Evaluation of The Endometrium For Malignant or Premalignant Disease
Evaluation of The Endometrium For Malignant or Premalignant Disease
Evaluation of The Endometrium For Malignant or Premalignant Disease
premalignant disease
Author
Sarah Feldman, MD, MPH Section Editors
Deborah Levine, MD
Barbara Goff, MD
Robert L Barbieri, MD Deputy Editor
Sandy J Falk, MD
Last literature review version 17.1: January 2009 | This topic last updated: February 2,
2009 (More)
INVASIVE METHODS
limitations, numerous studies have shown that the endometrium is usually adequately sampled
with biopsy techniques:
(1) A meta-analysis of 39 studies involving 7914 women compared the results of endometrial
sampling with histopathology at D & C, hysteroscopy, and/or hysterectomy [5] . The significant
findings from this analysis were: Pipelle sampling was more sensitive for the detection of
endometrial cancer and atypical hyperplasia than all other sampling devices. The detection rates
for endometrial cancer (by Pipelle) in postmenopausal and premenopausal women were 99.6 and
91 percent, respectively. The detection rate for atypical hyperplasia was 81 percent.
The specificity for all devices was 98 to 100 percent. Fewer than 5 percent of patients had an
insufficient or no sample.
(2) Another meta-analysis assessed the accuracy of outpatient endometrial biopsy in the
diagnosis of endometrial cancer [6] . The posttest probability of endometrial cancer after a
positive test was 82 percent (95% CI, 60 to 93 percent), and after a negative test it was 0.9
percent (95% CI, 0.4 to 2.4 percent). In this analysis, 15 percent (138 of 945) of specimens were
inadequate (no sample or insufficient tissue for assessment) and one case of cancer was
subsequently found among these patients.
All sampling devices perform better when pathology is global rather than focal; this has been
demonstrated repeatedly in multiple studies. As an example, the importance of the volume of
disease was illustrated in a study of 65 patients with known cancer who underwent Pipelle biopsy
prior to hysterectomy [7] . Endometrial sampling was most reliable when at least one-half of the
endometrium was affected by disease. Five patients had tumors present only in an endometrial
polyp.
For this reason, additional endometrial assessment should be performed if abnormal uterine
bleeding persists after a "benign" endometrial biopsy [8] . Benign endometrial histology includes
atrophy (absence of a hormonal effect), proliferative endometrium (estrogen effect), secretory
endometrium (progestin effect), disordered or dyssynchronous endometrium (implies irregular
shedding of the endometrium secondary to unopposed estrogen), and endometritis. Further
endometrial assessment should also be considered when the endometrial biopsy is
nondiagnostic, but a high suspicion of cancer remains [9] . Nondiagnostic biopsies may be
associated with polyps, fibroids, endometrial cancer or other lesions occupying an area of the
uterus that was not sampled. Additional endometrial assessment may include a combination of
repeat endometrial biopsy or D & C hysteroscopy combined with transvaginal ultrasonography
(TVUS) with or without sonohysterography.
If endometrial biopsy does not yield sufficient tissue for pathological diagnosis, then the clinical
setting should dictate further management. We suggest TVUS if the woman is postmenopausal
and the bleeding has not been persistent; a thin endometrial stripe in this setting is most
consistent with atrophy and does not require further invasive studies. A thick endometrial stripe,
persistent bleeding, or bleeding in a post or perimenopausal woman should be followed by
additional endometrial sampling, such as hysteroscopy with curettage.
Dilation and curettage As discussed above, endometrial biopsy in the office has generally
replaced diagnostic D & C, but there is still a role for this procedure for some women. Indications
for diagnostic D & C include: When a patient is not able to tolerate an office endometrial biopsy
(eg, due to pain or anxiety) After a nondiagnostic office biopsy in women who are at high risk of
endometrial carcinoma After benign histology on office biopsy in women who have persistent
abnormal uterine bleeding When there is insufficient tissue for analysis on office biopsy When
cervical stenosis prevents the completion of an office biopsy When a concomitant operative
procedure, such as laparoscopy, is deemed necessary
Patients who have one of the indications above and who proceed to diagnostic D & C should also
be consented for diagnostic hysteroscopy at the same time. Multiple studies have shown that
hysteroscopy can aid in the detection of focal lesions of the endometrial lining which may be
missed by D & C alone [10,11] . However, in one large study of 1286 women, endometrial cancer
was missed in 10 women (34.5 percent) using hysteroscopy alone. Therefore, for women at risk
for endometrial cancer, we advise performing BOTH procedures (visualization and biopsy). Small
polyps can usually be removed at the same time.
The decision as to whether to perform a diagnostic hysteroscopy in the operating room or in the
office depends upon provider preference and equipment, the indications for the procedure, and
patient comfort/preference. The option of operative/therapeutic hysteroscopy should also be
discussed with the patient. Because therapeutic hysteroscopy involves greater risk than
diagnostic hysteroscopy [12] , its use should be reserved for women at low risk of endometrial
cancer, and in whom the value of hysteroscopic resection of a lesion is clear (ie, heavy bleeding
in a premenopausal woman who desires to maintain her fertility).
