Atypical Glandular Cells of Endometrial Origin and the Risk of Endometrial Cancer Xuezhi Jiang, MD, PhD,1,2 Charmaine Anderson, MD,1 Kathryn E. Sharpless, MD, PhD,3 Jessica White, DO,1 Chevon Alderson, MPH,1 John Demko, BS,1 Bernice Robinson-Bennett, MD,1 and Peter F. Schnatz, DO1,2,4,5 Objectives: To assess the risk of endometrial cancer (EC) associated with atypical glandular cells of endometrial origin (AGC-EM) in 2 age groups (age younger than 51 vs 51 years or older). Methods: A retrospective case series was assembled identifying AGC from a pathology database between January 1, 2005 and January 1, 2009. Demographics, cervical cytology results, and final diagnoses (including clinically significant diseases and cancers) were recorded from the initial AGC diagnosis until August 30, 2011. Data were analyzed using the χ2 test to compare rates of disease between age groups. Results: Among the 444 patients with AGC, 41% (183/444) had AGCEM. Women younger than 51 years, compared to those 51 years or older, had significantly lower rates of AGC-EM (35% [105/296] vs 53% [78/148]; p < .001; odds ratio, 0.49; 95% confidence interval, 0.33–0.74). The rate of EC was significantly lower in those younger than 51 years, compared to those aged 51 or older (5% [8/158] vs 19% [18/95]; p < .001; odds ratio, 0.23; 95% confidence interval, 0.09–0.55) in women who underwent endometrial biopsy. In women younger than 51 years who underwent an endometrial biopsy, the rate of EC had a stepwise increase across 3 subclasses of AGC (from AGC of endocervical origin [AGC-EC] to AGC not otherwise specified to AGC-EM) ( p = .04). Conclusions: Women aged 51 years or older who have AGC are more likely to have AGC-EM and EC than women younger than 51 years. In women younger than age 51, AGC-EM is the subclass most associated with EC while compared to 2 other subclasses (AGC not otherwise specified and AGC-EC). Key Words: atypical glandular cells, atypical endometrial cells, endometrial cancer, dysplasia, cervical cytology (J Lower Gen Tract Dis 2015;19: 194–196)
A
typical glandular cells (AGCs) are one of 4 subcategories of glandular cell abnormalities on cervical cytology, along with AGC favor neoplasia, endocervical adenocarcinoma in situ, and adenocarcinoma. Atypical glandular cells are further categorized based on the origin of the glandular cells as endocervical, endometrial, or not otherwise specified (NOS).1 Although AGCs are found in less than 0.5% of all cervical cytology reports, they are associated with a high rate of clinically significant disease
1
Department of Obstetrics and Gynecology, The Reading Hospital, Reading, PA; Department of Obstetrics and Gynecology, Sidney Kimmel Medical College of Thomas Jefferson University, Philadelphia, PA; 3Department of Obstetrics and Gynecology, Baystate Medical Center, Springfield, MA; 4Department of Internal Medicine, The Reading Hospital, Reading, PA; and 5Department of Internal Medicine, Sidney Kimmel Medical College of Thomas Jefferson University, Philadelphia, PA Correspondence to: Xuezhi Jiang, MD, PhD, FACOG, NCMP, Sidney Kimmel Medical College of Thomas Jefferson University; and Department of Obstetrics and Gynecology, The Reading Hospital, PO Box 16052, Reading, PA 19612-6052. E-mail: [email protected] These data have been presented in poster format at the 2012 March Biennial Meeting of the American Society for Colposcopy and Cervical Pathology. However, these data and results have not been published in manuscript form. The authors have declared they have no conflicts of interest. © 2015, American Society for Colposcopy and Cervical Pathology
2
194
including premalignant and malignant disease (9%–50%), and approximately 5% of women with AGCs have an underlying malignancy.2 Previous studies have shown that of the patients with AGCs who have a diagnosis of cancer, approximately 58% are of endometrial origin, 29% are cervical, 6% are ovarian, and approximately 7% are from distant origins.2 Atypical glandular cells of endometrial origin (AGC-EM), which are reported as favoring endometrial or atypical endometrial cells, are a less commonly reported AGC finding than AGC of endocervical origin (AGC-EC) or AGC-NOS. However, AGC-EM is associated with higher rates of preneoplastic disease and endometrial cancer.3,4 The 2012 updated American Society for Colposcopy and Cervical Pathology abnormal cervical cytology guidelines recommend endometrial sampling as an initial workup for all patients with atypical endometrial cells, regardless of age. Several studies have reported that premenopausal women with AGC-EM are at a lower risk of endometrial cancer than postmenopausal women with AGC-EM.5,6 Does AGC-EM in younger or premenopausal women warrant the same degree of close workup as postmenopausal women? This study was designed to evaluate the clinical significance of AGC-EM by age group (younger than 51 years and 51 years or older).
