- Email: [email protected]
Effect of Cervical Disease History on Outcomes of Women Who Have a Pap Diagnosis of Atypical Glandular Cells of Undetermined Significance
Gynecologic Oncology 74, 460 – 464 (1999) Article ID gyno.1999.5512, available online at http://www.idealibrary.com on
Effect of Cervical Disease History on Outcomes of Women Who Have a Pap Diagnosis of Atypical Glandular Cells of Undetermined Significance Stephen S. Raab, M.D.,* ,1 N. Scott Bishop, Ed.S.,† ,2 and M. Sue Zaleski, CT(ASCP)† *Department of Pathology and Laboratory Medicine and the Division of Clinical Effectiveness and Outcomes Research, Allegheny General Hospital, 320 East North Avenue, Pittsburgh, Pennsylvania 15212; and the †Department of Pathology, University of Iowa Hospitals and Clinics, Iowa City, Iowa 52242 Received March 18, 1999
INTRODUCTION
Objective. To determine the relationship between history of cervical dysplasia or carcinoma and the development of cervical dysplasia or adenocarcinoma in women who have a diagnosis of atypical glandular cells of undetermined significance (AGUS), favor endocervical origin, or AGUS, not otherwise specified. Methods. A 6-year retrospective review of the pathology files was performed for 93 women who were diagnosed in 1992 with AGUS, favor endocervical origin, or AGUS, not otherwise specified. Data collected included previous history of cervical disease, follow-up diagnoses, time interval between follow-ups, and procedures performed. Results. Of women with follow-up who had or did not have a previous history of cervical dysplasia, 32.0 and 12.0%, respectively, developed a squamous dysplasia or adenocarcinoma in situ. This difference was statistically significant (P < 0.05). Of the women who had or did not have a previous history of cervical dysplasia and had Pap smear follow-up, only 4.2 and 4.3%, respectively, had a false-negative diagnosis on the most immediate subsequent smear. Conclusions. Women who have AGUS, favor endocervical origin, or AGUS, not otherwise specified, and no history of cervical dysplasia have a significantly lower risk of developing or having cervical dysplasia than women who have the same diagnosis and a history of cervical dysplasia. This may warrant different treatment protocols for these two groups. For the women with AGUS and no previous history of cervical dysplasia, a repeat Pap smear, rather than colposcopy with curettage, may be warranted. © 1999 Academic Press Key Words: atypical glandular cells of undetermined significance; Pap smear; cervical cancer; diagnosis; Bethesda system; dysplasia.
1
The frequency of a diagnosis of atypical glandular cells of undetermined significance (AGUS) ranges from 0.46 to 1.8% [1–11]. Conservatively assuming that one in four American women have an annual Pap smear (approximately 35 million women), at least 161,000 American women receive an AGUS diagnosis annually [12]. The AGUS category is viewed as high risk because up to 83% of women with AGUS have a squamous intra-epithelial lesion (SIL) or a glandular neoplasm on follow-up [1–11, 13]. Consequently, the American Society for Colposcopy and Cervical Pathology recently published guidelines that women with AGUS should undergo cervical and vaginal colposcopy with endocervical curettage [14]. Investigators have attempted to determine if there are subpopulations of women with AGUS who have different risks of having cervical dysplasia or carcinoma, obviating the need for aggressive follow-up for all women with AGUS [2– 4, 9, 15–20]. One focus has been on cytologically subclassifying AGUS lesions into probably neoplastic and favor reactive categories [1, 21]. Although criteria for this separation have been presented [22, 23], some researchers have shown that the favor reactive category has a specificity of only 58% [21], indicating that a significant number of clinically significant lesions are placed in the favor reactive category. A second focus of investigation has been on identifying demographic risk factors, although, recently, Zweizig et al. [9] found no association between age, parity, gravidity, menopausal status, bleeding, or contraceptive use and follow-up of cervical dysplasia or carcinoma. In this study, we measured the association of cervical dysplasia history and outcome of women who had a 1992 AGUS diagnosis, the population in which an AGUS diagnosis was most likely to occur, and the likelihood that clinically significant disease was missed on a repeat Pap smear.
