The importance of the Papanicolaou (Pap) smear in screening for cervical cancer is well established. Unfortunately, its usefulness in endometrial cancer screening is limited.³ Several studies have been performed to evaluate the use of Pap smears for endometrial cancer screening, but the average sensitivity is only about 43%.⁴ Several small studies report that abnormal Pap smears are found in 100% of papillary serous or clear cell endometrial cancer and in greater than 60% of endometrioid-type cancer.⁴ Another study showed a sensitivity of only 26% in detecting endometrial cancer overall; however, the authors note that the Pap smear may have improved sensitivity with higher grade disease.⁵ The huge variations in these findings make the clinical utility of the Pap smear in detection of endometrial cancer quite low.

Occasionally, cervical cytology samples show benign endometrial cells. Beal et al.⁶ showed that this occurred in approximately 0.9% of women 40 years of age or older; the majority of these women are completely asymptomatic.⁶ If benign endometrial cells are an
incidental finding in an otherwise asymptomatic premenopausal woman, it is exceedingly rare to find endometrial pathology.⁷ No further evaluation is indicated. However, if a woman is symptomatic with irregular or heavy vaginal bleeding, the finding of endometrial cells on Pap smear indicates the need for further investigation.⁷ In postmenopausal women, the finding of endometrial cells on Pap smear almost always necessitates endometrial sampling.⁸

The current American Society for Colposcopy and Cervical Pathology recommendations for atypical endometrial cells or abnormal glandular cells provide for further evaluation with colposcopy in conjunction with endocervical curettage, endometrial biopsy, and human papillomavirus testing. More often than not, cervical epithelial neoplasia is found in cases where atypical glandular cells are found on cervical cytology, but the endometrial cavity still warrants evaluation.⁸ Endometrial biopsy is indicated if a patient is less than 35 years of age and presents with vaginal bleeding that cannot be explained or has symptoms of chronic anovulation. If the patient is greater than 35 years of age, endometrial biopsy is almost always indicated with the finding of abnormal glandular cells but especially if atypical endometrial cells are noted on cytology. Less than 2% of women younger than 40 years of age have endometrial cells found on Pap smear requiring evaluation. Younger women rarely have pathology associated with the finding of endometrial cells on Pap smear; however, in menopausal women, it is critical to search for a pathologic cause.⁸

Sometimes, women who have had a hysterectomy for benign reasons have Pap smears of the vaginal cuff performed. If this occurs and benign glandular cells are found, no further investigation is needed.⁸ However, if a woman has a hysterectomy for cervical or endometrial cancer, Pap smear surveillance of the vaginal cuff is critical to evaluate for evidence of recurrence. Implications of Pap smear in evaluation of recurrent disease will be discussed later.

At time of initial presentation, a positive Pap smear is predictive of poor prognosis and future endometrial cancer recurrence. Histologic subtypes of papillary serous and clear cell endometrial cancer are more likely to be diagnosed after finding abnormal cells on cervical cytology than endometrioid-type endometrial cancer. The presence of abnormal endometrial cells on cervical sampling is also a predictor of poor prognosis, as indicated by the increased likelihood of severe myometrial invasion and nodal metastases.⁹ Increased grade of endometrial cancer regardless of histologic subtype is associated with having an abnormal Pap smear.⁹ Brown et al.⁹ determined that initial abnormal cytology leads to endometrial cancer recurrence in 12% of patients, as opposed to only 4% of patients with normal initial Pap smear findings. The same study showed that 19% of patients who were diagnosed with papillary serous or clear cell subtypes had abnormal cervical cytology compared with only 7% of endometrioid endometrial cancer patients.⁹

Endocervical curettage (ECC) is another method of evaluating women who present with abnormal cervical cytology or vaginal bleeding. ECC can be positive without any evidence of disease on imaging.² This finding may represent microscopic cervical involvement or possible contamination from the uterus.² ECC is included in a procedure called a fractional dilation and curettage (D&C), which attempts to isolate endometrial curettage from ECC. Distinguishing cervical involvement from endometrial contamination on fractional D&C is inaccurate. ECC is a poor method for evaluation of cervical extension of endometrial cancer because it has a false-positive rate of up to 50%.² False-negative results are also possible, so physicians must use other techniques to determine cervical involvement if there is high
suspicion.² Cervical involvement does not preclude surgical treatment, and often the presence of cervical involvement is only defined at the time of hysterectomy with frozen section pathology.

