Endometrial Hyperplasia and Cancer
Endometrial cancer is the most common malignancy of the female genital tract in the United States. Women have an overall lifetime risk of 2.5% of developing endometrial cancer. Fortunately, the majority of endometrial present at early stages with postmenopausal bleeding. Although obesity and estrogen excess remain the strongest risk factors for this disease, Lynch syndrome comprises the majority of inherited endometrial cancer cases; affected women have a 40% to 60% predicted lifetime risk of developing endometrial cancer.
For early-stage disease, standard management includes total abdominal hysterectomy, bilateral salpingo-oophorectomy, and staging. Management for women with advanced-stage disease primarily involves surgical resection and chemotherapy. Radiation therapy may be used for local control or in treating patients with positive lymph nodes, and hormonal therapies have been shown to be effective in a subset of patients. Several novel molecular-targeted therapies have been developed and evaluated for the treatment of endome-trial carcinoma. The principal benefit to these drugs, at this time, has been to prolong stable disease.
1. Endometrioid endometrial carcinomas account for approximately 80% of cases and typically occur in perimenopausal or postmenopausal women, are often of lower histologic grade, are often confined to the uterus and have a more favorable prognosis.
2. Risk factors associated with the development of endometrioid endometrial cancer include obesity, tamoxifen use, chronic anovulation, exogenous estrogen administration, nulliparity, early menarche, and/or late menopause.
3. Type 1 and 2 endometrial carcinomas exhibit distinct molecular alterations. The most common molecular alteration associated with type 1 tumors is loss of PTEN, whereas in type 2 tumors, it is p53 mutations.
Endometrial cancer is the most common gynecologic malignancy in the United States. An estimated 43,470 women will be diagnosed with uterine cancer in 2010, and it is estimated that 7950 of these women will die of the disease.1 Endometrial adenocarcinoma typically affects women in their perimenopausal or post-menopausal years and is most frequently diagnosed in women between the ages of 50 and 65 years. However, approximately 5% of cases are diagnosed in women before the age of 40 years, and approximately 10% to 15% of women are diagnosed before the age of 50 years.2 Women have an overall lifetime risk of 2.53% (1 in 40) of developing endometrial cancer.1 Approximately 90% of uterine tumors arise within the endometrium and are categorized as endometrial carcinomas. Of the endometrial carcinomas, 80% are endometrioid adenocarcinomas, and 15% to 20% are of more rare subtypes: papillary serous, clear cell, mucinous, or mixed carcinomas. These uncommon subtypes are associated with a poorer prognosis and greater risk of extra-uterine metastases when compared with endometrioid adenocarcinomas.
Multiple well-defined risk factors are associated with endometrial cancer and vary depending on the histologic subtype (Table 6-1). Type 1 endometrial cancer is associated with estrogenic stimulation; thus conditions that increase a patient’s level or duration of exposure to unopposed estrogen, in the absence of progesterone, will increase the risk of developing endometrial hyperplasia and ultimately carcinoma. Risk factors associated with excess or prolonged estrogen exposure include exogenous estrogen administration, chronic anovulation, obesity, tamoxifen use, nulliparity, early menarche, and/or late menopause.
Table 6-1 Risk Factors for Endometrial Cancer
Unopposed estrogen exposure is a well-established risk factor for the development of endometrial hyper-plasia and/or carcinoma. Estradiol acts as a mitogen in normal endometrial tissue. During the follicular phase, plasma estradiol levels predominate at normal premenopausal concentrations while progesterone levels are low. Endometrial proliferation rates remain high throughout the follicular phase. Plasma estradiol levels remain elevated until ovulation, when they begin to fall rapidly and the corpus luteum produces and secretes progesterone. During the luteal phase of the menstrual cycle, progesterone predominates and counters the estrogenic effects on the endometrium by promoting the local synthesis of 17β-hydroxysteroid dehydrogenase and estrogen sulfo-transferase, which favor the conversion of estradiol to the less potent estrogen (E1) and into estrogen sulfates that are quickly eliminated from the body.3 Thus any derangement in the normal balance between estradiol and progesterone will lead to continuous endometrial stimulation and proliferation. Over time, this persistent stimulation may lead to endometrial hyperplasia and/or carcinoma.
Tamoxifen is a selective estrogen receptor modulator that demonstrates either estrogenic or antiestrogenic effects in different tissues. In the uterus it functions as an estrogen agonist, whereas in the breast it functions as an estrogen antagonist. Tamoxifen is commonly used in the treatment of estrogen receptor–positive breast cancer, but is also associated with an increased risk of uterine cancer. The National Surgical Adjuvant Breast and Bowel Project (NSABP) B-14 trial compared rates of endometrial cancer in tamoxifen- and non–tamoxifen-treated patients and described the pathologic characteristics of the endometrial cancers.4 This study demonstrated a 7.5-fold increased risk of endometrial carcinoma in patients treated with tamoxifen relative to placebo controls. The annual hazard rate through all follow-up was 0.2 in 1000 in the placebo group and 1.6 in 1000 in the tamoxifen-treated group. The majority of endometrial cancer cases that developed during this study were of an early stage and low to moderate grade. In the NSABP P-1 chemoprevention trial, 13,388 women at increased risk for breast cancer were randomized to receive either tamoxifen (20 mg/d) or placebo for 5 years.5 Although tamoxifen reduced the risk of invasive breast cancer by 49% , the rate of endometrial cancer was increased in the tamoxifen group (risk ratio, 2.53; 95% confidence interval [CI], 1.35-4.97), particularly in women over the age of 50 years.5 Current recommendations for management of women treated with tamoxifen include performing an endometrial biopsy for women who develop irregular or postmenopausal vaginal bleeding.
