Endometrial carcinoma (EC) is the most common gynaecologic malignancy in the United States with approximately 52,630 diagnosed cases annually [
1]. In the Netherlands the incidence is about 1900 women, with a mortality rate of 480 [
2]. The incidence is still rising due to increased life expectancy and obesity as important risk factor [
3]. Although the majority of patients are diagnosed at an early stage with a favourable prognosis, still around 20 % of patients die from the disease [
4]. ECs are staged according to the 2009 Fédération Internationale de Gynécologie et d’Obstétrique (FIGO) classification. ECs are divided into two types. The majority of ECs are classified as type I and are related to unopposed oestrogenic stimulation resulting from obesity or exogenous hormone use and originate from hyperplastic endometrium. This tumour type is associated with early stage disease, endometrioid histology, and a favourable outcome after surgery [
5]. In contrast, type II carcinomas are unrelated to oestrogenic stimulation and arise in a background of atrophic endometrium. Type II carcinomas are associated with advanced stage, high grade, non-endometrioid histology, and an overall a poor prognosis [
5]. A recent study suggests the existence of a third type of EC characterised by low grade endometrioid endometrial carcinoma (EEC) and a background of atrophic endometrium [
6]. This third type of EC may have a poorer prognosis when compared to type I carcinomas [
6]. However, recently published data of The Cancer Genome Atlas (TCGA) Research Network, identified four subgroups of EC based on molecular classifiers such as
TP53,
PTEN and microsatellite instability [
7]. This supports the need for adjusting the currently used classification.
Primary treatment
Primary treatment is currently based on preoperative risk classification and consists of hysterectomy with bilateral salpingo-oophorectomy. In uterine papillary serous carcinoma (UPSC) and clear cell carcinoma (CCC) a complete surgical staging is mandatory because of the high risk of extra-uterine disease [
8‐
10]. Although the presence of lymph node metastasis is an unfavourable predictor for disease specific survival, data of Kwon et al. demonstrated that high-risk uterine factors including high grade tumour type, deep myometrial invasion, and cervical stromal involvement are more significant determinants of survival in EC than pelvic-node status [
11]. The current study focuses on diagnosis and preoperative risk assessment of patients with EC.
Preoperative diagnosis
During the last decades dilatation and curettage (D&C) has been replaced by minimally invasive techniques for endometrial sampling in an outpatient setting. The amount of tissue obtained from endometrial sampling is relatively small and there can be different subtypes of EC in one tumour, making routine histological discrimination between EEC and a high grade, UPSC or CCC difficult. Moreover, in 30 % the amount of tissue obtained with outpatient endometrial sampling is insufficient for diagnosis [
12]. Previous studies found discrepancy percentages between 15 and 40 %, including both grade and histological subtype [
13‐
17]. When preoperative diagnosis was based on D&C or endometrial sampling, a preoperative diagnosis of grade 1 was concordant with the final diagnosis in 85 % of cases. However, high grade lesions were more frequently underestimated by endometrial sampling compared to D&C [
18].
Immunohistochemical analysis in preoperative endometrial sampling
Identification of a panel of immunohistochemical (IHC) markers may be helpful to establish a reliable preoperative risk classification. A brief summary of the selected markers that will be tested is given in Table
1. P53 immunopositivity is associated with non-endometrioid EC [
19]. Negative IHC for oestrogen and progesterone receptors can predict lymph node metastasis and is associated with decreased survival [
20]. Double negative hormone receptor status and p53 immunopositivity correlates with lymph node metastasis, high FIGO stage, non-endometrioid histology, high grade and poor prognosis [
20]. Insulin-like growth factor II messenger RNA-binding protein 3 (IMP3) is a foetal protein not expressed in normal adult tissues. This oncoprotein plays an important role in tumour growth, migration and invasion. IMP3 could contribute to the preoperative identification of type II tumours, since it is more frequently expressed in UPSC and CCC when compared to EEC (resp. 78 %, 57 % and 15 %) [
21]. A recent study showed that L1CAM is the best predicting prognostic factor in FIGO stage I, type I EC and superior to the standard used multifactor risk score (myometrial invasion, tumour grade and lymph space or vascular invasion) [
22]. L1CAM immunohistochemistry can improve the identification of patients at risk for recurrent disease. However, all the mentioned biomarkers are lacking validation on pre-operative histological samples and are based on singles studies. Further research has to validate these promising results.
