported to show MMR deficiency (Addante et al. 2024; Vermij et al. 2020).

Compared to POLEmut ECs, MMRd ECs seem to be more prognostically affected by clinicopathological variables, although not as much as NSMP ECs. The ESGO-ESTRO-ESP guidelines substratify MMRd ECs into different risk groups based on pathological features, such as the depth of myometrial invasion, LVSI and histotype. It seems that grading does not matter in the MMRd molecular group (Concin et al. 2021). The overall prognosis of MMRd ECs is intermediate (Vermij et al. 2020).

Molecular features: MMRd/MSI status as a biomarker

The MMRd group is characterized by microsatellite instability (MSI) and displays a high mutational rate with frequent insertions and deletions but low copy-number variations. MSI is a condition of genetic hypermutability resulting from a defective DNA mismatch repair process, and the terms MSI and MMRd are often used interchangeably (Addante et al. 2024). The MMR system is one of the DNA repair pathways responsible for recognizing and repairing erroneous insertions, deletions, and mis-incorporations of bases that can occur during DNA replication. Deficiency in the MMR system leads to the accumulation of DNA mismatches in repetitive sequences known as microsatellites. These mismatches integrate into the genetic code as mutations, resulting in high levels of microsatellite instability (MSI-H). This signature arises from primary biallelic defects in genes that regulate DNA MMR, characterized by the loss of function of at least one of the following genes: MLH1, PMS2, MSH2, or MSH6 (Addante et al. 2024; Mendiola et al. 2023).

Sporadic and germline MMRd ECs

MSI can be caused by somatic or germline alterations. Somatic alterations, accounting for 85% of cases, include biallelic epigenetic MLH1 hypermethylation (in about 77% cases of sporadic ECs); downregulation of MMR genes by microRNAs; biallelic mutations; one somatic mutation and LOH; and secondary epigenetic MSH6 silencing induced by neoadjuvant Radiotherapy and Chemotherapy (RT/CHT) (Addante et al. 2024). Approximately 10% of MMRd ECs and 3% of all endometrial ECs are due to LS, a cancer susceptibility syndrome caused by germline mutations in the MMR genes (MLH1, PMS2, MSH2, MSH6) or EPCAM (Léon-Castillo 2023). Indeed, germline mutations in the MMR genes can determine two different types of clinical syndrome: (1) Constitutional mismatch repair deficiency (CMMRD), a rare childhood cancer predisposition syndrome with recessive inheritance, due to a biallelic MMR gene mutation in which MMR defects (occurring in MLH1, PMS2, PMS1, MSH2 or MSH6) are inherited from both parents; (2) LS, an autosomal dominant disorder characterized by the occurrence of multiple cancers, resulting from constitutional germline mutations, affecting the DNA MMR genes MLH1, MSH2, MSH6 and PMS2; constitutional MLH1 hypermethylation; or deletion of the stop codon (3′ end truncating) of the EPCAM gene causing the epigenetic silencing of the neighboring MSH2 (Addante et al. 2024).

The identification of MMR pathogenic variants in germline sequencing is the gold standard for the diagnosis of LS. However, the first step for LS screening in EC is represented by IHC (Addante et al. 2024). Universal tumor screening in endometrial cancer (EC) is increasingly being adopted to identify individuals at risk of LS (Kaya et al. 2024). Table 2 shows an algorithm of Lynch Syndrome Testing.

Table 2 Lynch syndrome testing algorithm in ECs

Molecular testing methods: MMRP IHC, MSI testing and next generation sequencing

EC is one of the types of cancer most commonly associated with the MMRd/MSI-H phenotype. There is currently a strong recommendation to assess the MMR status in EC, and the main guidelines recommend IHC for analyzing this biomarker (Mendiola M et al. 2023; Casey et al. 2021). Compared with others approaches, IHC is widely available and not expensive. The presence of stained non-neoplastic cells in the sample represents an internal positive control that easily identifies false negative results. Another advange of IHC over PCR is that the loss-of-expression pattern provides information on the particular altered gene, guiding later sequencing to rule out LS (Mendiola et al., 2023; Casey et al., 2021).

Identification of ECs with MMRd/MSI-H status has a threefold value: it (1) serves to detect patients at higher risk of presenting a LS, (2) provides prognostic information as a surrogate marker for MMRd EC (once POLE pathogenic mutations have been excluded), and (3) holds predictive value, as patients with MMRd endometrial carcinomas benefit from check-point inhibitor treatment (Léon-Castilho 2023; Mendiola et al. 2023).

