Sharpless, Kathryn E. MD, PhD*†; Schnatz, Peter F. DO*†; Mandavilli, Srinivas MD*; Greene, John F. MD*†; Sorosky, Joel I. MD*†
In 1988, the Bethesda system introduced the term atypical glandular cells of undetermined significance (AGUS) to describe endocervical or endometrial glandular cells with atypia more severe than that expected from a benign reactive process but lacking the diagnostic features of invasive adenocarcinoma.1 In 1991, the qualifiers “favor reactive” and “favor neoplasia” were added to the category to help differentiate between benign-appearing AGUS Pap tests and those more likely to correlate with malignant pathology.2 In 1998, the International Academy of Cytology Task Force further revised the category by designating endocervical adenocarcinoma in situ as a separate category.3 The 2001 Bethesda system changed the terminology to atypical glandular cells (AGC) and eliminated the misleading “favor reactive” subclassification.4
Atypical glandular cells at Pap test is a challenging classification because of the lack of well-defined cytomorphologic criteria for the interpretation and the difficulty in distinguishing glandular neoplasia from squamous intraepithelial lesions (SIL) with endocervical gland involvement.5 Clinically, the finding of AGC at Pap testing presents a diagnostic challenge because it is associated with a wide variety of squamous and glandular diseases, including many benign, premalignant, and malignant diseases.
The reported rates of dysplasia and malignancy associated with AGC vary widely, ranging from 17% to 59%.6–8 Clinicians who screen their patients for disease rely on these rates to assess their patients’ level of risk. The rates may vary because of study design and because of changing technologies, definitions, and recommendations. This study was designed to estimate the rate of disease in women with an initial AGUS Pap test from a cervical source lacking known cervical or uterine disease or concurrent atypical squamous cells of undetermined significance or SIL. This study also estimates the risk of disease based on the patient's age, Pap test subclassification, the presence of persistent AGUS Pap tests, and method of Pap test collection.
MATERIALS AND METHODS
After approval by the Hartford Hospital Institutional Review Board, the study sample was identified by retrospectively searching the pathology database at Hartford Hospital for cases of AGUS Pap tests during the years 1998 through 2001.9 From the 241,224 Pap tests analyzed and entered into the Hartford Hospital Pathology database during this period, 648 were identified as AGUS (from 634 patients). One hundred fifty-seven women were excluded from the study sample: 23 with a prior AGUS Pap test, 7 who had a total hysterectomy, 14 with known cervical or uterine disease likely accounting for the AGUS Pap test, 34 with concurrent atypical squamous cells of undetermined significance, 2 with concurrent SIL at cytology, and 77 who lacked any follow-up data. Four hundred seventy-seven subjects formed the final study sample, which consisted of women with an initial AGUS Pap test from a cervical source lacking known cervical or uterine disease or concurrent atypical squamous cells of undetermined significance or SIL. Although excluded from the main study, the 34 subjects with concurrent AGUS and atypical squamous cells of undetermined significance at Pap testing were analyzed separately.
The women's ages, city of residence, dates of Pap tests, and Pap test findings along with subclassifications were recorded. Specific demographic information was not available. The study sample was estimated to be approximately 70% suburban, 20% urban, and 10% rural, which is reflective of the patient population at Hartford Hospital.9 The sample of women studied included those managed in private practices, resident continuity practices, and those seen by extended care providers.
Based on their descriptive modifiers (ie, “favor reactive” or “cannot exclude significant endometrial or endocervical pathology”), AGUS Pap tests were categorized into 1 of the following 3 groups: benign-appearing, malignant-appearing, or unspecified. The findings and dates of all subsequent Pap tests and histologic samples were also recorded. The Pap tests were collected either by the liquid-based cytology or conventional method. Histologic specimens included endocervical, endometrial, and cervical samples. Cervical samples were from punch biopsies, cone biopsies (cold-knife or loop electrosurgical excision procedures), or from hysterectomy specimens. Follow-up Pap tests and histologic samples were evaluated up to 24 months from the initial AGUS Pap test in all cases. Human papillomavirus testing data were not obtained. Disease was defined as premalignant or malignant cervical or endometrial disease, high-grade squamous intraepithelial lesions (HSIL) or greater.
