The concept of atypical glandular cells of undetermined significance (AGUS) introduced by the Bethesda system (TBS) in 1988  has been used as a diagnostic category to denote cytologic changes in endocervical or endometrial glandular cells that exceed those typical of reactive changes but that are quantitatively and/or qualitatively not diagnostic of adenocarcinoma . The National Cancer Institute Workshop has suggested management protocols based on the presence of endocervical cells or endometrial cells . When possible, AGUS should be qualified further as to the origin of the glandular cells (endocervical or endometrial). When the atypical cells are endocervical, an attempt should be made to differentiate between a reactive and a neoplastic process. Atypical endometrial cells, however, are not subdivided further owing to the absence of good cytological criteria for differentiating atypical and reactive endometrial cells from those that are neoplastic.
AGUS represents a wide variety of conditions, ranging from clinically benign lesions such as inflammation, tubal metaplasia, microglandular hyperplasia, and endocervical/endometrial polyps to clinically significant preneoplastic or malignant lesions. Several studies have addressed the rate and the clinical significance of AGUS [4–12]. The origin of the benign or premalignant/malignant lesions that present as AGUS cytology vary with patient age. In younger patients, cervical pathology is more likely, whereas in older patients the endometrium is the more likely origin of atypical glandular cells . It is increasingly important to consider a patient's age when assessing the clinical implications of an AGUS smear.
Although age has been considered in some reports, most have not focused on patient age as a variable [4–12]. A recent study of the clinical significance of AGUS in postmenopausal women has been reported . However, a systematic study of the clinical implications of an AGUS diagnosis at different age groups is lacking. The current retrospective study was performed to compare the rate, accuracy, and histological correlates of cytological smears reported as AGUS in patients of different age groups.
MATERIALS AND METHODS
The computer-based files of the Cytopathology Department of Long Island Jewish Medical Center were searched for all diagnoses of AGUS (as defined by TBS) from January 1996 to December 2000. There is no overlap between the cases reported here and the previous publication from this institution . The diagnoses of AGUS were made using published TBS criteria [1,2]. Atypical glandular cells were classified into three categories: atypical endocervical cells, atypical endometrial cells, and atypical glandular cells not otherwise specified (AGNOS). Atypical endocervical cells were further subclassified as AGUS-favor reactive (AGFR) and AGUS-favor neoplastic (AGFN) based on whether a reactive or a preneoplastic/neoplastic process was suspected. The initial diagnoses and slides were reviewed and correlated with the subsequent histologic follow-up (including endocervical biopsy/curettage, cervical cone biopsy, and/or endometrial biopsy) and/or repeat smears when they were available. Follow-up histology/cytology was considered only if it was obtained within 4 months of the original smear. Discrepancies between original Pap smears and the subsequent histologic follow-up or repeat smears were reviewed.
A total of 40,464 Pap smears were performed at Long Island Jewish Medical Center during this 5-year period. Of that number, 163 (0.4%) smears were diagnosed as AGUS in patients ranging in age from 19 to 90 years. The mean age was 48 years (median age = 47 years). The highest AGUS rates in different age groups occurred in the fifth (0.6%; 46/7,649) and ninth (0.84%; 7/835) decades of life (Fig. 1). During the same 5-year period, there were 2,078 (5.13%) smears reported as atypical squamous cells of undetermined significance (ASCUS), 1,438 (3.55%) smears reported as low-grade squamous intraepithelial lesions (LGSIL), 419 (1.03%) smears reported as high-grade squamous intraepithelial lesions (HGSIL), and 26 (0.06%) smears reported as cancer.
Follow-up was available for 101 of 163 patients with AGUS smears (62%). Of the 101 patients with follow-up, 65 (64.4%) had histologic follow-up only (including endocervical biopsy/curettage, cervical cone biopsy, and/or endometrial biopsy), 13 (12.9%) had repeat cervico-vaginal smears only, and 23 (22.8%) had both histologic and cytologic follow-ups. No follow-up was available in 62 patients (38%; 62/163). The types of procedures performed in patients who underwent histologic evaluation are shown in Table 1.
