Latest updates in the histopathological and molecular approach of endometrial carcinoma

Endometrial carcinomas are one of the most frequent carcinomas of the female genital tract. This article aims to highlight the particularities of both histopathological, immunohistochemical and molecular aspects, according to the new World Health Organization (WHO) Classification of endometrial tumors. These tumors can be divided into 4 groups, depending on the mutation that is present, which impacts the prognosis and consequently also the treatment of these patients. Tumors with the best prognosis are part of the low-risk category and are those harboring the POLE mutation, while on the opposite spectrum, TP53 mutated tumors are known to have an aggressive evolution. A thorough literature search has been performed in order to extract the up-to-date information that has been furtherly synthesized. The articles obtained have been reviewed, correlating the histological and molecular features with treatment and prognosis. Although no known reliable immunohistochemical stain can identify the presence of the POLE mutation, molecular analysis is necessary in order to further classify these tumors and to identify the cases that need to be put in the high-risk or high-intermediate risk categories. Considering the good prognosis of carcinomas with POLE mutations, the identification of these cases is recommended due to the possible descaling of treatment.


INTRODUCTION
During the last decade, extraordinary progresses have been made in the field of endometrial carcinoma, especially regarding the molecular classification. In comparison to the previous edition, the current World Health Organization classification of tumors of the female genital tract has introduced a new diagnostic algorithm by which a specific molecular type can be attributed to each carcinoma using an integrated histomolecular classification [1]. According to this algorithm, endometrial carcinomas can be divided into: Polymerase Epsilon (POLE) ultramutated, MMR deficient, p53 mutated, or without any specific molecular alterations [1][2][3]. This classification cannot be predicted by histopathological means, but it has great impact on the prognosis and overall survival rate, thus being relevant in the current practice for both pathologists and clinicians.

HISTOLOGIC APPROACH OF ENDOMETRIAL CARCINOMA
Endometrial carcinoma has several tumoral subtypes, each with multiple histologic variants. For example, endometrioid carcinoma can have squamous differentiation, mucinous differentiation, tubal differentiation, secretory change, villoglandular pattern, small nonvillous papillae, sertoliform pattern or dedifferentiated areas. Other rare variants include: spindle cells, corded, hyalinized stroma, signet ring cells, trophoblastic differentiation, hepatoid differentiation or giant cells. Although extremely heterogeneous microscopically, the most important aspect in clinical practice is the distinction between endometrioid and non-endometrioid endometrial carcinomas.
Bokhman et al. have observed two distinct pathogenic types of endometrial carcinoma, which, to date, are still used in some pathology departments, especially in those where immunohistochemical tests or molecular investigations are not available [4]. Type I EC occurred in women with comorbidities (obesity, hyperlipidemia), or who have risk factors for hyperestrogenism. Upon histologic examination, these tumors are low-grade and have an endometrioid carcinoma histology [2]. Type II EC represents a broader category that incorporates non-endometrioid carcinomas, especially serous carcinoma, and clear cell carcinoma. These tumors have a more aggressive evolution and usually occur in women in the 6 th decade of life [2,5].
A more practical approach of endometrial carcinoma, especially on curettage biopsies, is represented by the evaluation of the predominating pattern of the neoplastic proliferation. Carcinomas with a glandular pattern of growth are most often of endometrioid type, a diagnosis that can be supported by the low-grade morphology of the cells. The most important differential diagnosis is represented by a serous carcinoma with a pseudo-glandular pattern of growth. The most reliable histological feature that can help distinguish between the two entities is the high-grade nuclei that are normally encountered in the latter [6]. Additionally, the expression of a mutational p53 pattern of staining can support the diagnosis. The expression in serous carcinoma can be either: block diffuse, diffuse cytoplasmic or loss of expression [7].
A papillary pattern of growth is usually seen in both clear cell carcinoma and serous carcinoma. However, the character of the papillae is different between the two entities: serous carcinomas have larger papillae lined by multiple layers of cells, while clear cell carcinomas have smaller papillae with a hyalinized core [8]. A clear cell-rich pattern can be seen either in clear cell carcinoma, which is characterized by cells with a hobnail morphology, but also in endometrioid carcinoma with secretory changes, or with pale/ clear squamous differentiation [9,10]. Lastly, a squamous pattern can be observed in endometrioid carcinomas, endometrial hyperplasia but also in atypical polypoid adenomyoma. The latter entity can be differentiated by the SATB2 immunohistochemical expression of the squamous morules [11].