Patients in whom a diagnosis of cancer is possible should have a diagnostic procedure followed
by definitive treatment.
In women at risk for cancer, the risk of possible cancer dissemination due to intraperitoneal
spillage of the endometrial cells is discussed below. (See "Risk of tumor dissemination" below).
Transvaginal ultrasound
The American College of Obstetricians and Gynecologists and the Society of Radiologists advise
that either TVUS or endometrial biopsy are effective as a first diagnostic step in women with
postmenopausal bleeding [15,16] . Both diagnostic modalities have similar sensitivity for
detecting endometrial carcinoma.
This conclusion was supported by a literature review including data from almost 6000 women,
which found that the probability of cancer in a postmenopausal woman with vaginal bleeding and
a 10 percent pretest probability of endometrial cancer (the average risk associated with
postmenopausal bleeding) was 1 percent following a normal (<5 mm endometrial thickness)
TVUS [17] . Given this low risk of endometrial cancer, the authors suggested that endometrial
sampling might be unnecessary in women with a negative sonogram.
However, two other meta-analyses came to a different conclusion. One such meta-analysis with
nine studies (3483 women without endometrial cancer and 380 with the disease) found that the
detection rate for endometrial cancer varied according to the criteria for judging the endometrial
thickness as abnormal [18] . The detection rate was 63 percent (95% CI 58-69) when considering
an endometrial thickness that resulted in a false positive rate of only 10 percent; however, the
detection rate increased to 96 percent (95% CI 94-98) if a false positive rate of 50 percent was
considered acceptable. The authors concluded that even with a lower threshold for calling results
suspicious, 4 percent of endometrial cancers would be missed; thus, further invasive diagnostic
testing was indicated in all symptomatic women with a "thin" endometrial thickness. They also
noted that median endometrial thickness varied among centers such that 5 mm may not be a
universally reliable threshold. The other systematic review consisted of 57 studies with 9031
patients [19] . Only four studies used the 5 mm cut-off level, which was the best-quality
criteria. Using the pooled estimates from these four studies only, a positive test result raised the
probability of carcinoma from 14 to 31 percent, while a negative test reduced it to 2.5 percent.
The authors concluded that a negative result ( 5 mm) usually excluded endometrial pathology,
but ultrasound measurement of endometrial thickness alone could not be used to reliably
exclude cancer.
As noted above, most clinical investigators have used either 4 or 5 mm as the threshold for
further assessment. A detailed discussion of choosing an approach to postmenopausal bleeding
can be found separately. (See "The evaluation and management of uterine bleeding in
postmenopausal women").
Women on HRT TVUS does NOT appear to be a useful screening tool for excluding endometrial
hyperplasia/cancer in women on estrogen replacement therapy (ERT) that is unopposed or given
with cyclic progesterone [23,24] . Thickness thresholds are not well established for such women;
as a result, endometrial sampling is still the gold standard to exclude endometrial hyperplasia
and/or carcinoma.
For women on HRT therapy, TVUS should be obtained only for standard clinical indications, such
as to assess adnexal pathology, or if abnormal bleeding occurs and an endometrial biopsy
cannot be easily obtained. In the case of abnormal bleeding in women on cyclic progesterone, it
is best to obtain the TVUS early in cycle, when the endometrium is expected to be at its thinnest.
Persistent bleeding always requires endometrial biopsy regardless of ultrasound findings. We tell
patients that bleeding is common when ERT is initiated and should decrease over time. If it does
not, and if it becomes heavier, or bleeding occurs after a long period of no bleeding, then a
biopsy is indicated.
Women on tamoxifen Women on tamoxifen, both for treatment and prevention of breast
cancer, are known to have thickened endometrial stripes. The appearance can be cystic on TVUS
due to reactivation of foci of adenomyosis. There is no clear cut-off for normal versus
pathological endometrial thickness in these patients. In general, we do not feel there is a role for
baseline endometrial assessment, but women should have an endometrial biopsy if any
abnormal bleeding occurs, as tamoxifen confers a slight increase in endometrial cancer risk
(roughly 1/1000). Since tamoxifen is associated with formation of endometrial polyps,
sonohysterography can be helpful in these women to identify those with focal disease that would
be better triaged to hysteroscopy rather than blind biopsy.
We also sample postmenopausal women without uterine bleeding with lesser degrees of
endometrial thickening if they have endometrial fluid. In most of these women, endometrial fluid
is related to cervical stenosis [26-28] . Observational studies have consistently found that
asymptomatic postmenopausal women with endometrial fluid and an endometrial thickness less
than 3 mm do not have endometrial cancer or hyperplasia [26,28-31] . However, there is an
increasing risk of endometrial cancer in women with endometrial fluid and endometrial
thickening >3 mm; therefore, we biopsy these women.
Pelvic pain, a common reason for pelvic imaging, is not an indication for endometrial sampling in
the absence of abnormal uterine bleeding or suspicious sonographic findings. Premenopausal
women In asymptomatic premenopausal women, endometrial thickness alone is not an
indication for biopsy. Endometrial evaluation should be based on a combination of factors (show
table 1), including cervical cytology results showing glandular abnormalities or endometrial cells,
a chronic history of estrogen excess or anovulation (polycystic ovary syndrome), as well as
endometrial thickness. (See "Premenopausal women" above and see "Terminology and
evaluation of abnormal uterine bleeding in premenopausal women").
