International Journal of Gynecological Cancer:
Outcomes of Women With Atypical Glandular Cells on Preoperative Cytology and Endometrial Cancer
Guy, Michael S. MD*; Cheng, Georgina MD, PhD*; Post, Miriam D. MD†; Spillman, Monique A. MD, PhD*; Behbakht, Kian MD*; Davidson, Susan MD*; Sheeder, Jeanelle MSPH, PhD*; Guntupalli, Saketh R. MD*
*Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, and †Department of Pathology, University of Colorado School of Medicine, Aurora, CO.
Address correspondence and reprint requests to Michael S. Guy, MD, Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, University of Colorado School of Medicine, Academic Office 1, Room 4409, B198-4 12631 E. 17th Ave, Aurora, CO 80045. E-mail: firstname.lastname@example.org.
The authors declare no conflicts of interest.
Received September 8, 2013
Received in revised form October 9, 2013
Accepted October 10, 2013
Objective: This study aims to examine the prognostic importance of preoperative cervical cytologic diagnosis with atypical glandular cells (AGC) or malignant cells (MC) as a predictor of poor outcomes in endometrial cancer.
Materials and Methods: A total of 563 patients were surgically staged for endometrial adenocarcinoma from 2002 to 2012 at our institution. Of these patients, 106 were included to perform a case-control study (39 patients with AGC or MC and 67 controls). Included patients were not significantly different from excluded patients and were matched for age, race, and body mass index. Outcome variables included presence of extrauterine disease (International Federation of Gynecology and Obstetrics stage ≥II) and high intermediate risk (HIR) disease. Further analysis sought to improve the prediction combining AGC or MC with other factors, such as grade and CA-125 levels. Standard statistical analyses were used.
Results: Among the patients with AGC or MC, 53.8% had HIR disease compared with 30.3% with normal cervical cytologic diagnosis (odds ratio [OR], 2.68; 95% confidence interval [CI], 1.18–6.09; P = 0.02). Extrauterine disease was found in 43.6% of patients with AGC or MC compared with that of 15.2% in patients with normal cervical cytologic diagnosis (OR, 4.33; 95% CI, 1.72–10.90; P < 0.01). Multivariate analysis confirmed that AGC or MC was an independent predictor of HIR disease (OR, 8.41; 95% CI, 1.34–52.78; P = 0.02) and extrauterine disease (OR, 4.78; 95% CI, 1.26–18.1; P = 0.02). The combination of elevated CA-125 levels with AGC or MC cervical cytologic diagnosis increased the statistical prediction of extrauterine disease (OR, 13.3; 95% CI, 3.1–56.8; P < 0.01) and HIR disease (OR, 5.83; 95% CI, 1.44–23.71; P = 0.02).
Conclusions: Patients with AGC or MC on preoperative cervical cytology are at risk for extrauterine and HIR disease. These preoperative findings should warn surgeons of the potential of extrauterine or occult metastatic disease.
Endometrial cancer is the most common gynecologic malignancy in the United States with 49 560 cases and 8190 deaths estimated in 2013.1 One of the most important prognostic factors for endometrial cancer is extrauterine disease. Gynecologic Oncology Group (GOG) protocol 33 demonstrated that preoperative findings, such as grade of tumor, combined with intraoperative findings, such as the depth of invasion into the myometrium, stratified patients into low-, intermediate-, and high-risk groups. These findings correlated with the likelihood of nodal involvement.2,3 In GOG protocol 99, the GOG protocol 33 stratification was further delineated, dividing the intermediate-risk group into “low intermediate risk” and “high intermediate risk” (HIR). These groupings were based on factors associated with the increased recurrence rate, including increasing age, moderate to poorly differentiated grade of tumor, presence of lymphovascular space invasion (LVSI), and outer third myometrial invasion.4 Gynecologic Oncology Group protocol 99 demonstrated that HIR disease confers a 22% risk of recurrence, and this group benefits from adjuvant radiation to decrease the risk of pelvic and vaginal recurrences. More recent studies have demonstrated that other histologic features, such as LVSI, also predict nodal disease.5 Hence, identification of patients with HIR or extrauterine disease is of paramount importance, but factors such as depth of invasion and LVSI are not known preoperatively.