To whom correspondence and reprint requests should be addressed at Medical College of Pennsylvania and Hahnemann University School of Medicine, Allegheny General Hospital, 320 East North Avenue, Pittsburgh, PA 15212– 4772. Fax: (412) 359 –3598. E-mail: [email protected]. 2 Current address: Department of Education, University of Iowa, Iowa City, IA 52242. 0090-8258/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
MATERIALS AND METHODS
460
In 1992 at the University of Iowa, 19,912 Pap smears were screened and the diagnoses were 0.3% unsatisfactory for diag-
461
CERVICAL DYSPLASIA HISTORY AND AGUS
nosis; 88.5% benign (e.g., within normal limits, benign cellular changes, repair); 5.0% atypical; 4% low grade SIL; 2.0% high-grade SIL; 0.2% suspicious for carcinoma or carcinoma. Of the 959 atypical diagnoses, there were 102 AGUS, favor endocervical origin or AGUS, not otherwise specified, and two atypical squamous cells of undetermined significance (ASCUS) combined with AGUS, favor endocervical origin [22, 23]. These latter two women were excluded to examine a pure AGUS population. Atypical endometrial cells of undetermined significance [22, 23] (a category we call AGUS, favor endometrial cell origin) and AGUS, suspicious for endometrial adenocarcinoma, cases were excluded to focus on the nonendometrial AGUS population. Subclassification of AGUS, favor endocervical origin, cases into favor reactive and probably neoplastic categories was not done in 1992. Of the 102 women who had an AGUS diagnosis, 9 women had a surgical biopsy performed on the same day as the Pap smear. These latter 9 women were excluded to examine a population who had AGUS and no other confirmation of disease (in addition, there may have been a bias in the diagnosis of the latter 9 women, since the surgical pathology specimens were interpreted prior to Pap smear interpretation [1]). Thus, Pap smears from 93 women served as the study set and for simplification in terminology were considered to be the AGUS study population. The University of Iowa Review Board approved this research. Pathology follow-up on these 93 women was obtained by review of the surgical pathology and cytology database (CoMed Utilities Version 6.5, Intersystems Co., Cambridge, MA, and PathNet Laboratory Information System, Cerner International, Kansas City, MO). The switch from CoMed Utilities to PathNet occurred in August 1996 and follow-up was obtained until January 1, 1998. Thus, possible follow-up of 6 years was available for all women. Follow-up data collected included the number of additional procedures after the AGUS diagnosis, types of procedures, follow-up diagnoses, time to follow-up diagnosis, patient age, and previous clinical history. Particular attention was paid to the most significant follow-up diagnosis. For women who had multiple follow-up tests over time (e.g., additional Pap smears, colposcopic biopsy specimens, or cone biopsy specimens), the most severe diagnosis (either by cytopathology or by surgical pathology) was considered to be the most significant follow-up. Choosing the most severe follow-up was a bias toward the AGUS category being of higher risk. Based on previous history, the women were classified into two risk groups: (1) a low-risk group composed of women who had no history of cervical dysplasia (i.e., no history of SIL) and (2) a high-risk group composed of women who had a history of cervical dysplasia (i.e., history of SIL) [24]. This separation was based on observation of a previous cervical dysplasia or carcinoma diagnosis in the cytology or surgical pathology database or the history of cervical dysplasia or carcinoma on either the 1992 or previous pathology requisition forms. The separation into risk groups was made to determine if the histologic and cytologic follow-ups were different for the two groups. Thus, we wanted to
TABLE 1 Number of Follow-up Procedures in Women Who Had a Diagnosis of Atypical Glandular Cells of Undetermined Significance Patient risk category Low risk (n 5 64)
High risk (n 5 29)
Variable
No. of women
Mean per woman
No. of women
Mean per woman
Follow-up procedures Follow-up Pap smears Follow-up biopsies
52 50 32
4.8 3.8 1.8
25 25 13
5.1 3.9 2.1