In order to classify MMRd ECs, either IHC or DNA-based methods, such as panels of microsatellite markers for detecting MSI, can be used (Léon-Castilho 2023). Since MSI typically results from a loss of function in the MMR system, leading to decreased corresponding protein expression, IHC for MMR proteins can serve as a surrogate test to identify the MMRd group (Addante et al. 2024). MMR IHC for MLH1, PMS2, MSH2, and MSH6 is the gold standard surrogate testing method and the MMRd ECs will have loss of nuclear expression of one or more of these four key MMR proteins, while the retention of expression of these four markers is known as MMR proficiency (Addante et al. 2024). Tumors with microsatellite stability (MSS) harbor an intact MMR machinery, which is call by the terminology “MMR proficiency” (MMRp). Dual loss of MLH1 and PMS2 is the most frequent pattern observed in daily IHC results (Addante et al. 2024). Figure 2 illustrates an example of dual loss of MLH1 and PMS2, in a case of low-grade endometrioid EC (likely sporadic tumor) and Fig. 3 shows a case of mixed carcinoma with isolated loss of MSH6 (likely germline tumor).

Fig. 2

Low grade EC. A Exophytic tumor located n the fundus of the endometrial cavity. B Endometrioid histology (HE, 40X). C MLH1: loss of nuclear expression. Postive internal control (IHC 40X). D PMS2: loss of nuclear expression. Positive internal control (IHC, 40X). E MSH2: intact nuclear expression (IHC, 40X). F MSH6: intact nuclear expression (IHC 40X)

Fig. 3

A Mixed EC (endometrioid-serous histotype) (HE, 4X). B Endometrioid component (HE, 10X). C Serous solid component (HE, 40X). D p53-abn with nuclear overexperssion in serous component (IHC, 40X). E MSH6: loss of expression in the two tumor components espacially distincts. (*) serous solid pattern (IHC, 10X)

MMR proteins occur as heterodimers in the cell. While MLH1 and MSH2 can stabilize in the cell by forming heterodimers with different partners in the absence of PMS2 and MSH6, respectively, this is not the case for PMS2 and MSH6. As a result, there will always be loss of PMS2 expression in absence of MLH1, and MSH6 will always be lost in absence of MSH2. Alternatively, it is possible to use a two-marker approach (IHC stain for only PMS2 and MSH6) to identify MMRd ECs (Léon-Castilho 2023).

The accordance between MMR IHC and MSI analysis is high (93–95%), but discrepancies can occur due to (1) subclonal MMR deficiency (a well-delimited area of the tumor with loss of expression of a major MMR protein can be overlooked by using MSI analysis), (2) ECs with MSH6 mutations that may appear as microsatellite stable (MSS), and (3) incorrect interpretation of MSI assay results, which requires highly trained personnel (Léon-Castilho 2023).

However, there are specific conditions requiring PCR or NGS instead of MMR protein IHC, namely: (1) Whenever IHC shows indetermined/ambiguous/equivocal results; (2) False negative IHC results due to pre-analytical tissue poor fixation; (3) Whenever IHC shows aberrant patterns (e.g., cytoplasmic, dot-like and perinuclear staining); (4) Whenever IHC shows the loss of only one heterodimer subunit, i.e., only MLH1 or PMS2 and not both (Addante et al. 2024).

Polymerase chain reaction (PCR) represents the first-line molecular analysis for MSI testing, also performed on FFPE tissue slides (Walsh et al. 2023). MLH1 promoter methylation testing, also performed on FFPE tissue slides, is another relevant molecular testing technique used in cases of absent MLH1 or absent MLH1 and PMS2 IHC. Most laboratories utilize a methylation-specific real-time PCR assay to determine the presence of methylation. It is used in cases of absent MLH1 or absent MLH1 and PMS2 IHC to differentiate between somatic loss due to promoter methylation and potential germline loss. Defective MMR in sporadic endometrial cancer is most often due to inactivation of the MLH1 gene promoter by methylation, by epigenetic silencing (Walsh et al., 2023).

NGS represents another type of molecular test that allows efficient sequencing of the entire genome, also performed on FFPE tissue slides. NGS has the advantage of allowing coupling MSI analysis with the determination of tumor mutational burden (Walsh et al. 2023).