Twenty-seven of 88 (30.7%) patients with subsequent histologic follow-up demonstrated clinically significant (premalignant or malignant) lesions. There was no difference in the number of significant lesions among any of the cytological categories (p = .113; Fisher exact test). Of those with significant lesions, 15 patients showed endometrial involvement, which included 12 (13.6%) endometrial adenocarcinomas and 3 (3.4%) endometrial hyperplasia. The youngest patient with endometrial carcinoma was 38 years old. The mean age for endometrial hyperplasia/carcinoma was 64 years. Significant cervical lesions occurred in 12 patients. These lesions included 4 that were cervical intraepithelial neoplasia (CIN) 1, 6 that were CIN 2,3, 1 that was adenocarcinoma in situ (AIS) of the cervix, and 1 that was metastatic poorly differentiated carcinoma. The mean age was 45 years, much lower than that of endometrial lesions. The number of cases in each category is summarized in Table 2. The ratios of significant endometrial to cervical lesions increase with age showing a bimodal age distribution pattern (Fig. 2).
Forty (45.5%) patients had nonneoplastic (benign) lesions, which included squamous metaplasia, tubal metaplasia, inflammation, reactive atypia, atrophy, microglandular hyperplasia, and endometrial and endocervical polyps. The majority of these cases were squamous metaplasia and/or inflammation (60%). The number of cases in each category is summarized in Table 3. The remaining 21 cases had normal histologic follow-up.
The diagnoses of AGUS-favor endometrial (AGFEM) cell origin were rendered in 34 patients. Twenty-five patients had either histologic (n = 23) or cytologic (n = 2) follow-up. Nine of 23 (39.1%) patients with histologic follow-up showed significant lesions (8 were endometrial carcinomas and 1was CIN 2,3 of the cervix) (Table 4). The mean age of these patients was 61 years.
Among the 88 patients with subsequent histologic follow-up, 22 patients were diagnosed as AGNOS. This group contained 8 (36.4%) significant lesions (3 were CIN 2,3, 3 were endometrial hyperplasia, and 2 were endometrial adenocarcinoma) (Table 4).
Thirty-two patients were diagnosed with AGFR. The subsequent histologic follow-up demonstrated 5 (15.6%) significant lesions: 2 were CIN 2,3 and 3 were CIN 1 (Table 4).
Eleven of 88 patients with subsequent histologic follow-up were diagnosed as AGFN. The AGFN correctly predicted the presence of 5 (45.5%) significant lesions: 1 was AIS, 1 was CIN 1, 2 were endometrial carcinomas, and 1 was metastatic poorly differentiated carcinoma. The histologic follow-up demonstrated benign lesions in the remaining 6 patients (Table 4).
Among those who had follow-up with cytology only (n = 13), 3 (23.1%) patients showed clinically significant lesions (1 was vaginal adenocarcinoma, 1 was LGSIL, and 1 was HGSIL) (Table 5). The remaining patients had either normal smears (n = 7) or benign lesions (n = 3).
The AGUS rate of 0.4% observed in the current study is within the range seen in the literature  and in agreement with our previous report . The rate of clinically significant lesions in patients with AGUS on subsequent follow-up is 30.7% in the current study, which is within the range of the reported rate of 17% to 80% [4–10].
Although AGUS was diagnosed in almost all the age groups, the current study shows that the highest AGUS rates occurred in the fifth and ninth decades of life (Fig. 1). Significant (premalignant or malignant) lesions occurred in 30.7% of the patients presenting with AGUS.
In the current study, significant cervical lesions occurred in 12 patients who had a mean age of 45 years. Endometrial lesions occurred in older women (mean age = 64;n = 15). The National Cancer Institute Workshop suggested management protocols based on the presence of endocervical cells or endometrial cells. This study indicates that the age of the patient should also be considered in investigating patients with AGUS Pap smears. In younger patients, the pathology is more likely to affect the cervix, whereas in older patients the endometrium is the more likely cause of the atypical glandular cells.
Our study confirmed the presence of these trends, which should affect the clinical management of patients. Although most of the cervical lesions were found in the younger age groups, we did find significant cervical lesions in the sixth, seventh, and eighth decades of life. Similarly, although most of the endometrial lesions were found in the older age groups, we did identify endometrial carcinomas in the fourth and fifth decades of life. Our data suggest that when an age appropriate diagnosis is not identified following the evaluation (cervical disease in younger women, endometrial disease in older women), the evaluation should be extended to include the rest of the genital tract.
Our data also identified another trend with respect to the National Cancer Institute recommendations. The current study shows that 39.1% of the patients presenting as AGFEM cell origin had significant lesions (endometrial carcinoma and HGSIL) as confirmed by histologic follow-up. The mean age of these patients was 61 years. The presence of AGFEM cell origin in older women should always be considered clinically significant, and the origin of the atypical glandular cells must be pursued aggressively. However, the finding of cervical lesions in this cytologic category suggests that if the evaluation fails to identify an endometrial origin of the cytologic abnormality, further workup should be performed to rule out the possibility of a cervical lesion. Similarly, we found one third of our endometrial cancers in smears suggesting a cervical origin of pathology (2 in AGFN and 2 in AGNOS). This suggests that if a cervical lesion cannot be identified in these women, the endometrium should be evaluated. These findings emphasize the point that although cytologic evaluation can direct the clinician to the likely origin of the atypical cells, the process is not perfect.