Historically, these studies have served to highlight the benefit of nodal dissection in comprehensive surgical staging for endometrial cancer, but routine lymphadenectomy is also problematic. Two large European studies, however, have called into question the therapeutic value of routine lymph node sampling in patients with this disease. These studies have shown no difference in survival when nodal staging is omitted.6,7 In addition, lymphadenectomy is associated with increased operative time and complications, including increased blood loss, length of hospital stay, lymphocysts, and lymphedema.8–10 Improved preoperative planning would allow the surgeon to identify patients requiring adjuvant therapy whereas minimizing the morbidity and mortality of surgery associated with lymph node dissection.
Cervical cytology (Papanicolaou test) is a screening test initially devised to detect potentially precancerous and cancerous processes of the uterine cervix.11 A meta-analysis shows that the incidence of atypical glandular cells (AGC) is between 0.1% and 2.1%.12 Atypical glandular cells represent a marker for significant glandular abnormalities, including complex endometrial hyperplasia with atypia, adenocarcinoma in situ of the cervix, and adenocarcinoma of the cervix or endometrium. Patients with AGC are also at risk for squamous pathology from cervical dysplasia.13 Previous studies have shown women with malignant preoperative cervical cytology may be at increased risk for extrauterine metastases of endometrial cancer, including pelvic and para-aortic lymph node metastases.14
In this study, we hypothesized that AGCs or malignant cells (MCs) found on cervical cytologic diagnosis independently serve as a predictor of extrauterine disease and identify patients with HIR disease. In addition, we statistically modeled combinations of known preoperative risk factors for extrauterine disease, such as the preoperative International Federation of Gynecology and Obstetrics (FIGO) tumor grade15 and CA-125 levels,16 in conjunction with AGC or MC to increase predictive values.
MATERIALS AND METHODS
Institutional review board approval was obtained for this single institution case-control study through the Colorado Multiple Institutional Review Board. Women aged 18 years and older with a diagnosis of endometrial adenocarcinoma that were staged at the University of Colorado Hospital (UCH) from January 2002 through June 2012 were identified from the UCH Tumor Registry. Patients without preoperative cervical cytologic diagnosis reviewed by the University of Colorado Pathology Department were excluded. Cases included patients with either AGC or MC diagnosed on preoperative cervical cytologic examination. Controls were identified using the same inclusion criteria as the case group but did not have AGC or MC on cervical cytology (Fig. 1).
Data on age, race, body mass index (BMI), preoperative CA-125 levels, surgical procedures completed, and risk category (based on the criteria from GOG protocol 994) were abstracted from the clinical chart. Follow-up data including date of recurrence, disease progression, or death were collected from the clinical chart or the UCH Tumor Registry. Data on histologic grade and cell type, lymph node counts and status, stage, and LVSI were abstracted from the pathology. Lymphovascular space invasion was defined as the presence of any adenocarcinoma in endothelium-lined channels of uterine specimens extracted at the time of surgery, as described by GOG protocol 99.4 Grade was defined by the standard FIGO criteria,17 and the stage was based on the FIGO 2009 staging system for endometrial cancer.18 Extrauterine disease was defined as FIGO stage II or greater. High intermediate risk status was determined using risk factors identified in GOG protocol 99,4 including FIGO grade 2 or 3 tumors, presence of LVSI, and outer third myometrial invasion. Patients were stratified into the HIR group if they were (1) at least 70 years with only 1 risk factor, (2) at least 50 years with any 2 risk factors, or (3) any age with all of the risk factors.4 Elevated CA-125 level was defined as a value greater than 35 U/mL.16 Study data were collected and managed using the Research Electronic Data Capture tools hosted at the University of Colorado.19 Statistical analysis was performed with SPSS package 20 (SPSS, Chicago, Ill). Statistical significance was calculated using the χ2 test for differences in dichotomous variables, and t tests were used to determine differences in continuous variables. Multivariate analyses, using logistic regression modeling, were used to evaluate possible predictors of HIR and extrauterine disease, including myometrial invasion, tumor grade, elevated CA-125 levels, presence of LVSI, and AGC or MC. Statistical significance was determined by P values less than 0.05.