Statistical analysis was performed using the χ2 test with Yates’ correction.10 P < .05 were considered statistically significant. Confidence intervals of proportions11 and odds ratios (ORs) and their 95% confidence intervals (CIs)12 were calculated using on-line software.
During the years of this study, there were 648 Pap tests reported as AGUS out of 241,224 Pap tests performed. This results in a 0.3% incidence of AGUS over all 4 years, ranging from 0.2–0.4% for each individual year. We identified disease in 42 of the 477 women with AGUS in our study (9%, 95% CI 7–12%; Table 1) and malignancy in 13 (3%, 95% CI 2–5%). Squamous abnormalities accounted for 55% of the diagnoses; 38% were glandular, and 7% had both squamous and glandular lesions.
Thirty-eight percent of the AGUS Pap tests were collected by the conventional method, whereas 62% were liquid-based cytology. The percent of AGUS Pap tests collected by the liquid-based cytology method increased from 40% in 1998 to 80% in 2001. The rate of disease among women with AGUS Pap tests collected by the liquid-based cytology method was 11%, whereas the rate of disease among women with AGUS Pap tests collected by the conventional method was 6%.
The average age of the 477 subjects in this study was 44 years. One hundred ten women (23%) were aged younger than 35 years, and 367 women (77%) were aged 35 years or older. The average age of the 42 women with disease was 40 years. The odds ratio of women aged younger than 35 years having disease as compared with women aged 35 years or older was 2.2 (95% CI 1.2–4.3, P < .02; Table 2). Squamous lesions accounted for 94% (15/16) of the diagnoses in women aged younger than 35 years. High-grade squamous intraepithelial lesions predominated the diagnoses among the women aged younger than 35 years, and the rate of HSIL in women aged younger than 35 years was more than 6 times that of women aged 35 years or older (OR 7.5, 95% CI 2.9–19.1, P < .01; Fig. 1). The rate of glandular lesions was 1% (1/110) in the younger than 35 years age group and 5% (18/367) in the age group 35 years or older; however, there were not enough cases to show a statistically significant difference. There was a greater diversity of diagnoses among the women aged 35 years or older. In those women, squamous lesions accounted for 31% (8/26) of the diagnoses, glandular lesions accounted for 58% (15/26), and 12% (3/26) were both squamous and glandular. All cases of endometrial cancer, adenocarcinoma in situ, cervical adenosquamous carcinoma, cervical adenocarcinoma, and extrauterine malignancies were in women aged 35 years or older. Women with endometrial cancer had an average age of 54 years and an age range of 35–75 years. The overall rate of malignancy was 3% (12/367) in women aged 35 years or older and 1% (1/110) in those aged younger than 35 years; however, there were not enough cases to show a statistically significant difference.
Based on their descriptive modifiers, we subclassified 12% (55/477) of AGUS Pap tests as malignant-appearing, 74% (351/477) as benign-appearing, and 15% (71/477) as unspecified (Table 3). The rate of disease among women with malignant-appearing AGUS Pap tests (29%) was greater than the rate in women with benign-appearing AGUS Pap tests (5%, OR 8.1, 95% CI 3.8–17.2, P < .01) and in women with unspecified AGUS Pap tests (13%, OR 2.8, 95% CI 1.1–7.0, P < .03). The rate of disease in women with unspecified AGUS Pap tests was greater than in women with benign-appearing AGUS Pap tests (OR 2.9, 95% CI 1.2–6.7, P < .03). The rates of malignancy were as follows: 16% (9/55) in malignant-appearing, 3% (2/71) in unspecified, and 1% (2/351) in benign-appearing Pap tests. The odds ratios of women with malignant-appearing AGUS Pap tests having a malignancy as compared with women with unspecified and benign-appearing AGUS Pap tests were 6.8 (95% CI 1.4–32.7, P < .02) and 34.1 (95% CI 7.2–162.9, P < .01), respectively. There were not enough cases to find a statistically significant difference between the rates of malignancy in the unspecified group and in the benign-appearing group. The rate of endometrial cancer in the malignant-appearing group (9%) was higher than the rate in the unspecified group (1%), but was not statistically significant. There were no cases of endometrial cancer in women with benign-appearing AGUS Pap tests.
