Premalignant Lesions of the Endometrium

Etiology

Endometrial carcinoma incidence is increasing worldwide as a direct result of an aging population and the rising incidence of obesity. There is a consensus that premalignant lesions develop several years before invasive endometrial carcinoma develops.[1] Atypical EH precedes invasive endometrioid carcinoma (type 1) and that the two share common predisposing factors as follows.[6] Among the non-modifiable risk factors are:

1. The age of the patient (>35 years)
2. Caucasian race

Other predisposing factors may also include a family history of endometrial hyperplasia or carcinoma, features of early menarche, prolonged perimenopause, or late menopause and postmenopausal status. Associated comorbidities that may play a role in predisposing to EH are obesity, diabetes mellitus type 2, and earlier or present diagnosis of granulosa cell tumor of ovary or HNPCC (hereditary non-polyposis colorectal cancer). In addition to this, exposure to tamoxifen therapy, hormone replacement therapy (especially estrogen-only), and any form of exogenous hormonal exposure also increase the risk of developing EH. Environmental factors like smoking and certain genetic mutations may also predispose to the development of EH.

On the other hand, endometrial intraepithelial carcinoma (EIC), the precursor lesion of type 2 ECs, tends to be estrogen-independent and progresses to the clinically aggressive ‘serous’ and ‘clear cell’ histological subtypes. Type 2 ECs are more often associated with endometrial atrophy in the postmenopausal woman as compared to endometrial hyperplasia (EH) seen in type 1 EC.[6]

The tumor progression model suggests that loss of PTEN occurs in sporadic glands in the normal endometrium, but this loss is insufficient to drive atypical glandular proliferation and progression. The PTEN negative glands need to acquire additional molecular ‘hits’ which contribute to the formation of AH/EIN.[7] The loss of PTEN function causes the PI3K/AKT pathway to become overactive. There is evidence that PI3K/AKT signaling enhances or turns on estrogen receptor-targeted gene expression. Therefore, loss of PTEN function may stimulate estrogen-dependent gene expression, leading to the overgrowth of cell types that depend on estrogen for trophic signals, such as endometrial and mammary epithelial cells.[8]

Longstanding molecular observations implicate PTEN inactivation, and it is seen in over 70% of the lesions. Thus, PTEN appears to be an early driver of atypical endometrial hyperplasia.[7] The systematic sequencing of endometrioid carcinomas has identified over 200 protein-encoding genes as pathogenic driver genes for type 1 EC.[3] Overall, there is dysregulation of the PI3K–PTEN–AKT–mTOR, RAS–MEK–ERK, and WNT–β-catenin pathways; microsatellite instability (MSI), mismatch repair (MMR) defects; and a relatively high occurrence of POLE mutations associated with a rare ultra-mutated phenotype.[5] The mutation of the ARID1A tumor suppressor gene may also be seen.[7]

TP53 inactivation as a major driver of most serous carcinomas (type 2), some high-grade endometrioid carcinomas (type 1), and many uterine carcinosarcomas; and inactivation of either gene as drivers of some clear cell carcinomas.[5][6] Mutation or inactivation of TP53 is an early event in the pathogenesis of serous tumors and is seen in nearly 75% of EICs, which precede serous carcinomas.[2] Other molecular events implicated in the pathogenesis of serous carcinoma include somatic mutations in PPP2R1A, FBXW7, SPOP, CHD4, and TAF1; amplification and/or overexpression of ERBB2, MYC, and CCNE1 (cyclin-E); and overexpression of p16 and synuclein-γ.[2][5] The altered PI3K pathway is also detected in serous carcinomas, mainly due to mutations in PIK3CA (in 17–43% of tumors), with some showing PTEN and PIK3R1 mutation.[3][5]

Epidemiology

The incidence of endometrial cancer has been shown to vary by race and ethnicity, with the highest rates among non-Hispanic White (NHW) women, and the lowest rates among Asian women with the worst survival rates among black women.[9] However, comparable data for premalignant lesions of the endometrium is not available.