In the current study, 62% of patients with AGUS had either histologic or cytologic follow-up or both. This follow-up rate is higher than the reported rate of 22% to 58% in the literature [3–9]. Therefore, this study may more closely reflect the clinical correlation of AGUS cytology. Although the study is small and may not have significant statistical power, the rate of AGUS smears is so low that few institutions have sufficient clinical material to report large studies, and so the number of correlational studies is few. Therefore, we must get as much information as we can from observational studies. It is our hope that this report will stimulate additional reporting to allow the collection of further information regarding AGUS cytology.
In conclusion, a finding of AGUS seems to represent 2 distinct disease states, 1 of cervical origin and 1 of endometrial origin. Although cytologic criteria and age can effectively direct the clinician toward the likely origin of the atypical cells, neither is perfect and clinicians should be cautioned not to rely on them exclusively to direct their evaluation. If the etiology of an AGUS smear cannot be discerned when looking in the most likely place, the clinician should pursue the evaluation further.
The authors thank Ms. Chiara Sugrue for her technical assistance in preparing this manuscript.
1. The 1988 Bethesda System for reporting cervical/vaginal cytological diagnoses. National Cancer Institute Workshop. JAMA 1989;262:931–4.
2. Kurman RJ, Solomon D. The Bethesda System for Reporting Cervical/Vaginal Cytologic Diagnoses. Definitions, Criteria, and Explanatory Notes for Terminology and Specimen Adequacy. New York: Springer-Verlag, 1994;64–73.
3. Kurman RJ, Henson DE, Herbst AL, Noller KL, Schiffman MH. Interim guidelines for management of abnormal cervical cytology: The 1992 National Cancer Institute Workshop. JAMA 1994; 271: 1866 – 9.
4. Bose S, Kannan V, Kline TS. Abnormal endocervical cells. Really abnormal? Really endocervical? Am J Clin Pathol 1994; 101: 708 – 13.
5. Duska LR, Flynn CF, Chen A, Whall-Strojwas D, Goodman A. Clinical evaluation of atypical glandular cells of undetermined significance on cervical cytology. Obstet Gynecol 1998; 91: 278 – 82.
6. Eddy GL, Strumpf KB, Wojtowycz MA, Piraino PS, Mazur MT. Biopsy findings in five hundred thirty-one patients with atypical glandular cells of uncertain significance as defined by the Bethesda system. Am J Obstet Gynecol 1997; 177: 1188 – 95.
7. Goff BA, Atanasoff P, Brown E, Muntz HG, Bell DA, Rice LW. Endocervical glandular atypia in Papanicolaou smears. Obstet Gynecol 1992; 79: 101 – 4.
8. Kennedy AW, Salmieri SS, Wirth SL, Biscotti CV, Tuason LJ, Travarca MJ. Results of the clinical evaluation of atypical glandular cells of undetermined significance (AGUS) detected on cervical cytology screening. Gynecol Oncol 1996; 63: 14 – 8.
9. Lee KR, Manna EA, St. John T. Atypical endocervical glandular cells: accuracy of cytologic diagnosis. Diagn Cytopathol 1995; 13: 202 – 8.
10. Nasu I, Meurer W, Fu YS. Endocervical glandular atypia and adenocarcinoma: a correlation of cytology and histology. Int J Gynecol 1993; 12: 208 – 18.
11. Manetta A, Keefe K, Lin F, Ahdoot D, Kaleb V. Atypical glandular cells of undetermined significance in cervical cytologic findings. Am J Obstet Gynecol 1999; 180: 883 – 8.
12. Zweizig S, Noller K, Reale F, Collis S, Resseguie L. Neoplasia associated with atypical glandular cells of undetermined significance on cervical cytology. Gynecol Oncol 1997; 65: 314 – 8.
13. Medalie NS, Wasserman P. Atypical glandular cells of undetermined significance: The experience at Long Island Jewish Medical Center. J Lower Gen Tract Dis 1998; 2: 127 – 31.
14. Chhieng DC, Elgert P, Cohen J-M, Cangiarella JF. Clinical significance of atypical glandular cells of undetermined significance in postmenopausal women. Cancer Cytology 2001; 93: 1 – 7.