MSI, PD-1/PD-L1, TMB and immunotherapy in EC

Concerning therapeutic options, MMRd ECs are considered optimal candidates for immunotherapy due to their high mutation load and rich immune infiltrate (Addante et al. 2024). All MMRd tumors exhibit high microsatellite instability (MSI-H), with a rapid accumulation of genomic mutations and high mutational load (H-TMB: 10–100 mutations per megabase). These tumors harbor hundreds to thousands of mutations and are highly immunogenic, with strong expression of immune checkpoints and high levels of lymphocyte infiltration. Cells from MMRd tumors may express programmed death ligand 1 (PD-L1) on their membranes, and infiltrating lymphocytes display upregulated checkpoint proteins, including programmed death 1 (PD-1), cytotoxic T-lymphocyte-associated protein 4 (CTLA-4), and lymphocyte-activation gene 3 (LAG-3) (Marabelle et al. 2020). Immune cell infiltration may result from the high number of mutations found in MMRd tumors, specifically frameshift mutations, which produce mutant protein neoantigens (Addante et al. 2024).

The Phase II KEYNOTE-158 study demonstrated the clinical benefit of anti–programmed death-1 (PD-1) therapy with pembrolizumab in patients with previously treated unresectable or metastatic MMRd colorectal cancer, showing a response rate of 48% and a median progression-free survival of 13.1 months. Other studies with immunotherapies such as dostarlimab, durvalumab, and avelumab in similar scenarios have also shown good clinical benefits. Pembrolizumab is a humanized immunoglobulin G4 monoclonal antibody that binds to the inhibitory immune checkpoint receptor PD-1 expressed on lymphocytes, blocking the binding of its ligands PD-L1 and PD-L2, thereby allowing reactivation of T-cell‒mediated tumor destruction (Marabelle et al., 2020). PD-L1 expression by IHC may predict the response to anti-PD-1/PD-L1 monoclonal antibodies (Addante et al., 2024; Mendiola et al. 2023). Given that MMRd/MSI status may serve as a suitable biomarker for response to PD-1/PD-L1 immunotherapy with pembrolizumab (Marabelle et al. 2020), ESGO/ESTRO/ESP guidelines recommend IHC for MMR proteins in all EC cases as a universal screening test to assess MMRd/MSI status and identify patients who could benefit from immunotherapy (Concin et al. 2021).

Tumor mutational burden (TMB) is a biomarker defined as the total number of somatic mutations per coding area of a tumor genome, obtained through NGS. Tumors with high TMB (TMB-H) often harbor elevated levels of neoantigens, making them targets for activated immune cells and potentially leading to superior and/or prolonged responses to ICIs, including anti-PD-1 agents. Cancers with > 10 mutations/Mb are referred to as TMB-H (Wash et al. 2023). POLE-mutant ECs and MMRd ECs are examples of TMB-H tumors, characterized by immunogenic microenvironments (Marabelle et al. 2020). A recently published biomarker analysis from KEYNOTE-158 reported that tumors with TMB-H status had a higher objective response rate to pembrolizumab monotherapy compared to non-TMB-H tumors, leading to the FDA approval of pembrolizumab for TMB-H solid tumors (Marabelle et al. 2020). TMB is a useful biomarker for selecting EC patients for ICI therapy (Walsh et al. ; Marabelle et al. 2020).

p53-abnormal (p53abn) group

Frequency, clinicopathologic and molecular features

Somatic mutations of the TP53 gene on chromosome 17p13.1 are present in up to 28% of all endometrial cancers (ECs) (Jamieson et al. 2021). Given that p53-abnormal (p53-abn) ECs represent the most aggressive molecular subtype with a poor prognosis, TP53 mutation status is a suitable biomarker for predicting unfavorable outcomes in EC patients (Jamieson et al. 2021; Arciuolo et al. 2022).

This group is characterized by a low mutation rate (TMB < 10 mutations per megabase) and high somatic copy-number alterations, with TP53 mutations present in 90% of cases (Jamieson et al. 2021). The majority of p53-abn ECs are high-grade serous carcinomas, along with other histotypes such as carcinosarcomas, mixed ECs, and clear-cell carcinomas (Jamieson et al., 2021). These other histotypes exhibit similarly aggressive biological behavior, resulting in the worst outcomes (Jamieson et al., 2021). Consequently, ESGO-ESTRO-ESP guidelines classify all p53-abn ECs as high-risk, except for non-myoinvasive cases (Arciuolo et al. 2022).