Obesity is an established risk factor for the development of multiple cancer types, cancer-related mortality, and all-cause mortality.6 Among all cancer types, increasing body mass index (BMI) and obesity is most strongly associated with endometrial cancer incidence and mortality.6 The development of endome-trial cancer in obese women is thought to be a result of the peripheral conversion of androstenedione to estrone by aromatase in adipose tissue. In a recent meta-analysis of 19 reviews and prospective studies, Renehan et al7 found that each increase in BMI of 5 kg/m2 significantly increased a woman’s risk of developing endometrial cancer (relative risk [RR], 1.59; 95% CI, 1.50-1.68). Endometrial cancer mortality is also adversely affected by obesity, both directly and indirectly. Calle et al,8 in a prospective study of more than 495,000 women followed for 16 years, examined the relationship between BMI and increased risk of cancer mortality. There was a clear trend associated with increasing BMI; the relative risk of uterine cancer–related death for women considered obese (BMI 30-34.9 kg/m2) was 2.53, whereas for morbidly obese women it was 6.25.8 The association between obesity and other medical comorbidities, such as diabetes mellitus and hypertension, adversely affects endometrial cancer–related mortality and all-cause mortality. In a retrospective review of 380 patients with early endometrial cancer, the Gynecologic Oncology Group (GOG) found that morbid obesity was associated with a higher mortality (HR, 2.77; 95% CI, 1.21-6.36) from causes other than endometrial cancer or disease recurrence.9 Unfortunately, public knowledge of the association between obesity and cancer risk is limited, with a recent survey indicating that up to 58% of women were not aware that obesity increased endometrial cancer risk.10
Approximately 5% of endometrial cancer cases are attributed to an inherited predisposition. Hereditary nonpolyposis colorectal cancer (HNPCC), now known as Lynch syndrome, comprises the majority of inherited cases. Affected women have a 40% to 60% predicted lifetime risk of developing endometrial cancer and a 10% to 12% lifetime risk of developing ovarian cancer.11,12 Lynch syndrome is inherited in an autosomal dominant pattern with incomplete penetrance due to a germline mutation in one of the mismatch repair (MMR) genes. The MMR genes include MLH1 on chromosome 3, MSH2 and MSH6 on chromosome 2, and PMS2 on chromosome 7. Loss of MMR gene function results in microsatellite instability (MSI), which leads to the accumulation of somatic mutations that are presumed to affect key regulatory genes related to cell growth and/or apoptosis. It is important to note that the incidence of endometrial cancer in female Lynch carriers actually equals or exceeds that of colorectal cancer.13 Moreover, carriers who develop cancer are also at an increased risk of developing second subsequent metachronous malignancies. In 50% of patients with both colon and endometrial cancers, endometrial cancer is the sentinel event. The risk for women to develop endometrial cancer appears to differ slightly based on the specific germline mutations in MMR genes. The estimated lifetime risk at age 70 years is 26% for MSH6 mutation carriers and 27% for MLH1 mutation carriers, whereas it is 40% for MSH2 mutation carriers.14–16
There are 2 primary guidelines used to identify families. The first is the Amsterdam Criteria, originally designed to diagnose Lynch syndrome in certain families based on clinical criteria. However, with the identification of MMR genes and the original criteria not accounting for extra-colonic cancers, the modified Amsterdam Criteria II was revised to be more inclusive (Table 6-2). The Bethesda guidelines were developed as a screening tool to identify which individuals should undergo MSI testing and then genetic testing. These guidelines focus predominantly on colon cancer. The goal of these guidelines was to determine which families should have MSI testing before screening for MMR mutations. The Society of Gynecologic Oncologists recently published a committee statement with guidelines for identifying women with Lynch syndrome (Table 6-3).17
Table 6-2 Amsterdam II Criteria
1. Three or more relatives with an associated cancer (colorectal or endometrial cancer, cancer of the small intestine, ureter, or renal pelvis)
2. Two or more successive generations affected
3. One or more relatives diagnosed before the age of 50 years
4. One should be a first-degree relative of the other two
5. Familial adenomatous polyposis should be excluded in cases of colorectal carcinoma
6. Tumors should be verified by pathologic examination
Table 6-3 Society of Gynecologic Oncologists Statement Guidelines on Risk Assessment for Lynch Syndrome
1. Family pedigree meeting Amsterdam Criteria
2. Patients with metachronous or synchronous colorectal and endometrial or ovarian cancers before age 50 years
3. Those with a first- or second-degree relative with a known germline mutation in a MMR gene.
For patients with a 5%-10% risk of having Lynch syndrome, genetic testing was classified as being “helpful.” These patients include:
1. Patients with endometrial or colorectal cancer diagnosed before age 50 years
2. Patients with endometrial and/or ovarian cancer and a synchronous or metachronous Lynch-associated malignancy before age 50 years
3. Patients with endometrial or colorectal cancer and a first-degree relative diagnosed with a Lynch-associated malignancy before age 50 years
4. Patients with endometrial or colorectal cancer at any age with ≥ 2 first- or second-degree relatives diagnosed with a Lynch-associated malignancy at any age
5. A patient with a first- or second-degree relative who meets the above criteria
There are limited data on the efficacy of endometrial cancer screening in Lynch syndrome carriers. Current recommendations advise women to undergo annual endometrial biopsies (EMB) beginning between the ages of 30 and 35 years or 10 years before the first endometrial cancer diagnosis in the family. Several studies have examined the utility of transvaginal ultrasound as a screening modality; however, as the only method, it is ineffective for detecting early endometrial cancer. Prophylactic hysterectomy with bilateral salpingo-oophorectomy has been recommended as a prevention strategy after completing childbearing; however, the specific age remains controversial.18 Thus counseling for prophylactic surgery in Lynch carriers having completed childbearing is on a case-by-case basis, balancing benefits of ovarian function and reducing cancer risk.