The demographic and clinical characteristics were evaluated using descriptive statistics and compared for cases and controls, and odds ratios (ORs) and their 95% confidence intervals (CIs) were computed. The primary research question was whether AGC or MC on preoperative cervical cytology is a predictor of extrauterine disease (FIGO stage ≥II) or HIR. Secondary analysis explored whether AGC or MC on preoperative cervical cytology predicts the presence of positive lymph nodes and sought to improve the prediction of poor prognosis variables by using combinations of AGC or MC with other preoperative factors such as preoperative tumor grade and CA-125 levels. A power analysis used HIR category and extrauterine disease as the primary outcome variables. Sample size estimates were calculated using α = 0.05 and β = 80% for each.
A total of 568 patients were surgically staged during the study. There were 462 women excluded because they did not have preoperative cervical cytology reviewed at the UCH. Thirty-nine women with endometrial adenocarcinoma who had preoperative cervical cytologic diagnosis with either AGC or MC (cases) were compared with 67 women with endometrial carcinoma who had normal preoperative cervical cytology (controls). There were no statistical differences in age, race, BMI, lymphadenectomy, or mean lymph node counts between the cases with AGC or MC and controls (Table 1). To control for the large number of patients excluded because their cervical cytology was not reviewed at our institution, a comparison was done between groups. The 106 study patients and the 462 excluded patients revealed no statistical difference in age, race, histologic cell type, and FIGO grade (Table 2).
Among the AGC or MC cases, 53.8% (21/39) had HIR disease compared with 30.3% (20/67) with normal cervical cytology (OR, 2.68; 95% CI, 1.18–6.09; P = 0.02; Table 3). Extrauterine disease was found in 43.6% (17/39) of patients with AGC or MC compared with that in 15.2% (10/67) of patients with normal cervical cytology (OR, 4.33; 95% CI, 1.72–10.90; P < 0.01; Table 3). Multivariate analyses confirmed that AGC or MC was an independent predictor of HIR disease (adjusted OR [aOR], 8.41; 95% CI, 1.34–52.78; P = 0.02) and extrauterine disease (aOR, 4.78; 95% CI, 1.26–18.07; P = 0.02; Table 4).
When elevated CA-125 and AGC or MC results were combined, the likelihood of extrauterine disease (OR, 13.3; 95% CI, 3.1–56.8; P < 0.01) and HIR disease (OR, 5.83; 95% CI, 1.43–23.71; P = 0.02) increased significantly. Atypical glandular cell or MC was not significantly associated with the presence of either a positive pelvic lymph node (P = 0.06), a positive para-aortic lymph node (P = 0.07), or any positive lymph node (P = 0.06; Table 3).
To further assess the validity of our data, the study population was then compared with previously reported risk factors for extrauterine disease, including LVSI, elevated CA-125 levels, and greater than or equal to 50% myometrial invasion (Table 5). Lymphovascular space invasion was significantly associated with extrauterine disease (OR, 9.13; 95% CI, 3.28–25.41; P < 0.01), HIR disease (OR, 13.77; 95% CI, 4.54–41.73; P < 0.01), and presence of a positive lymph node (P = 0.01). Elevated CA-125 levels were also significantly associated with extrauterine disease (OR, 7.40; 95% CI, 2.36–23.36; P < 0.01), HIR disease (OR, 3.22; 95% CI, 1.14–9.07; P = 0.02), and presence of a positive lymph node (P = 0.02). Multivariate analysis confirmed that the elevated CA-125 level was an independent predictor of extrauterine disease (aOR, 7.84; 95% CI, 2.03–30.22; P < 0.01; Table 4) but not HIR disease (P = 0.57). Greater than or equal to 50% myometrial invasion was also associated with extrauterine disease (OR, 9.00; 95% CI, 3.23–25.05; P < 0.01) and HIR disease (OR, 19.08; 95% CI, 5.78–63.00; P < 0.01). There was a trend toward predicting the presence of a positive lymph node (OR, 10.09; 95% CI, 0.95–107.00; P = 0.06; Table 5). Multivariate analysis confirmed greater than or equal to 50% myometrial invasion as an independent predictor of HIR disease (aOR, 47.50; 95% CI, 4.54–497.22; P < 0.01; Table 4).