There were 29 subjects in our study group who had more than one AGUS Pap test during the study period (persistent AGUS).9 Disease was identified in 31% (9/29) of those women. This rate was higher than the rate in women with a single AGUS Pap test (7% or 33/448), giving an odds ratio of 5.7 (95% CI 2.4–13.4, P < .01). The diagnoses among women with persistent AGUS Pap tests included 4 cases of HSIL and one case each of squamous cell carcinoma, cervical glandular dysplasia, adenocarcinoma in situ, cervical adenosquamous carcinoma, and endometrial adenocarcinoma. The rate of malignancy was 10% (3/29) in women with persistent AGUS Pap tests compared with 3% (11/448) in women with a single AGUS Pap test. The odds ratio of women with persistent AGUS Pap tests having a malignancy as compared with women with single AGUS Pap tests was 4.6 (95% CI 1.2–17.5, P < .05).
Thirty-four subjects with concurrent AGUS and atypical squamous cells of undetermined significance at Pap testing were analyzed separately from the 477 cases of AGUS. There were 6 cases of disease in the concurrent AGUS and atypical squamous cells of undetermined significance group, resulting in a rate of 18% with a 95% CI of 7–35%. The diagnoses in the concurrent AGUS and atypical squamous cells of undetermined significance group were predominately HSIL (67% or 4/6). There were not enough cases of concurrent AGUS and atypical squamous cells of undetermined significance to compare with the AGUS alone group.
The incidence of AGUS (0.3%) in our study was similar to the rates in previous studies, most of which were less than 1%.13 A comparison of our findings with prior large studies of women with AGUS Pap tests is presented in Table 4.14–22 Most large published studies of women with AGUS Pap test have included only women who have had at least 1 type of histologic evaluation. In the present study, we included women with histologic evaluation and women who were followed up by repeat Pap tests only. In Table 4 we present our data similar to prior studies, including only women who underwent histologic evaluation. In this cohort, the rate of disease was 14% (42/308), and the rate of malignancy was 4%. Of the studies we identified, our study had the lowest rate of disease (even when cases of LSIL are included). The rate of malignancy was similar to the rates of prior studies, which ranged from 1% to 12%.14–22
The wide range of rates of disease associated with AGUS may reflect differences in study design, time period, population, and institutional practices in reporting AGUS. We specifically excluded women with a prior AGUS Pap test and with cervical or uterine disease likely accounting for the AGUS Pap test to confine the study to women presenting with their first AGUS Pap test with no apparent underlying cause. Only two of the comparison studies reported similar exclusions.17,20
In this study, we found that a patient's age is helpful in predicting the range of underlying lesions and in directing the diagnostic evaluation. Women aged younger than 35 years had a higher rate of disease and HSIL, whereas women ≥ 35 had a greater variety of diagnoses, including all cases of adenocarcinoma in situ, cervical adenosquamous carcinoma, cervical adenocarcinoma, and extrauterine malignancies. In addition, there were no cases of endometrial cancer in the younger than 35 years age group, and the youngest woman with endometrial adenocarcinoma in the 35 years or older age group was diagnosed at age 35. These findings are consistent with previous studies, which have also found that women aged younger than 35 years with AGUS Pap tests have lower rates of endometrial cancer than women aged 35 years or older.23,24 Our findings, therefore, support the recommendation of evaluating all women aged 35 years or older with AGC at Pap testing with endometrial sampling.5,25
To evaluate the usefulness of AGUS Pap test subclassifications in predicting disease, we grouped the Pap tests based on their descriptive modifiers into 3 subclasses, benign-appearing, malignant-appearing, and unspecified, based on the 1991 Bethesda System in use during the study period. We found that women with malignant-appearing AGUS Pap tests were at greater risk for disease than women with either unspecified or benign-appearing AGUS Pap tests. The 2001 Bethesda System eliminated the misleading “favor reactive” category, effectively combining the “favor reactive” and unspecified categories.4 Based on the 2001 subclassification system, the rate of disease in the unspecified group was 6% (26/422), which is significantly lower than that in the malignant-appearing group (OR 0.2, 95% CI 0.1–0.3). However, the presence of disease, including malignancy, in all groups suggests that neither subclassification system is a completely reliable indicator of disease in women with AGUS Pap tests. Our findings support the practice of fully evaluating all women with AGUS Pap tests regardless of subclassification.