METHODS After approval by the Reading Hospital Institutional Review Board, a retrospective case series identified cases of AGC from a pathology database at the Reading Hospital between January 1, 2005, and January 1, 2009. All women with cervical cytological results of AGC-EM, which were reported as either favoring endometrial or having atypical endometrial cells, were reviewed. Through a review of the electronic medical database, demographics, cervical cytological results, colposcopic biopsy and/or endocervical curettage results, endometrial biopsy results, concurrent endometrial and cervical abnormalities, subsequent cytological and histological follow-up, and pathology of hysterectomy specimens were recorded from January 1, 2005, to August 30, 2011. Owing to a lack of information regarding menopausal status and based on the mean age of menopause, age 51, the study population was stratified into 2 groups, age younger than 51 years and age 51 years or older. Rates of endometrial cancer were calculated, and data were analyzed using the χ2 test to compare the rates of conditions and disease between the groups. The Fisher exact test was used if more than 20% of the expected cell frequencies were less than 5. SAS 9.3 was used for data analyses, with p < .05 being deemed statistically significant.
RESULTS During the 4-year study period, 444 cases of AGC were identified from the Reading Hospital pathology database. The incidence of AGC during this time period was 0.2% (444/230,780). Of the 444 AGC, 41% (183/444) were AGC-EM, 29% (127/444) were AGC-EC, and 30% (134/444) were AGC-NOS. Endometrial
Journal of Lower Genital Tract Disease • Volume 19, Number 3, July 2015
Copyright © 2015 American Society for Colposcopy and Cervical Pathology. Unauthorized reproduction of this article is prohibited.
Journal of Lower Genital Tract Disease • Volume 19, Number 3, July 2015
biopsies were performed in 57% (253/444). In women who had an endometrial biopsy, the rate of endometrial cancer was 10% (26/253) and the rate of atypical endometrial hyperplasia was 7% (18/253). There were 296 cases of AGC in women younger than 51, and 148 cases of AGC in women aged 51 years or older (Table 1). Rates of endometrial biopsies in the women aged 51 years or older were higher than those in the women younger than 51 years (64% [95/148] vs 53% [158/296]; p = .03), and rates of endometrial cancer were also higher in the women aged 51 years or older compared to the younger age group (19% [18/95] vs 5% [8/158]; p < .001). The proportion of AGC that were AGC-EM was higher in the women aged 51 years or older compared to the women younger than 51 years (53% [78/148] vs 35% [105/296]; p < .001). Although the percentage of endometrial biopsies performed in cases of AGC-EM in the women aged 51 years or older were similar to that in the younger age group (69% [54/78] vs 71% [75/105]; p = .75), the rates of endometrial cancer in those who had AGC-EM and endometrial biopsies were higher in the women aged 51 years or older compared to the women younger than 51 (20% [11/54] vs 8% [6/75]; p = .04) (Table 1). Of 183 who had AGC-EM, 30% (54/183) did not have endometrial biopsies performed after the initial cytological finding of AGC-EM. Of these 54, 48% (26) had either a cervical biopsy or an endocervical curettage after the initial diagnosis of AGC-EM. In addition, 78% (42/54) of these women had at least one cytological follow-up within 2 years. Most malignancies in the AGC-EM cases were endometrial cancer. In the women younger than 51 years with AGC-EM, 6 (86%) of 7 women with cancer had endometrial cancer and one had skin cancer. In the women aged 51 years or older with AGC-EM, 11 (92%) of 12 women with cancer had endometrial cancer and one had breast cancer. In Table 2, the rates of endometrial cancer were compared by age group and AGC subclass in women who had an endometrial biopsy. Overall, the rates of endometrial cancer were higher in both AGC-EM (13%) and AGC-NOS (13%) compared to that in AGC-EC (3%) (p = .06). The χ2 test for trend was conducted, and a significant trend existed in the rates of endometrial cancer over 3 AGC subclasses moving up by order from AGC-EC to AGC-NOS to AGC-EM; p = .03. While performing the test by age group, this significant trend was only seen in the women younger than 51, rather than in the elder counterpart (p = .04). In women younger than 51 years, the rate of endometrial cancer in women with endometrial biopsies
AGC of Endometrial Origin
was the highest in women with AGC-EM compared to the other AGC subclasses (p = .06; Table 2).