In the current study, we demonstrate that endometrial cancer patients with AGC or MC on preoperative cervical cytology are at increased risk for HIR disease and extrauterine disease. Atypical glandular cell or MC was an independent predictor of both HIR and extrauterine disease. The predictive ability of this preoperative cytologic diagnosis is increased when combined with an elevated preoperative CA-125. Our data also support previously identified risk factors for extrauterine disease, including greater than or equal to 50% myometrial invasion,3 elevated CA-125 levels,16 and LVSI.5 The controls included in the study have similar incidence patterns for HIR4 and extrauterine disease20 as previously reported in larger cohorts.
Although cervical cytologic diagnosis was initially designed for cervical cancer screening,11 it has demonstrated its utility in other gynecologic pathologies. Complex endometrial hyperplasia with atypia, adenocarcinoma in situ of the cervix, and adenocarcinoma of the cervix or endometrium were identified in a cohort of 492 patients initially diagnosed with AGC.13 Recently, in a small pilot study, analysis of DNA from cervical cytologic diagnosis has been shown to accurately identify 100% of endometrial cancers and 41% of ovarian cancers in women known to be affected with these diseases.21
Our study has several important implications. First, AGC or MC on cervical cytology adds to several already identified preoperative findings, such as elevated CA-125 levels, grade of tumor, and histologic cell type, that identify patients at risk for HIR, extrauterine disease, and aggressive tumor behavior. As shown by Keys et al4 in GOG protocol 99, patients with HIR disease benefit from adjuvant radiation to decrease pelvic and vaginal recurrence. Second, cervical cytology is commonly obtained preoperatively. When used in conjunction with CA-125 levels, it may more accurately identify patients requiring comprehensive surgical staging or postoperative adjuvant therapy. These findings are also not subject to issues inherent to final pathology, where risk factors including LVSI or a change in the grade of tumor or histologic cell type are revealed after the completion of surgery.
Strengths of our study include all procedures being conducted at a single institution and well-curated follow-up data through the tumor registry. Limitations of this study involve those common to a large tertiary referral center, including an inherent selection bias that may include more complex and advanced disease patients. In addition, potential cases may have been excluded in this study due to dependence on preoperative cervical cytology. We are currently designing a trial where all cervical cytology and surgical pathology specimens will be collected and reviewed prospectively. In conclusion, cervical cytology is an easily obtained specimen that when reported as AGC or MC should warn surgeons of extrauterine or occult metastatic endometrial cancer.
1. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2013. CA Cancer J Clin. 2013; 63: 11–30.
2. Boronow RC, Morrow CP, Creasman WT, et al. Surgical staging in endometrial cancer: clinical-pathologic findings of a prospective study. Obstet Gynecol. 1984; 63: 825–832.
3. Creasman WT, Morrow CP, Bundy BN, et al. Surgical pathologic spread patterns of endometrial cancer. A Gynecologic Oncology Group study. Cancer. 1987; 60: 2035–2041.
4. Keys HM, Roberts JA, Brunetto VL, et al. A phase III trial of surgery with or without adjunctive external pelvic radiation therapy in intermediate risk endometrial adenocarcinoma: a Gynecologic Oncology Group study. Gynecol Oncol. 2004; 92: 744–751.
5. Guntupalli SR, Zighelboim I, Kizer NT, et al. Lymphovascular space invasion is an independent risk factor for nodal disease and poor outcomes in endometrioid endometrial cancer. Gynecol Oncol. 2012; 124: 31–35.
