Complex atypical hyperplasia (CAH) of the endometrium is considered the precursor for endometrioid endometrial cancer, the most common gynecologic cancer in the United States. This disease is most frequently diagnosed by gynecologists who are evaluating symptoms of abnormal uterine bleeding in premenopausal women or in postmenopausal women who experience new bleeding. Medical therapies, typically progestin-based treatments, can be employed, particularly when fertility preservation is desired or among patients who are poor surgical candidates. However, the most definitive therapy remains surgery with total hysterectomy for two reasons: CAH is associated with a 28% risk for the development of invasive cancer, and occult invasive cancer frequently coexists with CAH.1,2 This raises a question for gynecologists: Given the risk for occult endometrial cancer, should patients be referred to a gynecologic oncologist for their surgery?

What is the risk for cancer?

Approximately 43% of patients with a preoperative diagnosis of CAH will have invasive cancer diagnosed on their hysterectomy specimen.2 In the majority of these cases these are low grade and minimally invasive tumors at low risk for lymph node metastases. However, approximately 12% are associated with deeply invasive, high grade tumors. Lymph node metastases have been observed in approximately 7% of patients with preoperative CAH who were staged at the time of hysterectomy.3

What is the significance of occult malignancy with CAH?

If surgeons are aware of endometrial cancer preoperatively or intraoperatively, decisions can be made about staging, particularly the need for lymphadenectomy. The virtues of staging in endometrial cancer is a controversial and frequently debated topic. No survival (therapeutic) benefit from lymphadenectomy has been observed in prospective trials when the information from staging results is not used to guide adjuvant therapy.4 However, the administration of adjuvant chemotherapy is associated with improved survival for patients with lymph node metastases.5 Therefore, if there is a benefit to staging with lymphadenectomy, it is its ability to identify patients who most need this life-saving systemic therapy.

Not all patients with endometrial cancer are at equal risk for harboring lymph node metastases and the majority may not benefit from lymphadenectomy. Patients with tumors that are deeply invasive, moderate or high grade, larger than 2 cm, or that have lymphovascular space invasion are at higher risk for lymph node metastases. Women with low grade, minimally invasive tumors that are smaller than 2 cm have extremely low risk for metastases.6 These criteria are commonly employed to stratify women at lowest risk and minimize unnecessary lymphadenectomy procedures. It should be noted that all three of these low risk features must be present to convey that negligible risk profile. The finding of a grade 1 invasive tumor alone is not enough to exclude potential lymph node metastases, particularly in the case of large or deeply invasive cancers.

How can the diagnosis be made preoperatively or intraoperatively?

The gold standard for discriminating between CAH and endometrial cancer is definitive surgical pathology. However, if surgeons wait until these results are available, they have lost the opportunity to stage the patient without subjecting them to a second surgery. The preoperative discovery of cancer may be increased by performing diagnostic curettage rather than relying on office endometrial biopsy sampling.7 This is likely due to the increased volume of tissue removed with dilation and curettage, and a reduction in the risk for sampling error. The addition of hysteroscopy to curettage does not improve upon the detection of cancer. Preoperative MRI to evaluate for depth of myometrial invasion has been described in cases of known endometrial cancer; however, its role in discriminating between CAH and invasive cancer is not well studied.

Intraoperative frozen section is commonly employed to evaluate the hysterectomy specimen for cancer in order to triage patients to staging during that same surgery. However, the accuracy of frozen section with definitive pathology is only approximately 50%.8 This means that at least half of women with CAH will have a false negative frozen section result and will have lost the opportunity for staging at the same procedure. The inaccuracy of frozen section is often overlooked by surgeons who may feel that it is a very straightforward diagnostic procedure. In reality, the characterization of CAH and invasive cancer is technically challenging and relies on multiple sectioning and significant experience in gynecologic pathology.9

Should all patients with CAH be referred and staged?