Molecular testing methods

The assessment of p53 by IHC has a high concordance with TP53 mutational status, making it a suitable surrogate method for TP53 mutation detection (Köbel et al. 2019; Köbel et al. 2021). Although p53 IHC is not a perfect substitute for copy-number analysis, its accuracy in routine practice remains adequate. Since 2020, the WHO has endorsed the inclusion of p53 IHC in diagnostic algorithms due to its accessibility, simplicity, and cost-effectiveness. However, the correct interpretation of p53 IHC by pathologists is crucial for accurate diagnostic reports, and the terminology used must be clearly understood by the multidisciplinary team (Arciuolo et al. 2022).

In a pathologist’s daily routine, the majority of TP53-mutant cases shows p53 overexpression on IHC, defined by a diffuse and uniformly strong nuclear staining in 80–100% of tumor cells. This aberrant pattern is commonly linked to missense mutations in the DNA-binding domain of TP53 (Vermij et al., 2022). Others aberrant patterns include complete loss of p53 nuclear expression (with a positive internal control) and unequivocal cytoplasmic expression (Jamieson et al. 2021). The "null" pattern often results from frameshift or nonsense mutations encoding truncated p53 protein, and the cytoplasmic overexpression is caused by mutations in the tetramerization or C-terminal domain of TP53. Köbel et al. (Köbel et al. 2019 and 2021) have described a heterogenous pattern ("subclonal"), where a distinct geographic area of the tumor exhibits both a abnormal p53 staining and a wild-type p53 expression (Vermij et al., 2022). Subclonal patterns are frequently associated with MMRd or POLEmut, where p53 abnormalities lack prognostic value (Arciuolo et al. 2022). Another subset of TP53-mutant tumors (~ 5%) harbor truncated TP53 mutations that do not manifest abnormal p53 expression on IHC (Arciuolo et al., 2022). Additionally, some copy number-high EC cases, defined solely by molecular analysis of copy-number variations, do not exhibit TP53 mutations (Arciuolo et al. 2022). Figure 4 illustrates three examples of abnormal p53 IHC patterns, considered indicative of mutation-type staining (overexpression, null-type and citoplasmatic stain). Figure 5 shows two cases of low-grade endometrioid tumors (same histology), with different p53 status (abnormal/mutated type and normal/wild type).

Fig. 4

A Case 1: CCC (HE, 40X). B CCC: NAPSIN A: Positive stain (granular citoplasmatic stain) (IHC, 40X). C CCC: p53 aberrant, nuclear overexpression (IHC, 40X). D Case 2: Serous carcinoma (ESC) (HE, 40X). E ESC: p16 overexpression (IHC, 40X). F ESC: p53 aberrant stain, null type. Positive estromal cells (IHC, 40X). G Case 3: ESC (HE, 40X). H ESC: p53 aberrant, nuclear and citoplasmatic stain (IHC, 40X)

Fig. 5

Two cases of low grade endometrioid tumor. A Case 1: G2 endometrioid tumor (HE, 40X). B Case 1: p53-abn (nuclear overexpression) (IHC, 40X). C Case 1: p16 aberrant, nuclear and citoplasmatic overexpression (IHC, 40X). D Case 2: G1 endometrioid tumor (HE, 40X). E Case 2: p53 wild type (IHC, 40X). F Case 2: p16 mosaic type (IHC, 40X)

Due to the aggressive nature of p53-abnormal (p53-abn) tumors, adjuvant treatment is often recommended for this group. For all myoinvasive cases, chemoradiotherapy is indicated (Arciuolo et al. 2022; Jamieson et al. 2021). Additionally, targeted therapy options may be available for a subset of cases based on specific characteristics such as homologous recombination deficiency (HRD) or human epidermal growth factor receptor 2 (HER2) overexpression/amplification. Patients with these features are potential candidates for HER2 inhibitors or PARP inhibitors (Arciuolo et al., 2022; Jamieson et al., 2021).