Based on epidemiologic, molecular, and prognostic factors, endometrial cancer can be subdivided broadly into 2 types.19 Type 1 carcinomas, accounting for approximately 80% of cases, are classically of endometrioid histology and are usually preceded by endometrial atypical hyperplasia. These tumors typically occur in perimenopausal or postmenopausal women, are often of lower histologic grade, are often confined to the uterus at presentation, and thus have a more favorable prognosis. Type 1 carcinomas commonly express estrogen and progesterone receptors and are associated with unopposed estrogen exposure. Up to 90% of type 1 endometrial cancer patients are obese.20
Type 2 carcinomas are often of nonendometrioid histology and typically arise in a background of atrophic endometrium. These lesions appear to be unrelated to estrogen stimulation and are not typically preceded by endometrial atypical hyperplasia. The precursor lesion of type 2 carcinoma is termed endometrial intraepithelial neoplasia. Type 2 lesions include serous and clear cell histologies and have a propensity for early metastatic spread and a poor prognosis. Type 2 tumors are not typically associated with obesity.
Molecular Biology of Endometrial Cancer
Type 1 and type 2 lesions exhibit distinct molecular alterations. The most frequent genetic alteration associated with type 1 lesions is a loss of function of the tumor suppressor PTEN. PTEN loss of function can be seen in up to 83% of endometrioid carcinomas and 55% of precancerous lesions.21 Mutations in PTEN have been documented in endometrial hyperplasia with and without atypia and thus have been postulated to be an early event in the endometrial tumorigenesis process.22 PTEN most notably plays a role in the regulation of the phosphotidylinositide 3-kinase (PI3K)-Akt-mammalian target of rapamycin (mTOR) pathway by inhibiting the downstream phosphorylation of AKT, but its loss of function has also has been shown to result in genomic instability by causing defects in either homologous recombination DNA repair or in cell cycle checkpoints. Mutations on PIK3CA are also relatively common and are seen in up to 36% of endometrioid cancers.23 These mutations tend to occur in tumors that also have PTEN loss.
Other common genetic alterations associated with type 1 endometrial carcinomas include MSI, mutations in K-ras, and mutations in β-catenin. Approximately 20% to 30% of type I lesions exhibit MSI. Microsatellites are repeated sequences of DNA of a set length predominately found in noncoding DNA that are variable from person to person. The propensity of a tumor to develop changes in the number of repeat elements compared with normal tissue due to defects in the DNA mismatch repair process is termed microsatellite instability. This leads to replication errors that may inactivate or alter tumor suppressor genes. MSI has also been shown to occur early in the tumori-genesis process and is associated with a higher rate of PTEN mutations.24,25
K-ras mutations have been found in up to 30% of endometrial cancers.26 When K-ras is mutated, it functions as an oncogene upregulating signaling through the mitogen-activated protein kinase pathway. As with PTEN, mutations in K-ras are more frequent in MSI-positive tumors.27
β-catenin plays a role in signal transduction as a transcriptional activator in the Wnt signaling pathway and is a member of the E-cadherin unit of proteins that is essential for maintenance of normal tissue architecture and cell differentiation. Gain of function mutations in β-catenin are seen in 25% to 38% of type 1 lesions.28 β-catenin mutations have been identified in atypical endometrial hyperplasia, indicating that it also may be an early event in endometrial tumorigenesis.29 However, whereas MSI and mutations in PTEN and K-ras tend to coexist, β-catenin gain of function mutations are usually seen alone.30
In contrast, the genetic alterations most commonly seen in type 2 lesions are p53 mutations, HER-2/neu amplification, and p16 inactivation. The most common of these is a mutation in the tumor suppressor gene, p53, which is found in up to 90% of serous carcinomas (compared with 10% of type 1 lesions).26 Mutations in p53 are also seen in up to 80% of endometrial intraepithelial lesions, the precursor lesions of serous carcinomas.31 HER-2/neu is an oncogene involved in cell signaling. Overexpression and amplification of HER-2/neu occurs in 43% and 29% of serous carcinomas, respectively.32 p16 is also a tumor suppressor gene involved in cell cycle regulation. Inactivation of p16 has been identified in 45% of serous carcinomas and less frequently in clear cell carcinomas.31 These genetic changes are also seen in preneoplastic atrophic endometrium, indicating that they are early events in type 2 endometrial tumorigenesis.33
The identification of these genetic alterations has led to the development and implementation of several targeted therapeutic strategies for the management of endometrial cancer, which are reviewed later in this chapter.
1. Postmenopausal bleeding is the most common presenting symptom of endometrial cancer.
2. Endometrial cancer screening strategies are unnecessary for women at general population risk.
4. Transvaginal ultrasound has been evaluated as a diagnostic tool for endometrial cancer in patients with postmenopausal or irregular vaginal bleeding.
5. Dilation and curettage remains the gold standard for diagnosing endometrial cancer; however, in-office endometrial sampling devices have been shown to be highly accurate in women with an endometrial stripe thickness of less than 7 mm.