DISCUSSION The incidence of AGC at our institution was similar to the incidence in previous reports.7 However, AGC-EM comprised 41% of the cases of AGC, which is higher than previous reports.3–5 The AGC-EM rate was 5.3% in the series of Chhieng et al5 collected from 1995 to 1999, 8.7% in that of Castle et al4 collected after 2003, and 20.2% in the series of Zhao et al 3 collected from 2005 to 2007. Whether this is an institutional reporting difference or whether it reflects an increase in the rate of AGC-EM over time is unclear. Previous studies have reported higher rates of endometrial cancer in postmenopausal women compared to premenopausal women.4,8 The rate of endometrial cancer in our study was significantly higher in the women aged 51 years or older (19%) compared to the women younger than 51 (5%) with AGC. The overall rate of endometrial sampling was 57% (253/444) despite published guidelines recommending endometrial sampling in women with atypical endometrial cells and in women with AGC aged 35 years or older or at risk for endometrial neoplasia.7,9 Older (51 years or older) and younger (younger than 51 years) women with AGC-EM were equally likely to have an endometrial biopsy (69% vs 71%) in this study, although AGC-EM and endometrial cancer were more common in older (53% and 20%, respectively) than younger women (35% and 8%, respectively). It was interesting to note that the rate of endometrial cancer in the women aged 51 years or older with AGC-NOS was similar to that in those with AGC-EM (20% vs 19%). Statistically speaking, the nonsignificant trend test may reflect the similar risk for endometrial cancer in older women across the 3 AGC subclasses. A significant trend and increased risk was seen in the women younger than 51 years, however, and may warrant closer attention and workup of AGC-EM than AGC-EC in this population. Regardless, these findings support the current guidelines that recommend endometrial sampling in women with atypical endometrial cells regardless of age.10 With the increasing incidence of endometrial cancer, it is also important to recognize the potential risk for endometrial cancer in women with AGC-NOS, which may be similar to that in women with AGC-EM, especially in older women aged 51 years or older. As a high-risk category for endometrial cancer, AGC-EM may
TABLE 1. Atypical Glandular Cells Summary Statistics by Age Groups Age Group Age
A
typical glandular cells (AGCs) are one of 4 subcategories of glandular cell abnormalities on cervical cytology, along with AGC favor neoplasia, endocervical adenocarcinoma in situ, and adenocarcinoma. Atypical glandular cells are further categorized based on the origin of the glandular cells as endocervical, endometrial, or not otherwise specified (NOS).1 Although AGCs are found in less than 0.5% of all cervical cytology reports, they are associated with a high rate of clinically significant disease
1
Department of Obstetrics and Gynecology, The Reading Hospital, Reading, PA; Department of Obstetrics and Gynecology, Sidney Kimmel Medical College of Thomas Jefferson University, Philadelphia, PA; 3Department of Obstetrics and Gynecology, Baystate Medical Center, Springfield, MA; 4Department of Internal Medicine, The Reading Hospital, Reading, PA; and 5Department of Internal Medicine, Sidney Kimmel Medical College of Thomas Jefferson University, Philadelphia, PA Correspondence to: Xuezhi Jiang, MD, PhD, FACOG, NCMP, Sidney Kimmel Medical College of Thomas Jefferson University; and Department of Obstetrics and Gynecology, The Reading Hospital, PO Box 16052, Reading, PA 19612-6052. E-mail: [email protected] These data have been presented in poster format at the 2012 March Biennial Meeting of the American Society for Colposcopy and Cervical Pathology. However, these data and results have not been published in manuscript form. The authors have declared they have no conflicts of interest. © 2015, American Society for Colposcopy and Cervical Pathology