No Specific Molecular Profile (NSMP) group

Most endometrial tumors (about 50%) with no molecular signatures are placed in this group. These tumors neither exhibit high mutational burden nor significant copy-number variations, been similar to the TCGA copy-number low group (Arciuolo et al. 2022). NSMP ECs represent a diagnosis of exclusion, assigned based on the absence of MMR deficiency, pathogenic POLE mutations, and p53 abnormalities. This group shows diverse clinicopathological and molecular features and is generally associated with intermediate patient outcomes, with strikingly different outcomes at the individual level (Aro et al. 2024). NSMP ECs encompass young women with low-grade indolent ECs arising in settings of obesity/excess estrogen and thin older women with rare histotypes historically categorized as high-risk and with unfavorable disease trajectories. A study showed that after the removal of high-risk NSMP cases, the low-grade (grade 1 or 2) and ER-positive endometrioid carcinomas, accounting for 84% of the NSMP group and corresponding to prototypical type 1 EC, was associated with a very favorable prognosis (Jamieson et al. 2023). Low ER expression is associated with hightumour grade, lymph node metastasis, deepmyo-invasion, high tumour stage, increased risk of disease recurrence, decreased overall survival and decreased disease-specific survival (Alafraidi et al. 2024). It seems that NSMP non-endometrioid tumors have a unfavored prognosis (like p53-abnormal ECs), while NSMP endometrioid ECs demonstrate a highly heterogeneous prognosis, ranging from favorable (like POLE-mutant ECs) to unfavorable (like p53-abnormal ECs) (Arciuolo et al., 2022). There is a suggestion that ER IHC can serve as a biomarker for stratification of NSMP carcinomas into two clinically distinct groups: low-risk carcinomas with high ER expression and an excellent prognosis similar to that of POLEmut endometrial carcinoma, and high-risk carcinomas with low ER expression (Alafraidi et al., 2024).

A model of sub-stratifying is currently being evaluated in the PORTEC-4a study (Van den Heerik et al. 2020), which integrates TCGA-based molecular subgroups with L1-CAM overexpression, substantial LVSI, and CTNNB1-exon 3 mutations. Such sub-stratification appears to be more applicable to the NSMP tumors (Arciuolo et al. 2022; Van den Heerik et al. 2020). This trial global study applies a molecular-integrated risk profile to the high-intermediate risk cohorts of the PORTEC-1 and -2 trials, utilizing L1CAM and CTNNB1 and delineating three risk categories (favorable, intermediate, and unfavorable) with significantly different recurrence-free survival rates. L1CAM serves as an independent risk factor for locoregional and distant spread and is associated with TP53 mutations, high tumor grade, and LVSI. L1CAM is a crucial indicator of high-risk disease. L1CAM expression is most frequent in p53-abnormal (p53-abn) tumors but also predicts worse outcomes among tumors with NSMP (Oaknin et al. 2022). CTNNB1-exon 3 mutations prognosticate a higher risk of recurrence, particularly in the copy-number-low group (Van den Heerik et al. 2020; Wash et al. 2023). In summary, L1CAM overexpression and CTNNB1 mutations appear to be significant prognostic biomarkers with clinical implications for deciding adjuvant treatment in patients with EC (Van den Heerik et al., 2020). ER positivity and L1CAM negativity are strongly associated with low grade, which limits the ability of these biomarkers to further improve prognostication in these patients. Notably, risk assessment for high-risk–advanced-metastatic NSMP and MMRd subtype carcinomas can be refined by ER status (Aro et al. 2024).

Tumors with multiple classifying alterations

Leon-Castillo et al. reported that approximately 3–5% of ECs are classified as “multiple classifiers,” exhibiting more than one molecular signature. They demonstrated that double classifiers such as MMRd/p53abn tumors behave like MMRd EC and POLEmut/p53abn behave like POLEmut EC in terms of morphology, prognosis and clinical behavior, and not like p53abn ECs (León-Castillo et al. 2020). The multiple classifiers represent a rare proportion (3–10%) of molecularly classified ECs from various previously published studies, such as the ProMisE confirmation and TransPORTEC early-stage EC cohorts, and efforts have been directed towards understanding their prognostic implications (León-Castillo et al. 2020; De Vitis et al. 2024). The most frequent subgroup was MMRd-p53abn, followed by POLEmut-MMRd, POLEmut-p53abn, and POLEmut-MMRd-p53abn (De Vitis et al., 2024). Indeed, studies indicate that POLE mutations demonstrate the most profound prognostic significance, surpassing both MMRd and TP53 mutations. Following POLE mutations, MMRd exhibits a lesser prognostic impact, followed by TP53 abnormalities. Therefore, both prognosis and genomic data strongly support that POLEmut trumps both MMRd and p53abn, and MMRd trumps p53abn (León-Castillo et al., 2020). These findings endorse the prioritization of POLE sequencing over MMR protein and p53 IHC, in the algorithms of molecular classification proposed in clinical oncology guidelines (León-Castillo et al., 2020).