Endometrial cancer is classically a disease of perimenopausal and postmenopausal women. The initial endometrial lesion arises in the glandular component of the uterine lining. As it forms a mass, it contains areas of superficial necrosis and becomes more friable (Figure 6-1). As a result, approximately 90% of women with endometrial cancer present with abnormal vaginal bleeding. Up to 20% of women with postmenopausal bleeding have an underlying endometrial carcinoma or hyperplasia. Although postmenopausal bleeding is most common, perimenopausal or anovulatory premenopausal women with intermenstrual bleeding or menometrorrhagia should also be evaluated for endometrial cancer. Unfortunately, as there are many potential causes for abnormal bleeding in this patient population, diagnosis may often be delayed (Table 6-4). In some cases, particularly in older women or women who have undergone prior cervical conization or loop electrosurgical excision procedure (LEEP), cervical stenosis may mask the development of postmenopausal bleeding. In this situation, patients may present with hematometra or pyometra. Postmenopausal women experiencing vaginal bleeding, perimenopausal women with heavy or prolonged bleeding, and anovulatory or oligovulatory premenopausal women with abnormal bleeding are considered high risk and warrant endometrial sampling. Furthermore, it is recommended that any woman over the age of 35 years with prolonged or heavy vaginal bleeding undergo endometrial sampling.
FIGURE 6-1. Endometrial carcinoma (with extension into the cervix).
Table 6-4 Differential Diagnosis of Postmenopausal Bleeding
Benign intracavitary lesions
Vaginal or cervical trauma
On physical examination, finding additional abnormalities is unlikely unless the patient is presenting at an advanced stage of disease with ascites or carcinomatosis. However, it is important to perform a thorough physical examination because a majority of these patients are obese and have other medical comorbidities, including hypertension and diabetes mellitus. On pelvic examination, examination of the vulva, vagina, and cervix is important to exclude metastatic disease. The appearance and patency of the cervical os should be noted, as stenosis may delay the manifestation of postmenopausal bleeding. On bimanual examination, the uterus may feel bulky or tender to palpation, particularly with hematometra, pyometra, or advanced disease. Rectovaginal examination should be performed to palpate the posterior cul-de-sac, adnexa, and parametria.
Endometrial cancer screening strategies are unnecessary for women at general population risk. Even women who are at increased risk, including those women being treated with tamoxifen, do not benefit from endometrial cancer screening. Pap smear screening for endometrial cancer is unreliable. The finding of atypical glandular cells on Pap smear should warrant further evaluation with endometrial biopsy and endocervical curettage. However, only 50% of patients with endometrial cancer will have an abnormality on Pap smear. Those patients who do are more likely to have more advanced disease. Compared with patients with normal cervical cytology, those with malignant endometrial cells have more than twice the risk of deep myometrial invasion, twice the risk of grade 2 or 3 tumor, and 3 times the risk of positive peritoneal washings.
Transvaginal ultrasonography to evaluate the thickness and contour of the endometrial stripe has been evaluated as a potential endometrial cancer screening technique, especially in women taking tamoxifen. However, Love and colleagues,34 in an investigation of 357 women treated with tamoxifen and 130 controls who were screened with transvaginal ultrasonography, found that although ultrasound identified a statistically significant positive correlation between length of time on tamoxifen and endometrial thickness, it was a poor screening tool because of the high false-positive rate (46%). Proponents of transvaginal ultrasonography for endometrial screening have argued that it is a relatively noninvasive technique that may help providers determine which patients should undergo endometrial sampling.
Transvaginal ultrasonography has been proposed to identify women with postmenopausal bleeding who are highly unlikely to have endometrial disease so that endometrial sampling may be unnecessary. In a multi-institutional study of more than 1100 women, Karlsson and colleagues35 established an endometrial stripe thickness cut-off of 5 mm to help triage patients with postmenopausal bleeding toward endometrial biopsy. This yielded a sensitivity of 94%, a specificity of 78%, a positive predictive value (PPV) of 69%, and a negative predictive value (NPV) of 96%. They found that no malignant endometrium was thinner than 5 mm and determined that the risk of finding any endometrial abnormality when the endometrial stripe thickness was ≤ 4 mm was 5.5%. As a result, they concluded that given the high NPV, it would be reasonable to refrain from endometrial sampling in women with postmenopausal bleeding who cannot undergo endometrial sampling. However, others have argued that the presence of a thin endometrial stripe does not reduce the need for endometrial sampling, because up to 4% of endometrial cancers would be missed using this strategy, with a false-positive rate as high as 50%.36 More recently, Timmermans and colleagues37 conducted a systematic review and meta-analysis of 90 studies reporting on endometrial stripe thickness in women with postmenopausal bleeding and a diagnosis of endometrial carcinoma. The authors concluded that previous studies had likely overestimated the diagnostic accuracy of endometrial stripe thickness in the detection of endometrial cancer and recommended reducing the cut-off to 3 mm.