determine if women who had an AGUS diagnosis and a history of cervical dysplasia had a higher, lower, or equal probability of having a SIL on follow-up as women who had an AGUS diagnosis and no history of cervical dysplasia. If there were significant differences between these groups, depending on risk, women possibly could be triaged to different follow-up protocols. A one-sided Fisher exact test was used to measure statistical significance, because the initial hypothesis was that high risk women would have a higher probability of clinically significant disease on follow-up. RESULTS Of the women who had an AGUS diagnosis, 64 (68.8%) were low risk and 29 (31.2%) were high risk. The mean ages of low- and high-risk women were 37.8 and 36.2, respectively. The number and type of follow-up specimens per woman are listed in Table 1. The percentage of low- and high-risk women who had follow-up was 78.5 and 86.2%, respectively. The mean number of months to the first and last follow-up for the low-risk women was 5.4 and 33.2, respectively. The mean number of months to the first and last follow-up for the high-risk women was 6.1 and 26.7, respectively. Follow-up of the low-risk women who had a diagnosis of AGUS is shown in Table 2. Of those with follow-up, 12.0% (6) had a diagnosis of SIL and 6.0% (3) had a diagnosis of high grade SIL [all 3 had either severe dysplasia or squamous carcinoma in situ (CIN III)]. Of the low-risk women classified as having benign follow-up based on Pap smears alone (i.e., no surgical biopsy follow-up), the mean number of subsequent Pap smears and follow-up time were 2.9 and 31.9 months, respectively. Follow-up of the high-risk women who had a diagnosis of AGUS is shown in Table 3. Of those with follow-up, 32.0% (8) had a diagnosis of SIL or adenocarcinoma in situ and 24.0% (6) had a diagnosis of a high-grade SIL [4 cervical intraepithelial neoplasia II (CIN II) and 1 CIN III] or adenocarcinoma in situ [1]. Of the high-risk women classified as having benign follow-up based on Pap smears alone (i.e., no surgical biopsy
462
RAAB, BISHOP, AND ZALESKI
TABLE 2 Follow-up of Low-Risk Women Who Had a Diagnosis of Atypical Glandular Cells of Undetermined Significance (n 5 50) Most severe follow-up Benign
LSIL a
HSIL
ASCUS
AGUS
20 21 41
0 3 3
0 3 3
2 0 2
1 0 1
Cytology alone b Biopsy alone or biopsy and cytology Total
a LSIL, low-grade squamous intraepithelial lesion; HSIL, high-grade squamous intraepithelial lesion; ASCUS, atypical squamous cells of undetermined significance; AGUS, atypical glandular cells of undetermined significance. b Cytology alone: The worst follow-up diagnosis was established on cytology alone (i.e., either no surgical pathology follow-up was obtained or the surgical pathology follow-up was negative). Biopsy alone or biopsy and cytology: The worst follow-up diagnosis was established on the basis of a surgical pathology procedure or on a combination of a cytology and a surgical pathology procedure.
follow-up), the mean number of subsequent Pap smears and follow-up time were 2.7 and 25.1 months, respectively. The difference between the high- and low-risk groups in the probability of having or developing cervical dysplasia (based on follow-up) was statistically significant (P 5 0.04). In the low-risk women, the mean and median time interval before the worst diagnosis was established were 374 and 197 days, respectively. In 5 of 7 women who had a subsequent follow-up Pap smear (rather than a surgical procedure), a nonbenign diagnosis was made; in 5 of the 7 women, the next Pap smear diagnosis was the worst diagnosis. Assuming that the 41 women with benign follow-up truly had no disease (including those with benign Pap smears and no surgical follow-up), 2 of 48 (4.2%) women had a false-negative diagnosis on the subsequent follow-up smear. In high-risk women, the mean and median time interval before the worst diagnosis was established were 150 and 98 days, respectively. In 7 of 8 women who had a subsequent follow-up Pap smear (rather than a surgical procedure), a nonbenign diagnosis was made; in 4 of the 8 women, the next Pap smear diagnosis was the worst diagnosis. Assuming that the 15 women with benign follow-up truly had no disease (including those with benign Pap smears and no surgical follow-up), 1 of 23 (4.3%) women had a false negative diagnosis on the subsequent follow-up smear.