The disadvantage of saline infusion sonography is that no tissue is obtained for histological
diagnosis and it can be uncomfortable. The lack of histologic material for diagnosis was
addressed in another study that performed both endometrial biopsy and saline infusion
sonography in 113 women age 25 to 69 years with persistent AUB who subsequently underwent
hysteroscopy/curettage or hysterectomy [33] . The sensitivity, specificity, positive and negative
predictive values of biopsy/saline infusion sonography were 97, 70, 82, and 94 percent,
respectively. False positive results were due to intrauterine debris, blood clot, thickened
endometrial folds, and misidentified endometrial fragments. The combined technique of blind
biopsy and saline infusion sonography permitted diagnosis of the cause of abnormal uterine
bleeding in most women, without the need for more invasive procedures such as hysteroscopy
[34] . A subsequent meta-analysis came to this same conclusion [35] . The combined technique
is most useful in women with a symmetrically thickened endometrium. If saline infusion
sonography shows a focal lesion, hysteroscopy should be considered.
In randomized trials, outpatient hysteroscopy was associated with more pain than saline infusion
sonography [36-38] .
In summary, because of the lack of a definitive diagnosis, the potential discomfort of the
procedure, and the cost, we feel saline infusion should not be used for primary evaluation of
abnormal bleeding, but only in cases in which a diagnosis is unclear after biopsy, and for whom
there is a relative contraindication to proceeding to D & C/hysteroscopy.
RISK OF TUMOR DISSEMINATION One concern with hysteroscopy and saline infusion
sonography is that fluid flushed from the endometrial cavity through the fallopian tubes can
contain endometrial cells, some of which remain viable [39-51] . (See "Overview of
hysteroscopy" and see "Saline infusion sonohysterography").
The presence of tumor cells in peritoneal washings changes the stage of a woman with
endometrial cancer to at least stage IIIA (show table 4). However, a meta-analysis of
observational studies reported no significant difference in the frequency of positive peritoneal
cytology in women with endometrial carcinoma who had or had not undergone diagnostic
hysteroscopy (OR 1.64, 95%CI 1.0-28) [51] .
The clinical significance of iatrogenic upstaging is not known, as transport of these cells does not
necessarily result in implantation and persistence. In the few studies that have reported longterm follow-up, there was no detriment to survival associated with pre-staging
hysterosalpingography (with confirmed intraperitoneal spill) [52] or hysteroscopy [53] . Further
study is needed to clarify the clinical impact of tumor cell dissemination.
INFORMATION FOR PATIENTS Educational materials on this topic are available for patients.
(See "Patient information: Endometrial cancer diagnosis and staging" and see "Patient
information: Abnormal uterine bleeding"). We encourage you to print or e-mail these topics, or to
refer patients to our public web site www.uptodate.com/patients, which includes these and other
topics.
Initial evaluation Endometrial biopsy is preferable to ultrasound as the initial test for women
with abnormal uterine bleeding due to its high sensitivity, low complication rate, and low cost.
Endometrial biopsy is a more cost-effective initial approach than ultrasound when the prevalence
of endometrial carcinoma is at least 15 percent [54] .
TVUS is an acceptable alternative initial test in postmenopausal women who cannot tolerate
office biopsy, or in women who need concurrent evaluation of the adnexae. A summary of the
literature suggests that using an endometrial cutoff of <4 mm would yield a false negative rate
for endometrial cancer of 0.25 to 0.5 percent [55] , which compares favorably with the false
negative rate reported for endometrial biopsy.
Endometrial biopsy is required for histological diagnosis if the endometrium is not adequately
visualized, the stripe is 4 mm (focal or global), and in women with persistent bleeding
[14,55,56] . Blind biopsy is most accurate in women with a globally thickened endometrium;
visually directed sampling (ie, hysteroscopy) is preferable for women with focal abnormalities.
Persistent bleeding is worrisome even when the endometrial thickness is <4 mm [14,59,60] ,
particularly if there are other risk factors for endometrial cancer (show table 1 and show table 3).
Endometrial cancer can still occur in this setting since serous carcinoma of the endometrium may
arise from atrophic endometrium [61] . Therefore, further diagnostic evaluation is indicated.
Depending upon the prior evaluation, a combination of repeat endometrial biopsy or D & C
hysteroscopy combined with transvaginal sonohysterography or ultrasonography should be
pursued.
Evaluation for a nonendometrial source of bleeding, such as fallopian tube or ovarian cancer,
should also be pursued.
Relative risk
(RR)
NA
2 to 10
2
2
3
2 to 4
2
22 to 50
percent
lifetime risk
2/1000
NA
NA
NA
Adapted from data in Smith, RA, von Eschenbach, AC, Wender, R, et al. American Cancer Society
Guidelines for Early Endometrial Cancer Detection: Update 2001.