Due to the unreliable results associated with Pap screening and the difficulties with conclusively interpreting endometrial stripe thickness, in-office endometrial sampling is a necessary step in the evaluation of women who present with abnormal or postmenopausal vaginal bleeding. Several devices have been developed and are commercially available. These devices are used for direct sampling of the endometrium and allow both cytologic and histologic evaluation of the uterine lining. A meta-analysis reported that the Cornier Pipelle (Prodimed, Neuilly-en-Thelle, France) was the most effective device, with detection rates for endometrial carcinoma in postmenopausal and premenopausal women of 99.6% and 91%, respectively.38 Use of the Pipelle to detect endometrial carcinoma or hyperplasia has been shown to be effective, with a sensitivity of 84.2%, specificity of 99.1%, accuracy of 96.9%, PPV of 94.1%, and NPV of 93.7%.39 In addition, the accuracy of in-office endometrial biopsy is comparable to the gold standard of dilation and curettage (D&C) only for an endometrial stripe thickness of less than 7 mm.40 For an endometrial stripe thickness greater than 7 mm, D&C may be superior. Thus women reporting postmenopausal bleeding with a negative endometrial biopsy warrant further investigation.
The usefulness of combining transvaginal ultrasound with endometrial biopsy has been investigated as a diagnostic schema for the detection of endome-trial hyperplasia and carcinoma, with excellent results. In a study of 552 women, Minagawa and colleagues41 found that the combined method achieved a sensitivity of 100%, a specificity of 99.1%, a PPV of 92.9%, and an NPV of 100% for the detection of endometrial carcinoma. For endometrial hyperplasia, the combined method resulted in a sensitivity of 100%, a specificity of 89.6%, a PPV of 40.0%, and an NPV of 100%.41
More recently, hysteroscopy has been combined with D&C in the diagnostic evaluation of women with a thickened endometrial stripe. Although hysteroscopy is generally used to identify benign lesions, such as endometrial polyps, it may also be used to examine the uterine lining and specifically biopsy suspicious lesions in the uterine lining under direct visualization.
Role of the General Gynecologist
Because the majority of women with endometrial hyperplasia or carcinoma initially will experience abnormal vaginal bleeding, they often first present to their general gynecologist or primary care provider for evaluation. It is important for the gynecologist to take a thorough history and perform a complete pelvic examination, as described previously. Key aspects of the patient’s history that should be considered when establishing a differential diagnosis include menopausal status, age at menarche and menopause, parity, history of infertility, history of hormone therapy use and duration, history of tamoxifen use and duration, and family history of uterine or colon cancers. Women with postmenopausal bleeding should be considered to have endometrial cancer until proven otherwise. The initial evaluation, including a transvaginal ultrasonography and endometrial biopsy as described earlier, is routinely performed by a general gynecologist. Patients with persistent postmenopausal bleeding and a negative endometrial biopsy should consider a definitive diagnostic evaluation with D&C to rule out the presence of malignancy. Referral to a gynecologic oncologist is recommended after a pathologic diagnosis of endometrial carcinoma.
Often women who are diagnosed with endometrial hyperplasia are managed and followed up by their general gynecologist. A discussion on the management considerations for women with endometrial hyperplasia is included later in this chapter.
1. Complex atypical hyperplasia is considered a premalignant lesion for endometrioid endometrial cancers.
2. Diagnostic criteria for endometrioid carcinoma include (1) back-to-back proliferation of endometrial glands occupying an area of 2 × 2 mm; (2) an extensive papillary pattern; and (3) a desmoplastic of fibroblastic stroma infiltrated by irregular glands.
3. Uterine serous carcinoma accounts for approximately 8% to 10% of endometrial cancers and are characterized by early extrauterine metastasis and a worse overall prognosis.
4. Endometrial cancer can spread by direct extension to adjacent structures, lymphatic dissemination, hematogenous dissemination, or transtubal passage of exfoliated cancer cells.
Endometrial hyperplasia is characterized by the proliferation and crowding of endometrial glands and stroma, resulting in an increased gland-to-stroma ratio. As mentioned previously, endometrial hyperplasia is believed to result from excessive or prolonged exposure to estrogen that is unopposed by the effects of progestin. In 1984, the International Society of Gynecologic Pathologists introduced terminology to classify endometrial hyperplasia that was adopted by the World Health Organization (WHO). According to the WHO definition, endometrial hyperplasia is subdivided into simple and complex on the basis of the architecture of endometrial glands. In simple hyper-plasia, the glands maintain round shapes and may be dilated, but there is abundant stroma (Figure 6-2). In contrast, with complex hyperplasia, the glands assume branched and complex outlines and may exhibit back-to-back crowding with little endometrial stroma. In both simple and complex hyperplasia, the cells can have cytologic atypia (Figure 6-3). Cytologic atypia is characterized by a loss of cellular polarity, an increase in the nuclear-to-cytoplasmic ratio, and prominent nucleoli. Thus endometrial hyperplasia can be further divided into 4 subclassifications: (1) simple hyperplasia without atypia, (2) simple atypical hyperplasia, (3) complex hyperplasia without atypia, and (4) complex atypical hyperplasia. Progression to carcinoma varies based on classification (Table 6-5), but is most prevalent in patients with complex atypical hyperplasia, occurring in up to 26.7% to 29% of cases.42 Furthermore, these 2 lesions often coexist in the same endometrium. Up to 42.6% of endometrial biopsies with complex atypical endometrial hyperplasia will result in a diagnosis of carcinoma in the subsequent hysterectomy specimen.43 In addition, a recent study by the GOG showed that the application of WHO criteria for endometrial hyperplasia can be somewhat variable among pathologists, resulting in a relatively low level of reproducibility of the diagnosis of complex atypical endometrial hyperplasia.44 They concluded that a new classification system, which is both highly reproducible among pathologists and more predictive of lesions on hysterectomy, is needed.