2
194
including premalignant and malignant disease (9%–50%), and approximately 5% of women with AGCs have an underlying malignancy.2 Previous studies have shown that of the patients with AGCs who have a diagnosis of cancer, approximately 58% are of endometrial origin, 29% are cervical, 6% are ovarian, and approximately 7% are from distant origins.2 Atypical glandular cells of endometrial origin (AGC-EM), which are reported as favoring endometrial or atypical endometrial cells, are a less commonly reported AGC finding than AGC of endocervical origin (AGC-EC) or AGC-NOS. However, AGC-EM is associated with higher rates of preneoplastic disease and endometrial cancer.3,4 The 2012 updated American Society for Colposcopy and Cervical Pathology abnormal cervical cytology guidelines recommend endometrial sampling as an initial workup for all patients with atypical endometrial cells, regardless of age. Several studies have reported that premenopausal women with AGC-EM are at a lower risk of endometrial cancer than postmenopausal women with AGC-EM.5,6 Does AGC-EM in younger or premenopausal women warrant the same degree of close workup as postmenopausal women? This study was designed to evaluate the clinical significance of AGC-EM by age group (younger than 51 years and 51 years or older).
METHODS After approval by the Reading Hospital Institutional Review Board, a retrospective case series identified cases of AGC from a pathology database at the Reading Hospital between January 1, 2005, and January 1, 2009. All women with cervical cytological results of AGC-EM, which were reported as either favoring endometrial or having atypical endometrial cells, were reviewed. Through a review of the electronic medical database, demographics, cervical cytological results, colposcopic biopsy and/or endocervical curettage results, endometrial biopsy results, concurrent endometrial and cervical abnormalities, subsequent cytological and histological follow-up, and pathology of hysterectomy specimens were recorded from January 1, 2005, to August 30, 2011. Owing to a lack of information regarding menopausal status and based on the mean age of menopause, age 51, the study population was stratified into 2 groups, age younger than 51 years and age 51 years or older. Rates of endometrial cancer were calculated, and data were analyzed using the χ2 test to compare the rates of conditions and disease between the groups. The Fisher exact test was used if more than 20% of the expected cell frequencies were less than 5. SAS 9.3 was used for data analyses, with p < .05 being deemed statistically significant.
RESULTS During the 4-year study period, 444 cases of AGC were identified from the Reading Hospital pathology database. The incidence of AGC during this time period was 0.2% (444/230,780). Of the 444 AGC, 41% (183/444) were AGC-EM, 29% (127/444) were AGC-EC, and 30% (134/444) were AGC-NOS. Endometrial
Journal of Lower Genital Tract Disease • Volume 19, Number 3, July 2015
Copyright © 2015 American Society for Colposcopy and Cervical Pathology. Unauthorized reproduction of this article is prohibited.