DISCUSSION The Pap smear diagnosis of AGUS, like all Bethesda System diagnostic categories [22, 23], confers a relative risk of having or developing a clinically significant lesion. The literature is controversial regarding the percentage of women with AGUS who have a clinically significant lesion on follow-up, with atypical follow-up rates ranging from 14 to 83% [1–11, 13]. Knowing the relative risk is important from a practical standpoint, because this risk affects subsequent follow-up protocols. Because the risk associated with an AGUS diagnosis is sometimes reported to be high, aggressive follow-up protocols (e.g., colposcopy with endocervical curettage) have been advocated [14]. The data from this study show that women who have a previous history of cervical dysplasia have a higher probability of having or developing a clinically significant disease than women who have no history of cervical disease. Women who have a history of cervical dysplasia also have a higher risk of having or developing high-grade SIL or adenocarcinoma in situ, abnormalities generally treated more aggressively than other abnormalities such as low-grade SIL and ASCUS. In fact, in women with AGUS who have no history of cervical dysplasia, the risk of having or developing a high-grade SIL is slightly lower than the risk associated with the diagnosis of
TABLE 3 Follow-up of High-Risk Women (n 5 25) Most severe follow-up
Cytology alone b Biopsy alone or biopsy and cytology Total
Benign
LSIL a
HSIL
AIS
ASCUS
AGUS
9 6 15
0 2 2
0 5 5
0 1 1
1 0 1
1 0 1
a LSIL, low-grade squamous intraepithelial lesion; HSIL, high-grade squamous intraepithelial lesion; AIS, adenocarcinoma in situ; ASCUS, atypical squamous cells of undetermined significance; AGUS, atypical glandular cells of undetermined significance. b Cytology alone: The worst follow-up diagnosis was established on cytology alone (i.e., either no surgical pathology follow-up was obtained or the surgical pathology follow-up was negative). Biopsy alone or biopsy and cytology: The worst follow-up diagnosis was established on the basis of a surgical pathology procedure or on a combination of a cytology and a surgical pathology procedure.
CERVICAL DYSPLASIA HISTORY AND AGUS
ASCUS [24, 25]. Clearly, additional studies are needed to determine if these women with AGUS should receive the same follow-up as women with ASCUS, which is itself controversial [25]. Some have argued that women who have an ASCUS diagnosis be followed with a repeat Pap smear rather than undergo colposcopy [25]. Based on these data, it appears that women with AGUS may be subclassified into different risk groups; those women who have a history of cervical dysplasia theoretically could have more aggressive follow-up. One difficulty in interpreting AGUS studies is that the Bethesda system recommendations for AGUS subclassification are not always presented. The Bethesda Committee recommended that AGUS be subclassified into favor endocervical, favor endometrial, and not otherwise specified categories [22]. The AGUS, favor endometrial, category was specifically excluded from this study, because some studies have shown the follow-up of these lesions to be different from the follow-up of the other categories (i.e, squamous dysplasias rarely are seen in the AGUS, favor endometrial, category) [3]. An attempt to specifically subclassify AGUS, favor endocervical, and AGUS, not otherwise specified, by cytologic means into different risk groups, thereby allowing for different follow-up protocols, was not successful [21]. Women who have AGUS, favor endometrial, may be older and have a risk of having endometrial hyperplasia or carcinoma [9]. Zweizig et al. showed no statistical association between several variables, including age, and outcome, but nonetheless recommended endometrial biopsy in the older population [9]. Zweizig et al. reported that other studies [5, 7, 8] showed that endometrial disease occurred in an older population, although none of these studies tested for statistical significance of age as a factor, and the Kennedy et al. study [8] showed only one endometrial adenocarcinoma. Our study did not examine age as a factor, although it is interesting to note that follow-up in our study showed no clinically significant endometrial disease. This may be related to our study inclusion criteria. In 1992, follow-up data for women with AGUS were scarce, and this may have led to less aggressive follow-up in the University of Iowa AGUS population. Of the 19 women who had nonbenign follow-up, a Pap smear was the subsequent most immediate means of follow-up in 16 (only 4 women had a biopsy instead of a repeat smear). In only 3 (18.8%) of these 16 women was the subsequent Pap smear interpreted as benign and only one of these 3 women had a SIL on extended follow-up (one had ASCUS and one had AGUS). Thus, in the majority of women who had AGUS, repeat Pap smears were able to establish if a lesion was present. In other words, a false negative was unlikely if the Pap smear was repeated, and in most cases, the true nature of the lesion declared itself. This was true for all the women, regardless of their cervical dysplasia history. This raises the possibility that women with AGUS could be followed by repeat smears, with those having atypical diagnoses on the second smear being referred for more aggressive follow-up. Clearly, this conclusion is contrary to the
463