Given the high mutational burden observed in POLEmut and MMRd tumours, secondary TP53 mutations/p53 IHC abnormalities can be found. These findings suggest the TP53 mutation is a later event during tumour progression in POLEmut and MMRd tumours and does not affect the clinical outcome. This also highlights the importance of interpreting p53 and MMR IHC in the context of POLE mutation status to avoid overtreatment in these patients with ‘multiple classifiers’. The 2020 ESGO/ESTRO/ESP guidelines for the management of patients with EC classify stage I-II POLEmut EC as low-risk and omission of adjuvant therapy should be considered (Concin et al. 2021). Information on p53abn status alone is not enough on its own to guide treatment decision-making as a proportion of these are double-classifiers with significant prognostic and predictive implicantions (León-Castillo et al. 2020). Table 3 shows subgroups of double classifiers and Table 4 shows a summary of the four-histomolecular groups, including their prognosis and clinical relevance.

Table 3 Multiple classifiers

Table 4 Histomolecular classification of ECs

Other selected biomarkers in EC

Human epidermal growth factor receptor 2 (HER2, HER2/Neu or ERBB2)

HER2/Neu, the human epidermal growth factor receptor 2, also called ERBB2, has been established as an important biomarker with both prognostic and therapeutic implications in breast and gastric cancers (Pina et al. 2024). Endometrial serous carcinoma (ESC) is an aggressive, high grade endometrial cancer subtype associated with poor clinical outcomes and significant mortality. Approximately 30% of ESC overexpress HER2 (Fader et al. 2020). Figure 6 illustrates an example of ESC with 3 + HER2 positivity.

Fig. 6

A SEC (HE, 40X). B p16: nuclear and citoplasmatic overexpression (IHC, 40X). C p53-abn: nuclear overexpression (IHC, 40X). D HER2: intense and complete membranous expression score 3 + (IHC, 40X)

ESC is typically treated with hysterectomy and surgical staging followed by platinum/taxane combination chemotherapy. Despite overexpression or amplification of HER2 in this subtype of cancer carries a poor prognosis (Ferriss et al. 2021), anti-HER2 therapy has emerged as an effective targeted treatment approach for patients with advanced stage and recurrent ESC, resulting in significantly prolonged progression-free and overall survival when combined with the standard chemotherapy regimen (Fader et al. 2020).

Most cases of HER2 overexpression are caused by HER2 gene amplification and can be detected either by IHC for the gene product, by fluorescent in situ hybridization (FISH) to determine the gene copy number, and by NGS (McNamara et al. 2023).

Tumors with intense complete or lateral/basolateral membranous HER2 immunostaining in more than 30% of tumor cells are considered positive and assigned a 3 + score, 2 + score, equivocal, is assigned when intense complete or lateral/ basolateral membrane staining was seen in 30% or less of tumor cells, or weak to moderate staining in greater than or equal to 10% of tumor cells. 0 score (no staining observed) and 1 + score (incomplete membrane staining that is faint/barely perceptible in any proportion of cells or weak complete staining in less than 10% of tumor cells) are considered negative. FISH is performed only on tumors with a 2 + immunohistochemical score on a large tumor area in direct correlation with the HER2 immunostained slide. HER2/CEP17 ratio of 2.0 or HER2/CEP17 ratio less than 2.0 with average HER2, copy number equal to or greater than 6 per nucleus is considered amplified (Buza 2021; Buza 2022). Table 5 shows the reporting results of HER2 testing by IHC and by FISH (dual-probe assay) for ESC.