Table 6-5 Comparison of Simple and Complex Hyperplasia With or Without Atypia and Progression to Endometrial Carcinoma
FIGURE 6-2. Simple hyperplasia without atypia. The glands are crowded, but intervening stroma is abundant. The glands maintain round to oval outlines, and there is no cytologic atypia.
FIGURE 6-3. Complex atypical hyperplasia. The glands are crowded, but intervening stroma is abundant. The glands show complex outlines and cytologic atypia.
Cancers of the uterine corpus can be divided into epithelial, mesenchymal, mixed epithelial and mesenchymal, and trophoblastic tumors. Mesenchymal uterine tumors and trophoblastic tumors are discussed in Chapter 7 and Chapter 8, respectively. The various histologic subtypes of endometrial carcinoma are listed in Table 6-6.45
Table 6-6 Histologic Subtypes of Endometrial Cancer
Uterine papillary serous adenocarcinoma (UPSC)
Clear cell adenocarcinoma
Squamous cell carcinoma
Mixed adenocarcinoma and other rare variants
Endometrioid carcinomas are the most common histologic subtype and comprise approximately 80% to 90% of all endometrial cancers. For a diagnosis of endometrioid endometrial adenocarcinoma, one of the following diagnostic criteria must be met: (1) back-to-back proliferation of endometrial glands occupying an area of 2 × 2 mm; (2) an extensive papillary pattern; and (3) a desmoplastic of fibroblastic stroma infiltrated by irregular glands.46 Endometrial tumors are graded based on their degree of differentiation and the amount of solid component present. Grade 1 tumors are well differentiated and have less than 5% of a solid component (Figure 6-4). Grade 2 tumors are of intermediate differentiation and have between 6% and 50% of a solid component (Figure 6-5). Grade 3 tumors are poorly differentiated and have more than 50% of a solid component (Figure 6-6). There are multiple recognized variants within this subtype, including variant with squamous differentiation, villoglandular (or papillary) differentiation, secretory variant, and ciliated-cell variant. These variants all have a similar clinical course and prognosis to that of typical endometrioid adenocarcinoma and as such are grouped together.
FIGURE 6-4. Grade 1 endometrioid cancer. The majority of the tumor is composed of back-to-back glands, with little to no intervening stroma.
FIGURE 6-5. Grade 2 endometrioid cancer. The tumor is a mixture of back-to-back glands and solid tumor nests.
FIGURE 6-6. Grade 3 endometrioid cancer. The majority of the tumor is composed of solid tumor nests.
Uterine serous carcinomas (USC) are a highly aggressive subtype that histologically resembles high-grade ovarian papillary serous carcinomas (Figure 6-7). USCs, however, are not graded. They account for approximately 8% to 10% of endometrial cancers and are characterized by early extrauterine metastasis and a worse overall prognosis, with a 5-year overall survival rate of approximately 50%.45,47 Several studies have shown that the depth of myometrial invasion does not correlate with the incidence of extrauterine metastasis. Slomovitz and colleagues47 found that among patients with no uterine invasion, 37% had extrauterine disease.
FIGURE 6-7. Uterine serous carcinoma. Tumor cells show high-grade cytologic atypia, loss of intraepithelial polarity, and formation of papillae without fibrovascular cores.
Clear cell carcinomas are less common than USCs, accounting for approximately 2% to 3.7% of endometrial cancers.48 Histologically they resemble clear-cell carcinoma of the ovary and vagina (Figure 6-8). Similar to USCs, uterine clear-cell carcinomas are not graded. They also have a higher frequency of extra-uterine metastases when compared with endometrioid carcinomas. There is a poor correlation between the depth of myometrial invasion and the presence of extrauterine disease; extrauterine metastases can be found in up to 50% of cases with clear cell carcinoma confined to the inner one-half of the myometrium. McMeekin and colleagues48 found that clear cell histology was an independent predictor of a worse progression-free survival. The 5-year overall survival rate is 62% for patients with clear cell histology.45
FIGURE 6-8. Clear cell carcinoma. The tumor is composed of sheets of cells with cytoplasmic clearing and high-grade nuclei.
Although carcinosarcoma (or malignant müllerian mixed tumor) is not part of the current WHO classification of endometrial carcinoma, clonality and mutational studies have shown that the carcinomatous and the sarcomatous components derive from the same precursor.49–52 In addition, based on patterns of recurrence and metastases, the behavior of this tumor is more akin to that of carcinoma than sarcoma. Tumors designated as carcinosarcoma must contain both a malignant epithelial and a malignant mesenchymal (sarcomatous) component, which can be clearly demarcated from each other on histologic examination (Figure 6-9). When matched for stage, age, patient performance status, and surgical procedure, carcinosarcomas have been found to have a worse outcome than endometrioid, clear cell, and serous carcinomas.53
FIGURE 6-9. Carcinosarcoma. The malignant epithelial component (left) is a high-grade endometrioid carcinoma. The malignant stromal component (right) is a high-grade sarcoma with liposarcomatous differentiation. Note the sharp transition between the malignant epithelial and stromal components.
Endometrial cancer can spread by direct extension to adjacent structures, lymphatic dissemination, hematogenous dissemination, or passage of exfoliated endometrial cancer cells through the fallopian tubes. The most common route of spread is through direct extension. Initially, the primary endometrial tumor will grow to involve the majority of the endometrial surface and extend into the lower uterine segment. Simultaneously, the tumor invades into the myometrium, extending to eventually involve the uterine serosa and the cervix. On endometrial biopsies, this can represent a diagnostic challenge for the pathologist, because endometrioid endometrial adenocarcinoma and endocervical adenocarcinoma can have overlapping histologic features.