6. Benedetti Panici P, Basile S, Maneschi F, et al. Systematic pelvic lymphadenectomy vs. no lymphadenectomy in early-stage endometrial carcinoma: randomized clinical trial. J Natl Cancer Inst. 2008; 100: 1707–1716.
7. Kitchener H, Swart AM, Qian Q, et al. Efficacy of systematic pelvic lymphadenectomy in endometrial cancer (MRC ASTEC trial): a randomised study. Lancet. 2009; 373: 125–136.
8. Abu-Rustum NR, Alektiar K, Iasonos A, et al. The incidence of symptomatic lower-extremity lymphedema following treatment of uterine corpus malignancies: a 12-year experience at Memorial Sloan-Kettering Cancer Center. Gynecol Oncol. 2006; 103: 714–718.
9. Conte M, Panici PB, Guariglia L, et al. Pelvic lymphocele following radical para-aortic and pelvic lymphadenectomy for cervical carcinoma: incidence rate and percutaneous management. Obstet Gynecol. 1990; 76: 268–271.
10. Larson DM, Johnson K, Olson KA. Pelvic and para-aortic lymphadenectomy for surgical staging of endometrial cancer: morbidity and mortality. Obstet Gynecol. 1992; 79: 998–1001.
11. Papanicolaou GN, Traut HF. The diagnostic value of vaginal smears in carcinoma of the uterus. 1941. Arch Pathol Lab Med. 1997; 121: 211–224.
12. Marques JP, Costa LB, Pinto AP, et al. Atypical glandular cells and cervical cancer: systematic review. Rev Assoc Med Bras. 2011; 57: 234–238.
13. Geier CS, Wilson M, Creasman W. Clinical evaluation of atypical glandular cells of undetermined significance. Am J Obstet Gynecol. 2001; 184: 64–69.
14. Larson DM, Johnson KK, Reyes CN, et al. Prognostic significance of malignant cervical cytology in patients with endometrial cancer. Obstet Gynecol. 1994; 84: 399–403.
15. Larson DM, Johnson KK, Broste SK, et al. Comparison of D&C and office endometrial biopsy in predicting final histopathologic grade in endometrial cancer. Obstet Gynecol. 1995; 86: 38–42.
16. Sood AK, Buller RE, Burger RA, et al. Value of preoperative CA 125 level in the management of uterine cancer and prediction of clinical outcome. Obstet Gynecol. 1997; 90: 441–447.
17. Announcements. FIGO stages—1988 revisions. Gynecol Oncol. 1989; 35: 125–127.
18. Creasman W. Revised FIGO staging for carcinoma of the endometrium. Int J Gynaecol Obstet. 2009; 105: 109.
19. Harris PA, Taylor R, Thielke R, et al. Research electronic data capture (REDCap)—a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009; 42: 377–381.
20. Lewin SN, Herzog TJ, Barrena Medel NI, et al. Comparative performance of the 2009 International Federation of Gynecology and Obstetrics’ staging system for uterine corpus cancer. Obstet Gynecol. 2010; 116: 1141–1149.
21. Kinde I, Bettegowda C, Wang Y, et al. Evaluation of DNA from the Papanicolaou test to detect ovarian and endometrial cancers. Sci Transl Med. 2013; 5: 167ra4.
Atypical glandular cells; Atypical glandular cells of undetermined significance; Cervical cytologic diagnosis; Endometrial cancer; Papanicolaou test
© 2014 by the International Gynecologic Cancer Society and the European Society of Gynaecological Oncology.
What does "Remember me" mean?
By checking this box, you'll stay logged in until you logout. You'll get easier access to your articles, collections,
media, and all your other content, even if you close your browser or shut down your
To protect your most sensitive data and activities (like changing your password),
we'll ask you to re-enter your password when you access these services.
What if I'm on a computer that I share with others?
If you're using a public computer or you share this computer with others, we recommend
that you uncheck the "Remember me" box.
Highlight selected keywords in the article text.
Data is temporarily unavailable. Please try again soon.