IMMUNOHISTOCHEMICAL APPROACH OF ENDOMETRIAL CARCINOMA
Immunohistochemical analysis can be used in two different scenarios: one in which there is a necessity to differentiate between an endometrial carcinoma and an endocervical carcinoma, while the other situation requires uncertainty regarding the histological subtype (e.g., high grade endometrial carcinoma versus serous carcinoma). The first scenario can be easily solved with the following classic panel: Vimentin, ER, monoclonal CEA, p16, the former two being positive in endometrial carcinoma, while the later favors an endocervical adenocarcinoma [12].
In order to differentiate between the subtypes of endometrial carcinoma, one should make use of both histological and immunohistochemical features. Endometrioid carcinoma shows diffuse reactivity for ER and PR, with only focal expression of p53 (wild type pattern of expression). The latter, together with the loss of PTEN, can help in the differential diagnosis with serous carcinoma, which shows a mutational pattern of staining for p53 (cytoplasmatic, strong diffuse or complete absence) and retained PTEN [12,13]. Serous carcinoma also features diffuse expression of p16, IMP3 and only focal, weak expression of ER and PR [1,12]. Clear cell carcinoma can show a similar immunohistochemical expression profile as serous carcinoma, because both can show diffuse or complete absence of p53 and loss of ER. However, HNF1beta and Napsin A are more commonly expressed in clear cell carcinoma [8].

