Transvaginal Ultrasound: Benefits

There is no evidence that screening by ultrasonography (e.g., endovaginal ultrasound or transvaginal ultrasound) reduces mortality from endometrial cancer. Most cases of endometrial cancer (85%) are diagnosed at low stage because of symptoms, and survival rates are high.

Transvaginal Ultrasound: Harms

Based on solid evidence, screening asymptomatic women will result in unnecessary additional biopsies because of false-positive test results. Risks associated with false-positive tests include anxiety and complications from biopsies.

• Study Design: Evidence obtained from cohort studies.
• Internal Validity: Fair.
• Consistency: One study for endometrial biopsy and one study for hysteroscopy.
• Magnitude of Effects on Health Outcomes: Small negative magnitude.
• External Validity: Fair.

Endometrial Sampling (Biopsy): Benefits

There is inadequate evidence that screening by endometrial sampling (i.e., biopsy) reduces mortality from endometrial cancer. Most cases of endometrial cancer (85%) are diagnosed at low stage because of symptoms, and survival rates are high.

Endometrial Sampling (Biopsy): Harms

Based on solid evidence, endometrial biopsy may result in discomfort, bleeding, infection, and rarely, uterine perforation.

• Study Design: Evidence obtained from cohort studies.
• Internal Validity: Fair.
• Consistency: One study for endometrial biopsy and one study for hysteroscopy.
• Magnitude of Effects on Health Outcomes: Small negative magnitude.
• External Validity: Fair.

Epidemiology of Endometrial Cancer

Incidence and mortality

Endometrial cancer is the most common invasive gynecologic cancer in U.S. women, with an estimated 67,880 new cases expected to occur in 2024 and an estimated 13,250 women expected to die of the disease. Endometrial cancer is primarily a disease of postmenopausal women, with a mean age at diagnosis of 60 years. Age-adjusted endometrial cancer incidence in the United States increased from the mid-1960s to 1975 and then declined from 1975 to 1980, with a transient increase in incidence occurring from 1973 to 1978, which was associated with estrogen therapy, also known as hormone therapy. There was no associated increase in mortality. Continuing with more recent trends, incidence rates increased by about 1% per year in White women and by 2% to 3% per year in women of all other racial and ethnic groups. Since the mid-2000s, death rates for endometrial cancer increased by 1.7% per year. Most cases of endometrial cancer are diagnosed because of symptoms, which are nonetheless early stage and have high survival rates.

Risk Factors

Estrogen therapy unopposed by progesterone therapy is a cause of endometrial cancer in women with an intact uterus. However, women taking combination estrogen-progesterone therapy (hormone therapy) exhibit similar risk to women who do not take postmenopausal hormone therapy. Tamoxifen therapy is also a cause of endometrial cancer. Results from the National Surgical Adjuvant Breast and Bowel Project P-1 trial, report a doubling of the risk of endometrial cancer associated with an annual rate of 2.30 per 1,000 for women taking tamoxifen compared with 0.91 per 1,000 for women receiving placebo; the increased risk was seen primarily in postmenopausal women.

In addition to the increased risk of developing endometrial cancer that is observed in women who use unopposed estrogen therapy or tamoxifen, a number of additional risk factors have been identified, and most appear to be related to estrogenic effects. Among these factors are obesity, a high-fat diet, and reproductive factors such as nulliparity, polycystic ovary syndrome, early menarche, and late menopause. Hereditary nonpolyposis colorectal cancer (HNPCC) syndrome is associated with a markedly increased risk of endometrial cancer compared with women in the general population. Among women who are HNPCC carriers, the estimated cumulative incidence of endometrial cancer ranges from 20% to 60% by age 70 years (for more information, see Genetics of Colorectal Cancer). This risk appears to differ slightly based on the germline mutation; for MLH1 carriers the lifetime risk at age 70 years is 25% while MSH2 mutation carriers have a 35% to 40% lifetime risk of endometrial cancer by age 70 years. The mean age of diagnosis for MLH1 or MSH2 carriers is 47 years compared with 60 years for noninherited forms of endometrial cancer. The prognosis and survival are similar between HNPCC-related and noninherited forms of endometrial cancer.

Major differences exist between Black and White women in stages of endometrial cancer at detection and at subsequent survival. Although the incidence of endometrial cancer is lower among Black women, mortality is higher. The National Cancer Institute initiated a Black/White Cancer Survival Study and concluded that higher-grade and more aggressive histologies appear to be related to excess risk of advanced-stage disease in Black women. It is difficult to disentangle the effects that biology and socioeconomic status may have on the lower survival rates of African American women with endometrial cancer. Evidence suggests that lower income is associated with advanced-stage disease, lower probability of undergoing a hysterectomy, and lower survival rates. Others, however, assert that there is no Black/White racial difference in the interval from patient-reported symptoms to initial medical consultation, making it unlikely that patient delay after onset of symptoms could explain much of the excess of advanced-stage disease found in Black women. Further research is necessary to understand why Black women tend to be diagnosed with more aggressive disease and have a higher probability of dying than White women, despite their lower incidence of endometrial cancer.