Endometrial sampling is the key step in the diagnosis of endometrial cancer. Premenopausal women who are anovulatory are at higher risk for endometrial neoplasia and require endometrial sampling to work-up abnormal uterine bleeding. All postmenopausal women with vaginal bleeding require endometrial sampling. Ten percent of postmenopausal women who present with this classic symptom will have cancer on biopsy.² Combined risk factors of age greater than 70 years, type 2 diabetes, and nulliparity increase risk of neoplasia in women with symptomatic bleeding.² Eighty-seven percent of these patients are diagnosed with atypical hyperplasia or endometrial cancer on biopsy.² Three percent of women with symptomatic bleeding and none of these risk factors are diagnosed with endometrial cancer on endometrial biopsy.² Accurate history and physical prior to endometrial sampling allows a physician to appropriately counsel patients.

Endometrial biopsy is often performed in an outpatient setting with minimal discomfort to the patient and no need for anesthesia. Adequate sampling is a must, and if not obtained, further investigation with more invasive techniques is warranted. Endometrial biopsy in the United States is most often performed with a disposable plastic endometrial suction curette (e.g., Pipelle; CooperSurgical, Trumbull, CT). The patient is placed in lithotomy position, and after placement of a speculum, a curette is inserted through the cervix into the uterine cavity, and a biopsy specimen is obtained.

Risks of endometrial biopsy include pain and bleeding. Infrequently, uterine perforation may occur, but this is almost always benign and does not require any action other than observation. Rarely, a specimen is unobtainable due to cervical stenosis. Endometrial biopsy in the outpatient office setting is a blind procedure and can produce a false-negative result due to lack of tissue representing the entire endometrium. Positive tests are diagnostic, but negative tests often require further endometrial cavity evaluation.¹⁰

Endometrial biopsy collection fails in approximately 8% of cases when a Pipelle device is used, according to a literature review by Clark et al.¹⁰ They evaluated 11 studies with more than 1,000 patients and found inadequate samples in up to 13% of specimens collected via the Pipelle and 15% of specimens for all methods.¹⁰ Other tools used to collect endometrial samples are the Vabra Aspirator (Berkeley Medevices, Richmond, CA) and the Z sampler device (Zinanti, Chatsworth, CA). The Vabra Aspirator has been found to be less accurate overall in diagnosing hyperplasia and endometrial cancer but is still widely used.³

In postmenopausal women, physicians may be unable to perform office endometrial sampling, and even when a sample is obtained, the specimen may still prove inadequate for analysis.¹⁰ Clark et al.¹⁰ found an overall failure rate of 12% but a higher inadequate sampling rate of 22% for all methods in postmenopausal women. The authors concluded that outpatient diagnosis via endometrial biopsy is a valid and valuable tool in the diagnosis of endometrial cancer, especially when negative results are further evaluated in cases where clinical suspicion is high.

The strength of Pipelle endometrial sampling is further demonstrated by Fakhar et al.,¹¹ who evaluated the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of endometrial biopsies in 100 patients. D&C following the biopsy served as the “gold standard” for comparison.¹¹ If Pipelle sampling indicated endometrial carcinoma,
endometrial hyperplasia, or secretory endometrium, then sensitivity, specificity, PPV, and NPV were each 100%.¹¹ For endometrial hyperplasia with atypia, sensitivity and NPV remained 100%, whereas specificity slightly decreased to 98% and PPV decreased to 80%.¹¹ These results further authenticate the use of outpatient endometrial sampling with a Pipelle device as a cost-effective, low-risk option with excellent sensitivity and specificity.

Office endometrial biopsy does have limitations in cases where an endometrial polyp is present. Tanriverdi et al.¹² suggest that if a polyp is suspected, D&C is necessary. They also confirmed previous findings that endometrial biopsy in postmenopausal women may produce an inadequate sample in up to 22% of cases and again suggest D&C in such cases.¹² They did not find the high 100% sensitivity and specificity for endometrial carcinoma found by Clark et al.,¹⁰ but they did evaluate the combination of transvaginal ultrasound and Pipelle biopsies, finding close to 100% sensitivity and specificity when these methods are combined.^11,12 Tanriverdi et al.¹² also proposed that outpatient biopsy be used only in low-risk patients without high suspicion of hyperplasia with atypia or endometrial cancer and that patient with high risk for cancer be more carefully and completely evaluated with formal endometrial curettage. Most physicians prefer an attempt at outpatient sampling in all patients, with appropriate follow-up if suspicion remains elevated despite a negative result.