general opinion in the literature [1] and needs further investigation. The 1992 University of Iowa frequency of AGUS, favor endocervical origin, and AGUS, not otherwise specified (0.51%), falls within the general AGUS frequency range reported in the literature (range: 0.17–1.8%, mean: 0.48%) [1–11]. However, only 25.3% (19 of 75 women with followup) of University of Iowa women had a clinically significant diagnosis, which is on the low end of the range reported by Veljovich et al. [1] (range: 13.9 –54.0%; mean: 43.9% [5–11]). Many of these studies used different inclusion criteria which affected overall AGUS follow-up results [5–11]. In our study, AGUS, favor endometrial origin, cases were specifically excluded because these cases have a different cytologic appearance from other AGUS cases and at the University of Iowa, follow-up for these AGUS women may include an endometrial curettage [14]. Women with AGUS, favor endocervical origin, or AGUS, not otherwise specified, generally do not have an endometrial curettage (although this is somewhat dependent on age and other clinical factors) [14]. AGUS cases that have adenocarcinoma or hyperplasia follow-up generally fall within the AGUS, favor endometrial origin, category [3], whereas the cases with adenocarcinoma in situ follow-up generally fall within the AGUS, favor endocervical origin, and AGUS, not otherwise specified, categories [3, 21, 24]. If the cases with follow-up of adenocarcinoma or hyperplasia are excluded from these other reports (assuming these cases were diagnosed or could have been diagnosed as AGUS, favor endometrial origin), our data appear even more in line with the other reported results (e.g., Veljovich et al. [1] reported that 25.1% of AGUS cases had follow-up of SIL or adenocarcinoma in situ). Including AGUS, favor endometrial origin, cases may lead to different overall follow-up results and the consequent conclusion that all AGUS women should have aggressive follow-up. It is uncertain why some laboratories do not use this AGUS subclassification system or separately report these categories in published reports. Some laboratories use the Bethesda Committee-recommended classification system of probably neoplastic and favor reactive [21, 22]. This designation was intended to subclassify AGUS, favor endocervical origin, cases and the Bethesda Committee made no recommendation regarding this designation for AGUS, favor endometrial origin, cases [21, 22]. The favor reactive and probably neoplastic categories generally are not used at the University of Iowa because of the low specificity and sensitivity of these categories [21]. The data from this study cannot be used to comment on this Bethesda recommendation, but it is important to know how laboratories use the AGUS categories before any study may be effectively interpreted. The ability to compare AGUS data across laboratories is controversial [21]. One study showed that when four expert cytopathologists reviewed 100 cases originally called AGUS, favor endocervical origin, or AGUS, not otherwise specified, all four cytopathologists concurred with the AGUS diagnosis
464
RAAB, BISHOP, AND ZALESKI
in only 15% of cases and all four cytopathologists used a different Bethesda System diagnosis (when reclassifying the AGUS cases) from each other in 45% of cases [21]. Interobserver variability using the AGUS category is significantly higher than the interobserver variability of the SIL categories [21, 26]. Some cytopathologists use the AGUS category more frequently than others, and it is possible that University of Iowa adenocarcinoma cases were missed or classified in other Bethesda categories (e.g., AGUS, favor endometrial origin). Similar to the ASCUS category, the AGUS category includes a spectrum of benign and neoplastic lesions. Confirming previously reported data, this study showed that women with AGUS with clinically significant follow-up are more likely to have squamous rather than glandular abnormalities [1–11]. However, in contrast to previously published reports, women who have a diagnoses of AGUS, favor endocervical origin, or AGUS, not otherwise specified, may be stratified into two risk groups on the basis of previous clinical history. Additional studies are necessary to determine the most appropriate follow-up protocols. Also, the initial follow-up smear in these AGUS women was highly unlikely to be a false negative. REFERENCES 1. 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 179:382–390, 1998 2. Bose S, Kannan V, Kline T: Abnormal endocervical cells: Really abnormal? Really endocervical? Am J Clin Pathol 101:708 –713, 1994 3. 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 104:574 –582, 1995 4. DiTomasso JP, Ramzy I, Mody DR: Glandular lesions of the cervix: Validity of cytologic criteria used to differentiate reactive changes, glandular intraepithelial lesions and adenocarcinoma. Acta Cytol 40:1127– 1135, 1996 5. Goff BA, Atanasoff P, Brown E, Muntz HG, Bell DA, Rice LW: Endocervical glandular atypia in Papanicolaou smears. Obstet Gynecol 79:101– 104, 1992 6. Taylor RR, Guerrieri JP, Nash JD, Henry MR, O’Connor DM: Atypical cervical cytology: Colposcopic follow-up using the Bethesda system. J Reprod Med 38:443– 447, 1993 7. Nasu I, Meurer W, Fu YS: Endocervical glandular atypia and adenocarcinoma: A correlation of cytology and histology. Int J Gynecol Pathol 12:208 –218, 1993 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 63:14 –18, 1996 9. Zweizig S, Noller K, Reale F, Collis S, Resseguie L: Neoplasia associated
10.