Measuring endometrial thickness (ET) with transvaginal ultrasound (TVU) and endometrial sampling with cytological examination have been proposed as possible screening modalities for endometrial cancer. The Papanicolaou (Pap) test, used successfully for screening for cervical cancer, is too insensitive to be used as a screening technique for the detection of endometrial cancer, although occasionally the Pap test may fortuitously identify endometrial abnormalities, such as endometrial cancer.

Routine screening of asymptomatic women for endometrial cancer has not been evaluated for its impact on endometrial cancer mortality. Although high-risk groups can be identified, the benefit of screening in reducing endometrial cancer mortality in these high-risk groups has not been evaluated. Using the same cutoffs to define an abnormal ET in asymptomatic women as used in symptomatic women would result in large numbers of false-positive test results and larger numbers of unnecessary referrals for cytological evaluations. Published recommendations for screening certain groups of women at high risk of endometrial carcinoma are based on opinion regarding presumptive benefit.

Modalities of Endometrial Cancer Screening

Ultrasonography in women with vaginal bleeding

TVU is used as a diagnostic tool to evaluate symptomatic women with vaginal bleeding. Among women with postmenopausal uterine bleeding and cancer, 96% will have an abnormal ET (>6 mm). The specificity varies by whether women used hormone therapy. Among nonusers, the specificity was 92%. Much less work has been done to evaluate the accuracy of TVU among asymptomatic women. If the same ET cutoff is used among asymptomatic women, the false positives will be extremely high, resulting in a very low positive predictive value. No studies have evaluated the efficacy of screening with TVU in reducing mortality from endometrial cancer.

A group of researchers used dilation and curettage (D&C) as a gold standard, to evaluate TVU measurement of ET as a predictor of endometrial cancer in women reporting postmenopausal bleeding (PMB) (estrogen-progestin therapy and nonhormone therapy users). Of the 339 participants, 39 (11.5%) were diagnosed with endometrial cancer (four had an ET of 5–7 mm and 35 had an ET >8 mm) based on histopathology from curettage. No cancers were detected in women with an ET of less than 4 mm. Using a cutoff point of 4 mm, TVU has 100% sensitivity and 60% specificity. In this population, 46% (156) of the women had an ET greater than 4 mm.

Ultrasonography in women without vaginal bleeding

A comparison of TVU and endometrial aspiration was conducted among asymptomatic postmenopausal women potentially eligible for an osteoporosis prevention trial as part of determination of eligibility for randomization. TVU was performed on 1,926 women. Of these, 93 women had ET greater than 6 mm. Among the 93 women with abnormal ET, 42 had endometrial aspiration with one finding of abnormal pathology (defined as adenocarcinoma or atypical hyperplasia). Of the 1,833 women with ET measuring 6 mm or less, 1,750 women had endometrial aspiration and five of these women had an abnormal pathological biopsy. Among this population of asymptomatic postmenopausal women, the estimated sensitivity for TVU with a threshold value of 6 mm was 17% and 33% for a threshold value of 5 mm.

One study assessed the usefulness of TVU among a cohort of postmenopausal, asymptomatic women receiving hormone therapy. Utilizing the Postmenopausal Estrogen and Progestin Interventions Trial participants who had undergone both TVU and endometrial biopsy, sensitivity, specificity, positive predictive value, and negative predictive value were determined for women who received placebo, estrogen alone, and estrogen-progestin therapy. At a threshold value of 5 mm for ET, TVU had 90% sensitivity and 48% specificity. Using this threshold, more than half the women would receive a biopsy while only 4% of them had serious disease.

Another study obtained endometrial biopsy specimens from 801 asymptomatic perimenopausal and postmenopausal women prior to enrollment in a hormone therapy study. Of the specimens, 75% of the samples contained sufficient tissue for diagnosis. Among these women, one case of endometrial cancer was diagnosed, illustrating the low yield of screening among asymptomatic women and the difficulty with endometrial cavity access.

Although TVU can be used to evaluate asymptomatic and occult endometrial pathology, the technique has not been evaluated as a screening method for reducing mortality in asymptomatic women.

Ultrasonography in women using tamoxifen

Tamoxifen is widely used as part of adjuvant therapy for breast cancer and as chemoprevention for women at increased risk of breast cancer. In addition to the protective effects for breast cancer, the biological and endocrine effects of tamoxifen increase patients’ risk of developing endometrial pathology, including endometrial polyps, endometrial hyperplasia, and endometrial carcinoma.

There is interest in trying to reduce the morbidity from endometrial cancer through early detection, and there has been interest in using endovaginal ultrasound as a method to screen women to detect endometrial cancer.

In a prospective, observational study of 304 women using tamoxifen over 6 years, women underwent annual endovaginal ultrasound screening; women with abnormal ultrasound findings and women who were symptomatic with bleeding all underwent endometrial biopsy. Thirty-two percent of the ultrasound examinations had associated significant uterine abnormalities identified that required further medical or surgical investigation and treatment. However, most abnormalities (80%) represented benign polyps for which no treatment was needed. Six cases of primary endometrial cancer were detected, and all cases presented with irregular bleeding. The sensitivity of ultrasound was only 63.3%, with a specificity of 60.4%, and had a low positive predictive value for cancer of only 1%.