Table 5 Reporting results of HER2 testing by IHC and FISH (dual-probe assay) for ESC

HER2 intratumoral heterogeneity by IHC is defined by the presence of at least two-degree difference in staining intensity (0 vs 2 + , 1 + vs 3 + , or 0 vs 3 +) involving at least 5% of tumor cells (Ross et al. 2022). With a tumor harboring both 3 + and 0/1 + HER2 tumor cells, HER2-directed therapy may not work up front or may diminish as HER2-negative tumor cells outgrow HER2-positive tumor cells due to trastuzumab’s selective pressure in HER2 heterogenous tumors over time (Buza 2021). Discordant HER2 status has also been observed between paired primary and metastatic tumors in close to 20% of breast, gastric, and most recently ECs (Buza 2021). HER2 testing of multiple specimens (endometrial biopsy, curetting, hysterectomy, and metastatic foci) is stimulated, as it may increase the rate of HER2 positivity and eligibility for targeted therapy (Rottmann et al. 2021).

Primary results from the DESTINY-PanTumor02 phase II assessed the efficacy of Trastuzumab deruxtecan (T-DXd), using ASCO/College of American Pathology guidelines for scoring HER2 in gastric cancer, and showed benefits observed in the endometrial, cervical, and ovarian cohorts. In the endometrial cohort, 77.5% of patients had ≥ two prior lines of therapy. The objective response rate in patients with HER2 IHC 3 + expression was 84.6%. In all patients with endometrial cancer, median PFS and OS were 11.1 months and 26.0 months, respectively. The clinically significant response and survival rates observed in this study are encouraging for HER2-expressing ECs, which are typically associated with high risk for progression and poor survival rates (Meric-Bernstam et al. 2024). Table 6 shows IHC scoring for HER2 in gastric and gastro-oesophageal junction cancer.

Table 6 IHC scoring for HER2 in gastric and gastro-oesophageal junction cancer

In high copy number/ p53abn group, patients can have mutations in FBXW7, PIK3 and PPP2R1A. PI3K/AKT/mTOR pathway is affected in 50–60% and ERBB2 alterations are observed in 25% of the patients, which can be potential targets for new combinations (Pina et al. 2024). Other notable recurrent genetic alterations included PIK3CA activating mutations (40%) and ERBB3 amplification (10%), both of which have been reported to impact the efficacy of anti-HER2 therapy in other cancer types (Ross et al. 2022). HER2-directed Antibody–Drug Conjugates show substantial promise and stand to provide significant improvements in effectivity and tolerability of cytotoxic therapy (McNamara et al. 2023). The correlation between HER2 positivity and p53abn EC is significantly stronger than the correlation between HER2 positivity and serous histology (Karpel et al. 2023; Vermij et al. 2020). It suggests that subsets of high-grade ECs with aberrant p53 expression may also be HER2 amplified, and these patients could benefit from the addition of targeted therapy (Joehlin-Price et al. 2023, Ross et al., 2022).

Estrogen and progesterone receptors (ER/PR)

Along the years, hormonal therapy has been employed in ER and PR-positive ECs, as viable management option, particularly in cases of recurrent or metastatic disease within the NSMP group, where the majority of endometrioid tumors exhibiting overexpression of ER and PR are placed (Karpel et al. 2023).

The hormone receptor status is determined in FFPE tissue sections using IHC and despite the absence of a consensus regarding the reporting of ER and PR IHC results, the College of American Pathologists (CAP) advocates for a standardized reporting approach in Ecs, akin to protocols employed in breast cancer diagnosis. The assessment involves the proportion of positive tumor cells and the intensity of immunoreactivity, akin to protocols employed in breast cancer diagnosis. Positivity is defined by nuclear staining in at least 1% of tumor cells (Walsh et al. 2023; Longacre et al. 2017).

Conclusion

The rise of molecular classification has profoundly changed the approach to EC management, with the emergence of biomarker-driven therapies guided by diagnostic, predictive, and prognostic biomarkers. The integration of these biomarkers as molecular tools in pathological reports represents a new histomolecular concept for EC. This integration enables clinicians to perform individualized risk stratification based on molecular profiles and to define tailored treatment strategies. This molecular approach is so applicable that expanding clinical trials are in progress, and new molecular biological tools, such as liquid biopsy biomarkers, are expected to emerge. Novel AI-based methods have been applied to enhance the molecular classification of EC, highlighting a broader range of morphological findings that might otherwise be classified as dichotomous by the human eye (Fremond et al. 2023). Additionally, clonal expression of markers or heterogeneous findings could be better integrated through computational imaging tools (Darbandsari et al. 2024). In the era of molecular pathology, pathologists have a new role, integrating molecular knowledge into pathological reports and guiding treatment decisions in collaboration with multidisciplinary teams, which is crucial for the success of precision medicine.