11.
12. 13.
14. 15.
16.
17.
18.
19.
20.
21.
22. 23. 24.
25.
26.
with atypical glandular cells of undetermined significance on cervical cytology. Gynecol Oncol 65:314 –318, 1997 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 177:1188 –1195, 1997 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 91:278 –282, 1998 U.S. Bureau of the Census: Current Population Reports, Series P23, No. 187, 1994 Currie MM, Cason Z, Baliga M, Lemos LB: The significance of atypical glandular cells on Papanicolaou smears: A 8-year follow-up study. Acta Cytol 38:810, 1994 (abstract) Cox JT: ASCP practice guidelines: Management of glandular abnormalities in the cervical smear. J Lower Genital Tract Dis 1:41– 45, 1997 Raab SS, Snider TE, Potts SA, McDaniel H, Robinson RA, Nelson DL, Thomas PA: Atypical glandular cells of undetermined significance: Diagnostic accuracy and interobserver variability using select cytologic criteria. Am J Clin Pathol 107:299 –307, 1997 van Aspert-van Erp AJM, van’t Hof-Grootenboer AB, Brugal G, Vooijs GP: Endocervical columnar cell intraepithelial neoplasia. I. Discriminating cytomorphologic criteria. Acta Cytol 39:1199 –1215, 1995 van Aspert-van Erp AJM, van’t Hof-Grootenboer AB, Brugal G, Vooijs GP: Endocervical columnar cell intraepithelial neoplasia. II. Grades of expression of cytomorphologic criteria. Acta Cytol 39:1216 –1232, 1995 Novotny DB, Maygarden SJ, Johnson DE, Frable WJ: Tubal metaplasia: A frequent potential pitfall in the cytologic diagnosis of endocervical glandular dysplasia on cervical smears. Acta Cytol 36:1–10, 1992 Pisharodi LR, Ramirez N, Gudlaugsson E, Lipinski D, Bedrossian C: The spectrum and significance of atypical glandular cells of undetermined significance (AGUS) on Papanicolaou smears. Acta Cytol 38:794 –795, 1994 Babkowski RC, Wilbur DC, Rutkowski MA, Facik MS, Bonfiglio TA: The effects of endocervical canal topography, tubal metaplasia, and high canal sampling on the cytologic presentation of nonneoplastic endocervical cells. Am J Clin Pathol 105:403– 410, 1996 Raab SS, Geisinger KR, Silverman JF, Thomas PA, Stanley MW: Interobserver variability of a Pap smear diagnosis of atypical glandular cells of undetermined significance (AGUS). Am J Clin Pathol 110:653– 659, 1998 The Bethesda Committee: The Bethesda system for reporting cervical/ vaginal cytologic diagnoses. Acta Cytol 37:115–124, 1993 The Bethesda Committee: The Bethesda System for Reporting Cervical/ Vaginal Diagnoses. New York, Springer-Verlag, 1994 Raab SS, Bishop NS, Zaleski MS: Long term outcomes and relative risk of women who have a diagnosis of atypical squamous cells of undetermined significance. Am J Clin Pathol 112:57– 62, 1999. Raab SS, Steiner AL, Hornberger J: The cost effectiveness of treating women with a cervical vaginal smear diagnosis of atypical squamous cells of undetermined significance. Am J Obstet Gynecol 179:411– 420, 1998 Sherman ME, Schiffman MH, Lorincz AT, Manos MM, Scott DR, Kurman RJ, Kiviat NB, Stoler M, Glass AG, Rush BB: Toward objective quality assurance in cervical cytopathology: Correlation of cytopathologic diagnoses with detection of high-risk human papillomavirus types. Am J Clin Pathol 102:182–187, 1994
Effect of Cervical Disease History on Outcomes of Women Who Have a Pap Diagnosis of Atypical Glandular Cells of Undetermined Significance Stephen S. Raab, M.D.,* ,1 N. Scott Bishop, Ed.S.,† ,2 and M. Sue Zaleski, CT(ASCP)† *Department of Pathology and Laboratory Medicine and the Division of Clinical Effectiveness and Outcomes Research, Allegheny General Hospital, 320 East North Avenue, Pittsburgh, Pennsylvania 15212; and the †Department of Pathology, University of Iowa Hospitals and Clinics, Iowa City, Iowa 52242 Received March 18, 1999