Other reports have noted similar results. Routine ultrasound surveillance in asymptomatic women using tamoxifen is not useful because of its low specificity and low positive predictive value. Evaluation of the endometrium in women taking tamoxifen should be limited to women symptomatic with vaginal bleeding.

Sonohysterography

Sonohysterography (hydrosonography) is a diagnostic test used to help guide biopsies in asymptomatic women that is able to separate space occupied by endometrial lesions from an abnormal endometrial-myometrial junction. There is no evidence that routine screening sonohysterography will confer clinical benefit.

Endometrial sampling in women with uterine bleeding

In the setting of abnormal uterine bleeding, endometrial sampling has gained favor largely as an alternative to more invasive procedures such as fractional D&C. Several methods of biopsy exist (e.g., Pipelle, Tao Brush, and Uterine Explora Curette) to identify endometrial pathology. Although endometrial sampling has largely replaced D&C as the first choice in the evaluation of women with bleeding, issues of access to the endometrial cavity and sampling error limit the clinical significance of a negative result. In the Arimidex, Tamoxifen, Alone or in Combination trial, 36% of biopsies had insufficient tissue for diagnosis. A meta-analysis of PMB reported that 91% (95% confidence interval , 87%–93%) of women with endometrial cancer reported PMB. However, among women with PMB, only 9% (95% CI, 8%–11%) were diagnosed with endometrial cancer. This report is limited by a lack of histology-specific estimates.

No studies have evaluated the use of endometrial sampling as routine screening in reducing endometrial cancer mortality.

Hysteroscopy

Hysteroscopy is used in the office setting to directly visualize the uterine cavity. A group of researchers noted that hysteroscopy is not as useful in detecting endometrial cancer as biopsy or D&C. It has not been evaluated as a screening tool.

Hormone Therapy

There is no evidence to suggest that screening women prior to or during estrogen-progestin therapy, also known as hormone therapy, would decrease endometrial cancer mortality. Thus, women on hormone therapy should have a prompt diagnostic work-up for abnormal bleeding. Although women using certain hormone regimens have an increased risk of endometrial cancer, most women who develop cancer will have vaginal bleeding. There is no evidence that screening these women would decrease mortality from endometrial cancer.

Hereditary Nonpolyposis Colorectal Cancer

The lifetime risk of endometrial cancer for women with hereditary nonpolyposis colorectal cancer (HNPCC) and for women who are at high risk for HNPCC is as high as 60%. These cases are often diagnosed in the fifth decade, 10 to 20 years earlier than sporadic cases. Based on limited evidence, it appears that 5-year survival among women with HNPCC diagnosed with endometrial cancer is similar to that of nonhereditary cases in the general population. Because the risk of endometrial cancer is so high among these women, international guidelines suggest gynecologic surveillance including annual transvaginal ultrasound with endometrial biopsy for women aged 25 to 35 years. The most recent American Cancer Society Cancer Detection Guidelines (updated January 2005) recommend annual screening with endometrial biopsy beginning at age 35 years.

Women Treated With Tamoxifen

The risk of endometrial cancer is increased in women who are treated with tamoxifen and is even greater in the subset of women who have a history of prior estrogen therapy. Beyond a routine gynecologic examination eliciting any history of abnormal bleeding, it has been recommended that screening studies and procedures for detecting endometrial pathology in women taking tamoxifen should be left to the discretion of the individual gynecologist. Commonly, there are endometrial abnormalities in women taking tamoxifen, especially in false-positive endovaginal ultrasound screening tests. More importantly, any abnormal uterine bleeding should be completely evaluated.

Endometrial cancers that occur in tamoxifen-treated women are very similar to those cancers occurring in the general population, with respect to stage, grade, and histology. Prognosis is good and not affected by early detection.

There have been no published studies evaluating the effect of endometrial cancer-screening modalities on mortality among women taking tamoxifen for breast cancer treatment or prevention.

Abnormal ultrasound typically requires further investigation including endometrial biopsy (sampling). The evidence is solid that endometrial sampling may result in discomfort, bleeding, infection, and rarely uterine perforation. A study designed to evaluate performance, patient acceptance, and cost-effectiveness of blind biopsy, hysteroscopy with biopsy, and ultrasound, in 683 women with vaginal bleeding, reported that minor events, including discomfort and distress, occurred in 16% of women who had hysteroscopy with biopsy, and in 10% of the women who had a blind biopsy. A group of researchers studied 13,600 diagnostic and operative hysteroscopic procedures and found a lower complication rate among diagnostic procedures (0.13%) compared with operative procedures (0.28%). Risks associated with false-positive test results include anxiety and additional diagnostic testing and surgery. Endometrial cancers may be missed on endometrial sampling and ultrasound.