The presence of concurrent complex atypical endometrial hyperplasia in the biopsy favors the diagnosis of endometrial adenocarcinoma, whereas the presence of concurrent adenocarcinoma in situ of the endocervix favors the diagnosis of endocervical carcinoma. In addition, the use of an immunohistochemical staining panel can be helpful to make this distinction. Endome-trial adenocarcinoma typically shows diffuse staining for estrogen receptor and vimentin and with patchy staining for p16, whereas the carcinoembryonic antigen is negative. In contrast, the endocervical adeno-carcinoma usually shows diffuse staining for p16 and carcinoembryonic antigen, but no staining for estrogen receptor and vimentin.54
The location of the primary tumor determines to some degree the timing of cervical involvement. Primary tumors in the uterine fundus often invade and extend to the uterine serosa before involving the cervix. Tumors that originate in the lower uterine segment tend to involve the cervix earlier. The mechanism of spread to the cervix likely involves a combination of surface spread, lymphatic spread, and invasion of deep tissue planes.55 Once tumors penetrate the uterine serosa, they may directly invade other pelvic structures such as the bladder, rectum, or adnexa, or cells may exfoliate into the peritoneal cavity to form metastatic implants throughout the abdomen.
The uterus has a complex lymphatic network that follows the major blood vessels supplying the uterus. The lymphatic channels that drain the fundal portion of the uterus pass through the infundibulopelvic ligaments and follow the ovarian vessels to the para-aortic lymph nodes. The lymphatic channels that drain the mid and lower portions of the uterus travel through the broad ligament while following the uterine vessels to the pelvic lymph nodes. Small lymphatic channels also travel through the round ligaments to the superficial inguinal lymph nodes. As a result of this complex lymphatic network, nodal metastases can theoretically occur in any combination of nodal basins. Creasman and colleagues56 have reported on the correlation between tumor histologic grade and depth of myome-trial invasion with the incidence of pelvic and para-aortic nodal metastases (Tables 6-7 and 6-8). Studies of endometrial cancer patients from the Mayo Clinic have elucidated the pattern of lymph node spread.57 The external iliac lymph nodes are the most commonly involved pelvic lymph nodes in tumors confined to the uterus or involving the cervix. However, when compared with tumors confined to the uterine corpus, tumors extending to the uterine cervix have a higher rate of common iliac nodal involvement. Furthermore, although lymphatic channels pass directly from the uterine fundus to the para-aortic lymph nodes, it is rare to find positive para-aortic lymph nodes in the absence of pelvic lymph nodes. In a separate consecutive series of 612 endometrial cancer patients from the Mayo Clinic, Mariani and colleagues58 identified 2 independent predictive factors of para-aortic nodal metastases: positive pelvic lymph nodes and the presence of lymph vascular space invasion (LVSI). Only 2% of patients with negative pelvic lymph nodes had positive para-aortic lymph nodes, compared with 47% of patients with positive pelvic nodes. Furthermore, when the pelvic lymph nodes and LVSI were both negative, only 0.8% of patients had positive para-aortic lymph nodes, compared with 31% in patients with at least 1 of these variables.
Table 6-7 Frequency of Positive Pelvic Nodes in Relationship to Tumor Grade and Depth of Myometrial Invasion
Table 6-8 Frequency of Positive Para-Aortic Nodes in Relationship to Tumor Grade and Depth of Myometrial Invasion
Hematogenous dissemination of endometrial carcinoma does occur, but is less common than lymphatic spread or direct extension. The most common site of hematogenous spread is to the lungs, and liver, brain, and bone are also less common sites of metastasis.
Transtubal migration of exfoliated endometrial cancer cells is a less common route of spread, but may explain the presence of positive peritoneal washings and/or disseminated intraperitoneal metastases in women with otherwise early endometrial cancer. In series of 87 patients who underwent hysterectomy for uterine serous carcinoma, Snyder and colleagues59 found tumor clusters within the fallopian tube lumen of 16 patients, all of whom had peritoneal spread. Three of these 16 patients had no evidence of myome-trial invasion or LVSI. There have also been concerns that hysteroscopy may facilitate transtubal passage of endometrial cancer cells, resulting in peritoneal dissemination. Obermair and colleagues60 reported on a retrospective analysis of 113 women with stage I endometrial carcinoma confined to the inner half of the myometrium. They noted that 10 patients (9%) had suspicious or positive peritoneal cytology, and this was significantly associated with a history of hysteroscopy. In general, the prognostic significance of this finding is uncertain. Most recently, positive washings have been removed from surgical staging.
Endometrial carcinoma is staged surgically according to the International Federation of Gynecology and Obstetrics (FIGO) staging system. Comprehensive surgical staging includes a hysterectomy, bilateral salpingo-oophorectomy, bilateral pelvic and para-aortic lymphadenectomy, and peritoneal washings. This procedure can be accomplished through either a laparotomy incision or by a laparoscopic or robotically assisted approach.
Before 1988, endometrial cancer was staged clinically based on the depth of the uterine cavity, physical examination findings, and fractional biopsy specimens from the endocervix and endometrium. However, several studies comparing the accuracy of clinical versus surgical staging consistently demonstrated the superiority of surgical staging.61–63 As a result, clinical staging was abandoned, and in 1988, FIGO approved a surgical staging system for carcinoma of the uterine corpus. This was felt to be inadequate by many, given that uterine sarcomas have a different histologic appearance, clinical behavior, and prognosis than endometrial adenocarcinomas and should, therefore, not be grouped together in the same staging system. In response, FIGO revised the surgical staging system for endometrial carcinomas in 2009 (Table 6-9) and developed a separate staging system for uterine sarcomas (reviewed in Chapter 7).