MOLECULAR APPROACH OF ENDOMETRIAL CARCINOMA
According to the Cancer Genoma Atlas, endometrial carcinomas can be divided into four large groups. Group 1 is associated with a good prognosis and is characterized by mutations of the POLE gene [14]. Epidemiologically, studies have showed a lower incidence of this mutation in Chinese patients [15]. From a morphological point of view, these tumors are frequently grade 3 endometrioid carcinoma, with an ambiguous morphology and large front of invasion. Intratumoral or peritumoral lymphocytes, as well as giant cells, can also be observed [16]. This mutation is encountered in approximate-ly 10% of all endometrioid tumors and in 16% of all clear cell carcinomas [17,18]. To date, there are no reliable immunohistochemical stains that can be used in order to predict the presence of the POLE mutation [19]. This group is also called "ultramutated" because the number of alterations is greater than 100 mutations/ mb [16,19]. If the tumors are classified as stage I or II, then the patient can be placed in the low-risk category [16]. However, POLE mutations have also been identified in patients who developed synchronous ovarian and endometrial carcinoma [20]. Leon-Castillo et al. have also identified cases with POLE exonuclease domain mutation and MMRd, which can be separated into two categories. Those cases with a pathogenic POLE exonuclease domain mutation (EDM) and MMRd can be considered as POLE ultramutated due to a 5-year disease free survival rate of 92.3. On the other hand, cases with non-pathogenic POLE mutation have a 5-year recurrence free survival rate of 76,2%, thus being more similar in their behavior to MMRd endometrial carcinoma [21]. Alexa et al. have observed that the patients with POLE mutation have received the second most aggressive adjuvant treatment after cases harboring TP53 mutations. This is most probably because POLE mutated carcinomas have a high grade endometrioid morphology [16].
Group 2 tumors present microsatellite instability, which can be easily identified by immunohistochemical staining for the following markers: MLH1, MSH2, PMS2 and MSH6 [22,23]. Raffone et al. have observed that even a panel made up of PMS2 and MSH6 can show similar results to using all four markers, but considerably reducing the costs of the laboratory [24]. From a morphological point of view, these tumors are frequently grade 3 endometrioid carcinoma, have multiple foci of lymphovascular invasion, low-uterine segment involvement and MELF-type invasion [16]. This mutation can be identified in about 40% of all endometrioid carcinomas [17,25]. Kaur et al. have reported the expression of INI1 and aberrant expression of the MMR proteins in a case of dedifferentiated endometrial carcinoma [26]. This group has a hypermutated state, meaning that there have between 10 and 100 mutations/ mb [16]. The hypermethylation of MLH1 has been known as a key factor in the tumorigenesis of endometrial carcinoma since 1999 [27,28]. The prognosis in the group 2 tumors is intermediate between group 1 and group 4 [2]. Stage IA low grade endometrioid carcinomas with no or only focal lymphovascular invasion can be placed in the low-risk category, while patients with low-grade endometrioid carcinoma -stage IB and high-grade endometrioid carcinoma stage IA need to be placed in the intermediate risk grade [16,29]. Cases of endometrioid carcinoma with diffuse lymphovascular invasion and microsatellite instability are automatically considered high-intermediate risk. High-grade stage IB and stage II endometrioid carcinomas are also placed in this risk category [16]. However, stage III or IV MMR deficient endometrioid carcinomas represent high-risk tumors [10,16]. However, Makker et al. have observed that a treatment with Lenvatinib and pembrolizumab in patients with advanced endometrial carcinoma can show promising results [30].
Group 3 has the lowest mutation burden, a moderate prognosis, and is not characterized by any of the mutations encountered in the other three groups (POLE, P53 or MSI-H) [2,19]. Clinically, these patients tend to have a higher BMI than the other tumor groups [16]. These tumors have less than 10 mutations/ mb [16]. They usually express ER and PR and from a histopathological point of view they represent an endometrioid-type carcinoma [2,16]. Although there are no specifical immunohistochemical stains that can be used to identify these cases, if none of the other available markers reveals a specific type of mutation, one can assume at least that these cases will probably have a better prognosis [19]. Most common mutations identified in this category are the following: PTEN, PIK3CA, PIK3R1, CTNNB1 and ARID1A [16].
Group 4 tumors are characterized by the presence of p53 mutations and have a high copy-number alterations, although the mutational frequency is less than 10 mutations/ mb [2,16]. P53 mutations are encountered in up to 92% of all group 4 cases, and, from a histological point of view, they have a serous or clear cell carcinoma (grade 3) morphology, feature numerous tumoral emboli, with a destructive pattern of invasion [16,31]. Cytologically, there is high cytonuclear atypia, with large, irregular nuclei, relatively frequent giant cells, and occasional hobnailing of nuclei [16,18,32]. Travaglino et al. have observed that the TP53 mutation represents the most frequent mutation also in carcinosarcomas [33]. In order to identify this category, one can use the p53 immunohistochemical stain, which can function as a reliable surrogate marker for the presence of this mutation. Both overexpression and missense mutations are associated with an extremely poor prognosis [16,17,34]. Cases with p53 mutations of any stage, that do not show any residual disease, are automatically placed in the high-risk group, while those with residual disease are placed in the advanced risk group [16,35]. It is important to note that studies have revealed that grade 3 endometrial carcinomas, which harbor both PT53 and MMRd alterations, behave like other MMRd tumors [36]. However, the survival between TP53 mutated grade 3 endometrial carcinomas and serous carcinoma with the same mutation have a similar survival, jus-tifying a similar therapy for both entities [36]. Akiyama et al. have also observed that cases with p53 overexpression, from patients aged 70 years or more, with advanced stage of disease or featuring a non-endometrioid subtype, are associated with a worse prognosis and a poor overall survival [37]. Additionally, these tumors appear not to respond as well to radiotherapy as other endometrial carcinomas that do not share this mutation [37].

CONCLUSIONS
Endometrial carcinoma encompasses a heterogenous group of tumors that can harbor different and even multiple mutations, including TP53, POLE and MSI-H. The morphological aspects do not give away the mutation associated, although a histo-pathologic aspect of serous carcinoma hints towards the presence of p53 mutation. Reliable immunohistochemical stains are available only for group 2 (MMRd carcinomas) and for group 4 (TP53 mutated carcinomas). However, the favorable outcome and better prognosis of cases with POLE EDM mutation require the identification of these cases either through immunohistochemical stains or through molecular analysis. Given the differences in the prognosis, placing the patients into one of the four categories should be a priority in their treatment and follow up. The favorable prognosis can impact the treatment and warrant its descaling for patients with POLE mutation, while patients with TP53 protein mutation need to be considered as high-grade and require a more aggressive treatment.