An alternative to relying on the frozen section process and its inherent inaccuracies would be to routinely stage all women with CAH, knowing that approximately 40% of them have occult cancer, and more than a third of those will have high risk features for lymph node metastases. However, due to the risks associated with lymphadenectomy, particularly lymphedema, most gynecologic oncologists do not routinely stage patients with preoperative CAH with complete lymphadenectomy.

An alternative to the all (complete lymphadenectomy) or none (hysterectomy alone) approach is to perform sentinel lymph node (SLN) biopsy for patients with CAH. SLN biopsy involves removing scant, but high yield lymphatic tissue, and has been shown to be extremely sensitive in detecting metastatic disease.10 This approach is commonly employed by surgeons in the treatment of ductal carcinoma in situ of the breast which, like CAH, is a stage 0 cancer that can be associated with invasive carcinoma on final pathology. In the case of ductal carcinoma in situ, the risk for upstaging is actually substantially lower (25%) than what is observed in CAH.11 Therefore, it would seem even more compelling to apply this approach for endometrial pathologies. The ability to apply the SLN technique is lost after hysterectomy is performed, as there is no longer the target organ into which tracer can be injected; therefore, if SLN biopsy is to be offered to these patients, it needs to be performed using only the preoperative diagnosis of CAH. In this approach, there will be overtreatment of approximately two-thirds of patients, albeit with a less radical and morbid staging procedure.

Making the decision to refer

Ultimately, decisions to refer or not are guided by comprehensive discussions between patient and provider that outline the potential risks and benefits of various approaches. Patients frequently have strong relationships with confidence in their gynecologists who may have cared for them for many years, and may be motivated to have them perform their surgery. For others, the uncertainty and possibility of an unstaged cancer and the potential of a second surgery drives their decision to seek an oncology consultation. Clinicians should discuss the inherent uncertainties in the diagnosis of CAH and the potential for underlying cancer and lymph node metastases, and help patients determine the balance of their underlying competing concerns regarding the risk for inadequate surgery versus the risk of unnecessary surgical procedures.

Summary of recommendations

Invasive endometrial cancer will be identified in the hysterectomy specimens of approximately 40% of women with a preoperative diagnosis of complex endometrial hyperplasia. Preoperative dilation and curettage may reduce the potential for missed occult cancer. Frozen section is an option for determining which patients might benefit from staging but is associated with significant inaccuracies. Failure to diagnose malignancy pre- or intraoperatively handicaps postoperative decision making regarding the necessity of adjuvant chemotherapy, and prevents the ability to offer patients potentially less morbid staging techniques such as SLN biopsy. When gynecologists without oncology training perform these hysterectomies, they should discuss these scenarios to patients and consider referral to gynecologic oncology for patients who desire the potential for comprehensive staging if necessary.

Dr. Rossi is an assistant professor in the division of gynecologic oncology at the University of North Carolina, Chapel Hill. She reports no relevant financial disclosures.


1. J Clin Oncol. 2010 Feb;28:788-92 .

2. Cancer. 2006 Feb;106:812-9 .

3. Int J Gynecol Cancer. 2005 Jan-Feb;15:127-31 .

4. Lancet. 2009 Jan;373(9658):125-36 .

5. J Clin Oncol. 2006 Jan;24:36-44 .

6. Gynecol Oncol 2008 Apr;109:11-8 .

7. Am J Obstet Gynecol. 2010 Oct;203(4):349. e1-6 .

8. Am J Obstet Gynecol. 2007 May;196(5):e40-2 .

9. Obstet Gynecol. 2012 Nov;120(5):1160-75 .

10. Lancet Oncol. 2017 Mar;18(3):384-92 .

11. Radiology. 2011 Jul;260:119-28 .


You May Also Like

Half of newly detected antimicrobial antibodies do not lead to PBC

FROM HEPATOLOGY Nearly half of newly detected antimitochondrial antibodies (AMAs) in clinical practice do ...

Candida colonization raises risk of Acinetobacter-based VAP

FROM MEDICAL MYCOLOGY Acinetobacter baumanii was the most common cause of ventilator-associated pneumonia in ...