Table 1
Immunohistochemical analysis
IMP3 | Insulin-like growth factor II mRNA-binding protein 3 | IMP3 is more frequently expressed in UPSC and CCC than in EEC (resp. 78 %, 57 % and 15 % of the tumours were positive). | |
P53 | | P53 is more expressed in non-endometrioid endometrial carcinomas than in EEC. Expression is also related to higher tumour grade. | |
ER and PR | Oestrogen and progesterone receptor | Negative receptors were associated with lymph node metastasis and decreased survival. ER and PR expression is lower in non-endometrioid endometrial carcinomas than in EEC. | |
MLH1 | MutL homolog 1 | Loss of expression of mismatch repair proteins is seen in high grade EEC and not in UPSC and CCC. Loss of MLH1 expression is associated with longer survival. | |
PTEN | Phosphatase and tensin homologue | PTEN positivity is more frequently found in UPSC than EEC. | |
Beta-catenin | | Positive beta-catenin expression is associated with decreased stage, decreased grade and negative lymph node status | |
P16 | | Loss of p16 expression is significantly correlated with high FIGO stage and serous and clear cell histological subtype. | |
Ki-67 | | Higher Ki-67 expression is associated with higher tumour grade. UPSC and CCC show higher Ki-67 proliferation index than EEC. | |
Stathmin | | Stathmin overexpression was associated with non-endometrioid histology, high grade and poor disease-specific survival. | |
ARID1A | AT-rich interactive domain 1A gene | Loss of ARID1A expression is significantly more frequent in high grade EEC compared to UPSC. | |
L1CAM | L1 cell adhesion molecule | L1CAM is associated with higher grade and non-endometrioid histology. Moreover, L1CAM positive EC have statistical significant poorer disease-free survival and overall survival. | |
Preoperative diagnosis of EC in cervical cytology
The presence of endometrial cells in cervical cytology in postmenopausal women is strongly associated with endometrial pathology [
23]. Abnormal cervical cytology is associated with extra-uterine disease in patients with UPSC and with cervical involvement in patients with EEC [
24]. A combination of preoperative cervical cytology with endometrial sampling might better predict final histology and risk for extended disease. In a study of Kinde et al. DNA was extracted from cervical smears to detect genetic disorders present in EC [
25]. The mutation profile found in the primary tumour was found in all of the cervical smears [
25]. These results indicate that cervical cytology might be a reliable and minimal invasive source of material for detection of EC.
Comorbidity and EC
The impact of comorbidity on cancer outcome has been underestimated for a long time. Recently published data demonstrated that EC patients with cardiovascular disease, previous malignancy and diabetes have a significantly decreased survival of 15–17 % compared to patients without comorbidity [
26]. Additionally, patients with diabetes and EC have more comorbidities, higher body mass index (BMI) and higher FIGO stage, compared to those without diabetes [
27]. There is also a significant increase in the risk of EC-specific mortality among women with diabetes [
28]. Although obesity is a risk factor for development of EC, obesity seems not related to overall survival [
29]. Yet, comorbidity has demonstrated to influence the outcome in EC [
30].
In summary, the main challenging issue concerning clinical management of EC patients is underscored by the discordance between the preoperative risk classification of the tumour and the final surgical pathology. At the moment a subgroup of patients needs either a secondary surgical staging procedure or additional chemotherapy and/or radiation therapy. With the current study we want to select a panel of the most accurate biomarkers that can be used in daily practice for preoperative diagnosis of EC. This will aid in improving the concordance between preoperative and final histological diagnosis and thus prevent over and under treatment. Incorporating cervical cytology and comorbidity could potentially improve a proper risk classification in EC patients.