Endometrial cancer is the most common gynecological malignancy in the Western world, and endometrial hyperplasia is its precursor.  Although historically considered a disease of postmenopausal women, the incidence of EC is rising, with alarming increases described in the 40–44-year-old age group. In the USA, EC is the fourth most common malignancy in women, with 1 in 37 women getting EC. According to the American Cancer Society (ACS), a diagnosis of about 65,620 uterine cancer cases will be made in 2020 in the US.[10] 

The incidence of endometrial hyperplasia is estimated to be at least three times higher than endometrial cancer, and if left untreated, it can progress to cancer.[11] Based on epidemiological registry data, it is estimated that in developed countries, there are about 200,000 new cases of EH per annum.[12] This is thought to be an underestimation.

A nested case-control study of EH progression, using extensive histopathology reports, concluded that AH was 14 times more likely to progress to endometrial carcinoma as compared to the women that presented with disordered proliferative endometrium without hyperplasia. Endometrial hyperplasia without atypia was unlikely to develop into endometrial carcinoma.[13]

Pathophysiology

There is a general consensus that accurate diagnosis of premalignant endometrial lesions and mitigation of risk factors can prevent progression to invasive carcinoma. During the cyclical hormonal endometrial remodeling, the endometrium responds to the estrogenic stimulation in a uniform programmed manner. The proliferative response is regulated by progestogenic effects. However, unopposed estrogenic stimulation of the endometrium can lead to the hyperproliferative glandular epithelium. This may lead to glandular remodeling and crowding of glands impinging on stromal spaces. Additionally, metaplastic changes may occur in glandular epithelium. If sporadic mutations accumulate in these hyperproliferative glands, prolonged or unopposed estrogen exposures act as a selection factor. The positively selected hyperproliferative clones then crowd out the normal glands.[14] 

Thus, two biologically distinct types of lesions, those that represent hormonal field effects i.e., AH and true precancerous lesions i.e., EIN, may be seen. They may either present independently or coexist in the same patient. Biologically, they have differing cancer risks and must be matched with appropriate intervention to avoid under- or over-treatment.[15]

Histopathology

EH is characterized by an increase in the gland-to-stroma ratio as compared to normal proliferative endometrium. Premalignant endometrial lesions, such as AH/EIN and EIC are clonal neoplasms that arise focally and can be diagnosed using specific criteria.

The precursor lesion of type 1 EC is AH. The WHO classification of endometrial hyperplasia includes two major categories, hyperplasia and atypical hyperplasia/endometrial intraepithelial neoplasia (AH/EIN).[16] While simple hyperplasia has a wide range of appearances, including an increased gland-to-stroma ratio and dilated glands; these lesions rarely progress to EC (1% to 3%). They are caused by persistent estrogen stimulation and may evolve into cystic atrophy when estrogen is withdrawn. AH/EIN shows proliferating glands with nuclear atypia.

The glands are crowded together and have complex branching patterns. Individual cells are rounded and haphazardly arranged on the basement membrane, having lost the perpendicular orientation seen in proliferative endometrial glands. The nuclei have prominent nucleoli and open or vesicular chromatin. It may not be possible to differentiate AH from well-differentiated endometrioid adenocarcinoma on biopsy. Up to 50% of women diagnosed with AH may have the final diagnosis of invasive EC when the uterus is examined after hysterectomy.[17]

The precursor lesion of type 2 serous carcinoma is EIC (serous endometrial intraepithelial carcinoma). It arises in the small atrophic uterus; and consists of malignant cells identical to those of serous carcinoma that is confined to the epithelial surface. The cells show marked cytologic atypia, including high nuclear-to-cytoplasmic ratio, atypical mitotic figures, hyperchromasia, and prominent nucleoli. Immunohistochemistry may be helpful in the differential diagnosis of AH and EIC. Atypical hyperplasia/EIN frequently shows PTEN and/or Pax-2 negativity and low Ki-67 and differs from EIC, which shows strong diffuse p53 staining and high Ki-67 staining index.[1][18]

History and Physical

Patients with endometrial hyperplasia mostly present with abnormal uterine bleeding. Women in the reproductive age group present in the form of menorrhagia and metrorrhagia, while postmenopausal bleeding is the presentation in the post-reproductive period. Occasionally patients may present with vaginal discharge. Some patients may be detected due to abnormal glandular or endometrial cells in the Pap smear.[19]