INTRODUCTION
Objective. To determine the relationship between history of cervical dysplasia or carcinoma and the development of cervical dysplasia or adenocarcinoma in women who have a diagnosis of atypical glandular cells of undetermined significance (AGUS), favor endocervical origin, or AGUS, not otherwise specified. Methods. A 6-year retrospective review of the pathology files was performed for 93 women who were diagnosed in 1992 with AGUS, favor endocervical origin, or AGUS, not otherwise specified. Data collected included previous history of cervical disease, follow-up diagnoses, time interval between follow-ups, and procedures performed. Results. Of women with follow-up who had or did not have a previous history of cervical dysplasia, 32.0 and 12.0%, respectively, developed a squamous dysplasia or adenocarcinoma in situ. This difference was statistically significant (P < 0.05). Of the women who had or did not have a previous history of cervical dysplasia and had Pap smear follow-up, only 4.2 and 4.3%, respectively, had a false-negative diagnosis on the most immediate subsequent smear. Conclusions. Women who have AGUS, favor endocervical origin, or AGUS, not otherwise specified, and no history of cervical dysplasia have a significantly lower risk of developing or having cervical dysplasia than women who have the same diagnosis and a history of cervical dysplasia. This may warrant different treatment protocols for these two groups. For the women with AGUS and no previous history of cervical dysplasia, a repeat Pap smear, rather than colposcopy with curettage, may be warranted. © 1999 Academic Press Key Words: atypical glandular cells of undetermined significance; Pap smear; cervical cancer; diagnosis; Bethesda system; dysplasia.
1
The frequency of a diagnosis of atypical glandular cells of undetermined significance (AGUS) ranges from 0.46 to 1.8% [1–11]. Conservatively assuming that one in four American women have an annual Pap smear (approximately 35 million women), at least 161,000 American women receive an AGUS diagnosis annually [12]. The AGUS category is viewed as high risk because up to 83% of women with AGUS have a squamous intra-epithelial lesion (SIL) or a glandular neoplasm on follow-up [1–11, 13]. Consequently, the American Society for Colposcopy and Cervical Pathology recently published guidelines that women with AGUS should undergo cervical and vaginal colposcopy with endocervical curettage [14]. Investigators have attempted to determine if there are subpopulations of women with AGUS who have different risks of having cervical dysplasia or carcinoma, obviating the need for aggressive follow-up for all women with AGUS [2– 4, 9, 15–20]. One focus has been on cytologically subclassifying AGUS lesions into probably neoplastic and favor reactive categories [1, 21]. Although criteria for this separation have been presented [22, 23], some researchers have shown that the favor reactive category has a specificity of only 58% [21], indicating that a significant number of clinically significant lesions are placed in the favor reactive category. A second focus of investigation has been on identifying demographic risk factors, although, recently, Zweizig et al. [9] found no association between age, parity, gravidity, menopausal status, bleeding, or contraceptive use and follow-up of cervical dysplasia or carcinoma. In this study, we measured the association of cervical dysplasia history and outcome of women who had a 1992 AGUS diagnosis, the population in which an AGUS diagnosis was most likely to occur, and the likelihood that clinically significant disease was missed on a repeat Pap smear.