We found that women with persistent AGUS Pap tests had higher rates of disease (31%) and malignancy (10%). Multiple AGUS Pap tests in a woman, therefore, suggest that the underlying lesion has yet to be identified. We also found a higher rate of disease in women with concurrent atypical squamous cells of undetermined significance and AGUS compared with women with AGUS alone; however, these results were not statistically significant. Other studies have reported higher rates with concurrent AGUS and atypical squamous cells of undetermined significance,20,23,26 whereas some studies found similar rates of disease between AGUS as the only diagnosis and concurrent AGUS and atypical squamous cells of undetermined significance.7,15,24,27,28 The reasons for the discrepancies are unclear.
We also found that the disease rate associated with AGUS varied depending on the type of Pap test collection method used. The rate of disease associated with the liquid-based cytology method of Pap test collection was twice that of the conventional method, supporting previous research that shows the liquid-based cytology method is a more sensitive test than the conventional method.29
This study is among the largest retrospective reviews of women with AGUS at Pap testing. Like other similar studies, the study is limited by its retrospective design. Women were not subject to a uniform clinical evaluation, and the lack of thorough evaluation may have missed some diagnoses. The rate of disease was 9% in the total study sample, whereas in the cohort with histologic analysis, the rate was 14%. The former rate may be an under-representation of the actual rate, because many women did not have a thorough evaluation, whereas the latter rate may be inflated, because cases with higher clinical suspicion may have been more likely to have histologic analyses. The actual rate may lie somewhere between 9% and 14%.
The purpose of this study was to estimate the rate of disease associated with AGUS to provide clinicians with information for patients who present with unexplained AGC at Pap test screening. Our findings suggest that the rate of disease may not be as high as previously reported in prior studies and that the rate of disease varies with method of Pap test collection, patient's age, presence of persistent AGUS Pap tests, and AGUS subclassification. Regardless of risk factors, all women with AGUS at Pap testing require evaluation with colposcopy, directed biopsy and endocervical curettage.9 Women aged 35 years or older or with other risk factors for endometrial cancer should have an endometrial biopsy. For women with persistent AGC, or AGC favor neoplasia, a secondary evaluation consisting of a cone biopsy and an evaluation for an extrauterine malignancy should be considered.
1. National Cancer Institute Workshop. The 1988 Bethesda System for reporting cervical/vaginal cytological diagnoses. JAMA 1989;262:931–4.
2. 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.
3. Solomon D, Frable WJ, Vooijs GP, Wilbur DC, Amma NS, Collins RJ, et al. ASCUS and AGUS criteria. International Academy of Cytology Task Force summary. Diagnostic cytology towards the 21st century: an international expert conference and tutorial. Acta Cytol 1998;42:16–24.
4. Solomon D, Davey D, Kurman R, Moriarty A, O'Connor D, Prey M, et al. The 2001 Bethesda system: terminology for reporting results of cervical cytology. JAMA 2002;287:2114–9.
5. Levine, L., Lucci JA III, Van Dinh T. Atypical glandular cells: new Bethesda terminology and management guidelines. Obst Gynecol Surv 2003;58:399–406.