A detailed history and physical examination are required to determine and rule out other sources and causes of the presenting complaints or bleeding. A thorough and focused examination of the lower genital tract for pathologies and lesions of the lower genital tract, e.g., vulva, vagina, cervix, along with uterus and ovaries, should be done. This should include a pelvic examination. In obese patients where the pelvic examination may be difficult to interpret radiological investigations such as pelvic ultrasonography may be used. Pelvic ultrasonography is helpful in ruling out ovarian and other pathologies like uterine fibroids etc.[20]

Evaluation

Imaging the endometrium for diagnostic purposes is done on days 5-10 of the patient’s cycle. Endometrial hyperplasia can be reliably excluded if the endometrium measures less than 6 mm in thickness in pre-menopausal women and less than 5 mm thick in post-menopausal women.[20][21] The endometrium may appear smooth or more often will show an irregular surface with asymmetric/focal thickening. Cystic changes may be present. However, it is not possible to reliably differentiate EH from EC by ultrasonography;[20] unless frank myometrium invasion is seen.

In pre-menopausal patients who present with abnormal bleeding, a pregnancy test should always be done to exclude pregnancy. In all patients (pre- and post-menopausal) with abnormal vaginal bleeding, assessment of the coagulation profile with total platelet counts, prothrombin time, and partial thromboplastin time should be done to exclude coagulation and platelet disorders. A complete blood count should be ordered to assess anemia resulting from excessive blood loss (both chronic and acute).[22]

The final diagnosis of endometrial hyperplasia is usually made by hysteroscopy and direct biopsy sampling of the endometrium. Obtaining tissue via biopsy ideally needs to be done when the patient is not taking progestin-based hormones. If a patient is currently on hormonal therapy, a biopsy should be obtained 2 to 4 weeks after stopping therapy. Assessment of premalignant etiology may include molecular analysis (see etiology section).[23]

Treatment / Management

According to the American College of Obstetrics and Gynecologists (ACOG), the management plan for patients diagnosed with premalignant endometrial lesions includes:[24](B3)

• Ruling out the concurrent presence of endometrial carcinoma
• Strategic management plan designed to accommodate late diagnosis of occult carcinoma
• Prevention of progression to endometrial cancer

Management options of premalignant endometrial lesions include 1) surgical and 2) non-surgical options. A total hysterectomy (not supracervical hysterectomy) is the treatment of choice for premalignant endometrial lesions. In younger patients wanting to have children, high-dose progestin therapy with appropriate close histological monitoring is preferred.[24](B3)

Surgical Management: Total hysterectomy is the current standard of care for premalignant lesions of the endometrium. ACOG recommends that in clinically appropriate situations, total hysterectomy provides definitive treatment and assessment for a possible concurrent carcinoma. The supracervical hysterectomy, morcellation, and endometrial ablation are not recommended because of concerns of underlying carcinoma, which may be missed. Total hysterectomy with removal of the cervix with or without bilateral salpingo-oophorectomy (BSO), along with peritoneal washings, can provide proper staging of any incidentally discovered endometrial neoplasia and mitigates the risk of leaving behind any undetected or residual disease.

Surgical options include abdominal, vaginal, and minimally invasive procedures (such as laparoscopic or robotic approach) and may be chosen depending on the extent of the planned surgery and surgical skills of the operating surgeon. In contrast to abdominal hysterectomy, laparoscopic or vaginal hysterectomy is associated with less pain, earlier hospital discharge, and quicker recovery. However, the vaginal approach may not permit a comprehensive surgical staging, and removal of ovaries should intraoperative assessment require it. The scope of the surgery may change based on intraoperative frozen section guided pathologic review. Pathologic evaluation includes an assessment of myometrial invasion.[24] Lymphadenectomy is not recommended and is considered overtreatment and increased surgical risk for a majority of patients with premalignant lesions of the endometrium.(B3)

Non-surgical Management: Non-surgical management is offered to younger patients wanting to preserve fertility or patients with medical comorbidities; therefore, poor surgical candidates due to increased risk of complications. The therapeutic goal is 1) complete clearance of disease, 2) reversion to normal endometrium, and 3) prevention of invasive carcinoma.