Routine screening for endometrial cancer does not exist, but certain groups warrant some method of evaluation. Patients with known HNPCC should undergo annual endometrial biopsies starting at age 35 according to the American Cancer Society.² HNPCC increases the lifetime risk of endometrial cancer from 3% to 20%.² Some physicians advocate that women on tamoxifen treatment for breast cancer should undergo endometrial sampling for screening, whereas most physicians advise prompt evaluation only in symptomatic women.² Most oncologists and other authorities do not advocate routine evaluation with ultrasound or endometrial biopsy in women on tamoxifen.² The American Cancer Society does not recommend endometrial cancer screening for women on tamoxifen.²

D&C is the gold standard for diagnosis of endometrial cancer.³ A D&C requires general or regional anesthesia in an operating room and is often performed with hysteroscopy to completely evaluate the endometrial cavity. Women with endometrial polyps, cervical stenosis, or insufficient sampling on endometrial biopsy who remain at high suspicion for endometrial cancer require endometrial curettage in the operating room. For patients contemplating medical therapy of endometrial cancer to preserve future fertility, complete evaluation of the endometrium with curettage is crucial.¹ Hysteroscopy involves using a camera inserted into the cervix after adequate dilation. The entire endometrial cavity may be visualized, and directed biopsies may be taken.

The accuracy of office endometrial sampling versus endometrial curettage has been compared, and results are mixed. Larson et al.¹³ showed that D&C correctly identified the tumor grade of 77% of endometrial cancers compared with 58% of office samples taken with the Z sampler, one device for outpatient endometrial sampling. Often, the cancers were undergraded (i.e., given a lower tumor grade than the grade found at hysterectomy).¹³ Preoperatively, because there is no definitive way to evaluate depth of invasion, tumor grade is of utmost importance to guide the decision of whether or not to perform complete lymphadenectomy for surgical staging.¹³ For example, patients with grade 3 endometrial carcinoma, regardless of myometrial invasion, would undergo surgical staging with both pelvic and paraaortic
lymphadenectomy due to a higher risk of nodal metastasis. In contrast, a patient with grade 1 endometrioid endometrial cancer may be managed with a more conservative surgical procedure.¹³

Hysteroscopy is often performed with D&C. Ben-Yehuda et al.¹⁴ showed that hysteroscopy does not increase the diagnostic accuracy of the endometrial curettage. They reviewed medical records of 403 patients who underwent both uterine curettage and hysteroscopy and noted that there was only 52% concurrence in the diagnosis of hyperplasia via hysteroscopic impression versus final pathology findings. Only 10 patients in their study had endometrial cancer found on D&C, but hysteroscopic impression of carcinoma was only noted in the operative notes of 2 of these patients. Two patients in their study had endometrial cancer diagnosed within 6 months but were not diagnosed with either modality, refuting the idea that hysteroscopy might catch some cancers that curettage may miss¹⁴ (Figs. 5-2,5-3 and 5-4).

In benign gynecologic conditions, when hysteroscopy is used to detect polyps and submucosal leiomyomata, research shows that diagnosis of abnormal uterine bleeding does improve with hysteroscopy combined with endometrial curettage.¹⁴ Uterine curettage was only able to detect 32% of polyps and 6% of leiomyomata in this review, confirming that hysteroscopy aids in diagnosis in some situations.¹⁴ The review by Ben-Yehuda et al.¹⁴ of 403 patients identified three uterine perforations, none of which necessitated exploratory laparotomy. Occasionally, diagnostic laparoscopy may be necessary after uterine perforation to evaluate the intra-abdominal organs for damage.¹⁵ Complications due to general or regional anesthesia can also be encountered during hysteroscopy. In addition, there is the added possibility of complication from the distending media used to visualize the endometrial cavity.¹⁴

Distending media can be isotonic or hypotonic fluid or carbon dioxide gas. Most gynecologists prefer to use normal saline to distend the uterine cavity to improve visualization. If uterine perforation is encountered, normal saline will not harm the patient unless a large amount is deposited. Pulmonary or cerebral edema from extensive intravascular absorption is a rare complication of excess fluid balance after hysteroscopy, and the patient should be observed postoperatively if this is encountered. Hypotonic fluids such as glycine and sorbitol can lead to electrolyte abnormalities with a relatively minimal amount of positive fluid balance.¹⁵ Accurate assessment of fluid deficits is important in preventing these complications. Carbon dioxide may also be used as a distending media. If the intrauterine pressure is allowed to get too high, venous gas embolism may occur because the gas is forced into the cervical or endometrial vasculature.¹⁵