6. Kennedy AW, Salmieri SS, Wirth SL, Biscotti CV, Tuason LJ, Travarca MJ. Results of the clinical evaluation of atypical glandular cells of undetermined significance (AGCUS) detected on cervical cytology screening. Gynecol Oncol 1996;63:14–8.
7. Burja IT, Thompson SK, Sawyer WL Jr., Shurbaji MS. Atypical glandular cells of undetermined significance on cervical smears: a study with cytohistologic correlation. Acta Cytol 1999;43:351–6.
8. Meath AJ, Carley ME, Wilson TO. Atypical glandular cells of undetermined significance: review of final histologic diagnoses. J Reprod Med 2002;47:249–52.
9. Sharpless KE, Schnatz PF, Mandavilli S, Greene JF, Sorosky JI. Lack of adherence to practice guidelines for women with atypical glandular cells on cervical cytology. Obstet Gynecol 2005;105:494–500.
13. Kaferle JE, Malouin JM. Evaluation and management of the AGUS Papanicolaou smear. Am Fam Physician 2001;63:2239–44.
14. Raab SS, Isacson C, Layfield LJ, Lenel JC, Slagel DD, Thomas PA. Atypical glandular cells of undetermined significance: cytologic criteria to separate clinically significant from benign lesions. Am J Clin Pathol 1995;104:574–82.
15. 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.
16. Bennett BB, Takezawa K, Wilkinson EJ, Drew PA, Hardt NS. Atypical glandular cells of undetermined significance and other glandular cell abnormalities in a high-risk population. J Lower Genit Tract Dis 1998;2:132–5.
17. Veljovich DS, Stoler MH, Andersen WA, Covell JL, Rice LW. Atypical glandular cells of undetermined significance: a five-year retrospective histopathologic study. Am J Obstet Gynecol 1998;179:382–90.
18. Dinh TV, Haque M, Gracia JM, Lucci JA, Smith ER. Papanicolaou smears of atypical glandular cells of undetermined significance: histological correlations and suggestions for management. J Lower Genit Tract Dis 1999;3:73–6.
19. Chhieng DC, Elgert P, Cangiarella JF, Cohen JM. Variation in the incidence of AGUS between different patient populations. Acta Cytol 2001;45:287–93.
20. Geier CS, Wilson M, Creasman W. Clinical evaluation of atypical glandular cells of undetermined significance. Am J Obstet Gynecol 2001;184:64–9.
21. Koonings PP, Price JH. Evaluation of atypical glandular cells of undetermined significance: is age important? Am J Obstet Gynecol 2001;184:1457–61.
22. Nasuti JF, Fleisher SR, Gupta PK. Atypical glandular cells of undetermined significance (AGUS): clinical considerations and cytohistologic correlation. Diagn Cytopathol 2002;26:186–90.
23. Chhieng DC, Elgert PA, Cangiarella JF, Cohen JM. Clinical significance of atypical glandular cells of undetermined significance: a follow-up study from an academic medical center. Acta Cytol 2000;44:557–66.
24. 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.
25. Wright TC Jr., Cox JT, Massad LS, Twiggs LB, Wilkinson EJ. 2001 consensus guidelines for the management of women with cervical cytological abnormalities. JAMA 2002;287:2120–9.
26. 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.
27. Chin AB, Bristow RE, Korst LM, Walts A, Lagasse LD. The significance of atypical glandular cells on routine cervical cytologic testing in a community-based population. Am J Obstet Gynecol 2000;182:1278–82.
28. Goff BA, Atanasoff P, Brown E, Muntz HG, Bell DA, Rice LW. Endocervical glandular atypia in Papanicolaou smears. Obstet Gynecol 1992;79:101–4.
29. Abulafia O, Pezzullo JC, Sherer DM. Performance of ThinPrep liquid-based cervical cytology in comparison with conventionally prepared Papanicolaou smears: a quantitative survey. Gynecol Oncol 2003;90:137–44.