Acceptable non-surgical treatments include hormonal therapy. Endometrial ablation is not advised as the completeness of ablation cannot be assessed. Additionally, it may cause fibrosis with adhesions in the endometrial cavity and making subsequent follow-up surveillance difficult or impossible. However, there are no standardized guidelines for optimal dosage and duration.[24][15][21](A1)

Hormonal therapy includes mainly progestins or suppression of estrogenic effects. In addition to this, selective estrogen receptor modulators (SERMS) may be used. Other agents used to reduce estrogenic effects are aromatase inhibitors, gonadotropin-releasing hormone (GnRH) antagonists, and sulfatase inhibitors. Hormonal therapy using progesterone derivatives is used extensively for any patient wanting to preserve fertility. Medroxyprogesterone acetate (MPA) and megestrol acetate, with different doses and schedules, are the most commonly used progesterones. Nearly 75% of atypical hyperplasia resolve with long-term high dosage progestin therapy.[24][21][15](A1)

Progesterone acts as a counterbalance to the mitogenic effects of estrogens. It induces secretory differentiation of the glandular epithelium.[25] Progesterone is normally well-tolerated and with side effects e.g., infrequent edema, gastrointestinal disturbances that are well tolerated. However, occasionally it may lead to thromboembolic events. It is also a reasonable option for any patient with a hyperplastic or pre-cancerous lesion who desires uterine retention. It can also be used for most elderly patients with medical comorbidities with either premalignant endometrial hyperplasia or low-grade malignancy.[15] (B3)

A full and detailed histopathologic examination of the endometrium is required to measure regression, persistence, or progression. Therefore examination of the entire uterus after hysterectomy is considered the “gold standard.” However, this option is not available for patients who receive non-surgical management in order to preserve fertility. These patients should be followed up with hysteroscopy and direct endometrial sampling when undergoing fertility-sparing management. No serum or tissue marker is available for follow up. Imaging modalities do not have the required specificity, so again follow up with hysteroscopy, and direct endometrial sampling is needed.[21] A repeated endometrial biopsy may lead to the elimination of the premalignant lesion nidus, yielding false-positive responses for hormonal therapy.(A1)

Hormonal therapy resistance is seen in nearly 30% of patients. This is thought to be due to a decreased availability of progesterone receptors and alteration of the apoptotic signaling pathway of the endometrial glandular cells.[21][15] Additional factors that may be responsible include the down-regulation of progesterone receptor and activation of the transforming growth factor signaling pathway. Less likely, resistance to hormonal therapy could result from mutations in PR or possibly paracrine effects. The histologic response of the glands of AEH/EIN is strongly coupled to the decidual response in the stroma, so the possibility of a paracrine effect is plausible, but the epithelial-stromal interactions of the endometrium are incompletely understood.[21][15][26][27](A1)

Differential Diagnosis

Although the criteria for AH/EIN are well defined, several conditions may mimic the cytological picture. Many of these lesions are interpreted by the pathologist as 'gland crowding.' Localized groups of altered and crowded endometrial glands may be misdiagnosed as premalignant or malignant lesions. In one study, follow-up outcomes of "gland-crowding" reports show 77% benign lesions (proliferative endometrium, secretory endometrium, endometrial polyp, etc.), 19% premalignant lesions, and 4% EC.[28] 

Other benign lesions that may mimic premalignant and malignant lesions when biopsies are examined may include mucinous metaplasia, papillary mucinous metaplasia, and endometrial polyps. If the history of pregnancy is not available, then the Arias-Stella reaction may be misinterpreted as the premalignant lesion. Endometrial hyperplasia with secretory changes may occasionally be difficult to distinguish from the "cork-screw" shaped extensive and proliferative glands of a late secretory endometrium.[28][29]

Prognosis

AH has been reported to progress to endometrial carcinoma in 29% of cases, with a mean duration of the progression being 4.1 years.[11] Additionally, AH may carry a higher risk of coexistent invasive carcinoma.[17] Recent studies report that up to 50% of women with atypia are found to have EC in subsequent hysterectomy specimens.[17][15] On the other hand, EC with concomitant hyperplasia tends to be of a lower grade and stage and higher 5-year survival rates.[30]

Complications

Complications of atypical hyperplasia may include severe uterine hemorrhage, requiring emergency medical or surgical management. Acute hemorrhage may be severe enough to require a blood transfusion.