Table 6-9 FIGO 2009 Surgical Staging for Endometrial Carcinoma
In 2009, the FIGO staging system was again revised (Table 6-9). The key differences between the 1988 staging system and the revised 2009 staging system include changes in reporting the depth of myometrial invasion, endocervical glandular involvement, and the extent of advanced (stage III) disease. According to the 1988 staging system, tumors confined to the uterine corpus (stage I) were divided into 3 subgroups based on the depth of myometrial invasion. Stage IA was defined as no myometrial invasion, whereas stage IB was defined as the presence of less than 50% myometrial invasion and stage IC as greater than 50% myometrial invasion. However, as data from the FIGO Annual Report showed no significant difference in 5-year survival rates between stage IA grade 1, stage IB grade 1, stage IA grade 2, or stage IB grade 2 (93.4%, 91.6%, 91.3%, and 93.4%, respectively), the previous stage IA and IB were combined into stage IA.64 Stage IB is now defined by the presence of ≥ 50% myome-trial invasion.
The staging of cervical involvement has also changed for in the FIGO 2009 criteria. Previously, stage II endometrial cancer was divided into 2 subgroups based on endocervical glandular involvement (stage IIA) or invasion into the cervical stroma (stage IIB). The revised staging system eliminated these subgroups and classifies only tumors involving the cervical stroma as stage II. Tumors involving the endocervical glands are now classified as stage I and are subdivided based on the presence and depth of myometrial invasion.
The definition of stage III disease has also been refined based on the importance of prognostic features. Positive peritoneal cytology (previously stage IIIA) appears to worsen prognosis when combined with other poor prognostic features, but does not appear to be an independent poor prognostic feature. 65,66 For this reason, it was removed from the revised staging system. It is still an important part of the endometrial cancer staging, but should be reported separately. The presence of parametrial extension was added to stage IIIB. The classification of nodal involvement was also revised. Previously, stage IIIC encompassed the presence of either positive pelvic and/or para-aortic lymph nodes. However, several studies have shown that involvement of para-aortic lymph nodes carry a worse prognosis than involvement of pelvic lymph nodes alone.67–69 In a recent study of patients with stage IIIC disease, 5-year overall survival and recurrence-free survival with involvement of only the pelvic lymph nodes was 69.7% and 65.6%, respectively, compared with 48.8% and 44.4% when the para-aortic nodes were involved.68 For this reason, stage IIIC was subdivided into stage IIIC1 (pelvic lymph node involvement) and stage IIIC2 (para-aortic lymph node involvement with or without pelvic lymph node involvement).
In 2009, FIGO also developed a separate staging system for uterine sarcomas (reviewed in Chapter 7). The 1998 FIGO staging system was also felt to be inadequate by many, given that uterine sarcomas have histologic appearance, clinical behavior, and prognosis that are different from those of endometrial adenocarcinomas and should, therefore, not be grouped together in the same staging system.
1. Women with complex atypical hyperplasia have an increased risk of concurrent endometrial cancer, as well as the increased risk of progression to carcinoma. Although hysterectomy is considered standard treatment, progestin therapy may be an option in women who wish to preserve fertility.
2. Early-stage endometrial cancer is often curative with surgical resection alone. Adjuvant radiation therapy has not been shown to improve overall survival in patients with early-stage disease, but may have a role in reducing vaginal cuff recurrence or targeting occult nodal disease in patients at high risk for recurrence.
3. Controversies remain regarding the extent of surgical staging of endometrial cancer, primarily, which patients require lymphadenectomy and what is considered an adequate para-aortic lymph node dissection.
4. Primary radiation therapy is a viable option for patients who are considered to have medically inoperable disease. However, this treatment approach may not be as effective as primary surgery, particularly in stage II and/or high grade tumors.
5. Advanced-stage endometrial cancer is treated primarily with surgical resection (when feasible) and chemotherapy. Radiation therapy has a role in local control and in treating patients with positive lymph nodes. Hormonal therapies have been shown to be effective in patients with grade 1 tumors.
Primary Treatment Modalities
The mainstay of curative therapy for women with endometrial carcinoma is surgical resection, which includes complete hysterectomy, bilateral salpingo-oophorectomy, and comprehensive surgical staging. External pelvic radiotherapy and/or vaginal brachytherapy have not been shown to decrease mortality in early-stage disease. In certain cases, it is used to reduce the risk of vaginal cuff recurrence or to target occult disease in patients at high risk of disease recurrence. In patients who are considered inoperable secondary to significant medical comorbidities, external radiotherapy and/or intracavitary brachytherapy can be used as first-line treatment. The role of chemotherapy has evolved over the last several decades. Historically, chemotherapy was used in the treatment of recurrent disease; however, in the past decade, chemotherapy has played a larger role in the upfront and adjuvant treatment of patients with advanced (stage III and IV) endometrial cancer. Finally, hormonal therapy has also been used in a variety of settings, including in patients with early-stage and low-grade endometrial carcinoma who desire to preserve fertility, in patients with recurrent disease, and in patients with significant comorbidities who are not surgical candidates. Treatment recommendations for endometrial carcinoma by stage and histology are summarized in Tables 6-10 and 6-11.
Table 6-10 Treatment Summary for Endometrioid Endometrial Carcinoma After Comprehensive Surgical Staging