Journal of Lower Genital Tract Disease • Volume 19, Number 3, July 2015
biopsies were performed in 57% (253/444). In women who had an endometrial biopsy, the rate of endometrial cancer was 10% (26/253) and the rate of atypical endometrial hyperplasia was 7% (18/253). There were 296 cases of AGC in women younger than 51, and 148 cases of AGC in women aged 51 years or older (Table 1). Rates of endometrial biopsies in the women aged 51 years or older were higher than those in the women younger than 51 years (64% [95/148] vs 53% [158/296]; p = .03), and rates of endometrial cancer were also higher in the women aged 51 years or older compared to the younger age group (19% [18/95] vs 5% [8/158]; p < .001). The proportion of AGC that were AGC-EM was higher in the women aged 51 years or older compared to the women younger than 51 years (53% [78/148] vs 35% [105/296]; p < .001). Although the percentage of endometrial biopsies performed in cases of AGC-EM in the women aged 51 years or older were similar to that in the younger age group (69% [54/78] vs 71% [75/105]; p = .75), the rates of endometrial cancer in those who had AGC-EM and endometrial biopsies were higher in the women aged 51 years or older compared to the women younger than 51 (20% [11/54] vs 8% [6/75]; p = .04) (Table 1). Of 183 who had AGC-EM, 30% (54/183) did not have endometrial biopsies performed after the initial cytological finding of AGC-EM. Of these 54, 48% (26) had either a cervical biopsy or an endocervical curettage after the initial diagnosis of AGC-EM. In addition, 78% (42/54) of these women had at least one cytological follow-up within 2 years. Most malignancies in the AGC-EM cases were endometrial cancer. In the women younger than 51 years with AGC-EM, 6 (86%) of 7 women with cancer had endometrial cancer and one had skin cancer. In the women aged 51 years or older with AGC-EM, 11 (92%) of 12 women with cancer had endometrial cancer and one had breast cancer. In Table 2, the rates of endometrial cancer were compared by age group and AGC subclass in women who had an endometrial biopsy. Overall, the rates of endometrial cancer were higher in both AGC-EM (13%) and AGC-NOS (13%) compared to that in AGC-EC (3%) (p = .06). The χ2 test for trend was conducted, and a significant trend existed in the rates of endometrial cancer over 3 AGC subclasses moving up by order from AGC-EC to AGC-NOS to AGC-EM; p = .03. While performing the test by age group, this significant trend was only seen in the women younger than 51, rather than in the elder counterpart (p = .04). In women younger than 51 years, the rate of endometrial cancer in women with endometrial biopsies
AGC of Endometrial Origin
was the highest in women with AGC-EM compared to the other AGC subclasses (p = .06; Table 2).
DISCUSSION The incidence of AGC at our institution was similar to the incidence in previous reports.7 However, AGC-EM comprised 41% of the cases of AGC, which is higher than previous reports.3–5 The AGC-EM rate was 5.3% in the series of Chhieng et al5 collected from 1995 to 1999, 8.7% in that of Castle et al4 collected after 2003, and 20.2% in the series of Zhao et al 3 collected from 2005 to 2007. Whether this is an institutional reporting difference or whether it reflects an increase in the rate of AGC-EM over time is unclear. Previous studies have reported higher rates of endometrial cancer in postmenopausal women compared to premenopausal women.4,8 The rate of endometrial cancer in our study was significantly higher in the women aged 51 years or older (19%) compared to the women younger than 51 (5%) with AGC. The overall rate of endometrial sampling was 57% (253/444) despite published guidelines recommending endometrial sampling in women with atypical endometrial cells and in women with AGC aged 35 years or older or at risk for endometrial neoplasia.7,9 Older (51 years or older) and younger (younger than 51 years) women with AGC-EM were equally likely to have an endometrial biopsy (69% vs 71%) in this study, although AGC-EM and endometrial cancer were more common in older (53% and 20%, respectively) than younger women (35% and 8%, respectively). It was interesting to note that the rate of endometrial cancer in the women aged 51 years or older with AGC-NOS was similar to that in those with AGC-EM (20% vs 19%). Statistically speaking, the nonsignificant trend test may reflect the similar risk for endometrial cancer in older women across the 3 AGC subclasses. A significant trend and increased risk was seen in the women younger than 51 years, however, and may warrant closer attention and workup of AGC-EM than AGC-EC in this population. Regardless, these findings support the current guidelines that recommend endometrial sampling in women with atypical endometrial cells regardless of age.10 With the increasing incidence of endometrial cancer, it is also important to recognize the potential risk for endometrial cancer in women with AGC-NOS, which may be similar to that in women with AGC-EM, especially in older women aged 51 years or older. As a high-risk category for endometrial cancer, AGC-EM may
TABLE 1. Atypical Glandular Cells Summary Statistics by Age Groups Age Group Age