To whom correspondence and reprint requests should be addressed at Medical College of Pennsylvania and Hahnemann University School of Medicine, Allegheny General Hospital, 320 East North Avenue, Pittsburgh, PA 15212– 4772. Fax: (412) 359 –3598. E-mail: [email protected]. 2 Current address: Department of Education, University of Iowa, Iowa City, IA 52242. 0090-8258/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
MATERIALS AND METHODS
460
In 1992 at the University of Iowa, 19,912 Pap smears were screened and the diagnoses were 0.3% unsatisfactory for diag-
461
CERVICAL DYSPLASIA HISTORY AND AGUS
nosis; 88.5% benign (e.g., within normal limits, benign cellular changes, repair); 5.0% atypical; 4% low grade SIL; 2.0% high-grade SIL; 0.2% suspicious for carcinoma or carcinoma. Of the 959 atypical diagnoses, there were 102 AGUS, favor endocervical origin or AGUS, not otherwise specified, and two atypical squamous cells of undetermined significance (ASCUS) combined with AGUS, favor endocervical origin [22, 23]. These latter two women were excluded to examine a pure AGUS population. Atypical endometrial cells of undetermined significance [22, 23] (a category we call AGUS, favor endometrial cell origin) and AGUS, suspicious for endometrial adenocarcinoma, cases were excluded to focus on the nonendometrial AGUS population. Subclassification of AGUS, favor endocervical origin, cases into favor reactive and probably neoplastic categories was not done in 1992. Of the 102 women who had an AGUS diagnosis, 9 women had a surgical biopsy performed on the same day as the Pap smear. These latter 9 women were excluded to examine a population who had AGUS and no other confirmation of disease (in addition, there may have been a bias in the diagnosis of the latter 9 women, since the surgical pathology specimens were interpreted prior to Pap smear interpretation [1]). Thus, Pap smears from 93 women served as the study set and for simplification in terminology were considered to be the AGUS study population. The University of Iowa Review Board approved this research. Pathology follow-up on these 93 women was obtained by review of the surgical pathology and cytology database (CoMed Utilities Version 6.5, Intersystems Co., Cambridge, MA, and PathNet Laboratory Information System, Cerner International, Kansas City, MO). The switch from CoMed Utilities to PathNet occurred in August 1996 and follow-up was obtained until January 1, 1998. Thus, possible follow-up of 6 years was available for all women. Follow-up data collected included the number of additional procedures after the AGUS diagnosis, types of procedures, follow-up diagnoses, time to follow-up diagnosis, patient age, and previous clinical history. Particular attention was paid to the most significant follow-up diagnosis. For women who had multiple follow-up tests over time (e.g., additional Pap smears, colposcopic biopsy specimens, or cone biopsy specimens), the most severe diagnosis (either by cytopathology or by surgical pathology) was considered to be the most significant follow-up. Choosing the most severe follow-up was a bias toward the AGUS category being of higher risk. Based on previous history, the women were classified into two risk groups: (1) a low-risk group composed of women who had no history of cervical dysplasia (i.e., no history of SIL) and (2) a high-risk group composed of women who had a history of cervical dysplasia (i.e., history of SIL) [24]. This separation was based on observation of a previous cervical dysplasia or carcinoma diagnosis in the cytology or surgical pathology database or the history of cervical dysplasia or carcinoma on either the 1992 or previous pathology requisition forms. The separation into risk groups was made to determine if the histologic and cytologic follow-ups were different for the two groups. Thus, we wanted to
TABLE 1 Number of Follow-up Procedures in Women Who Had a Diagnosis of Atypical Glandular Cells of Undetermined Significance Patient risk category Low risk (n 5 64)
High risk (n 5 29)
Variable
No. of women
Mean per woman
No. of women
Mean per woman
Follow-up procedures Follow-up Pap smears Follow-up biopsies
52 50 32
4.8 3.8 1.8
25 25 13
5.1 3.9 2.1
determine if women who had an AGUS diagnosis and a history of cervical dysplasia had a higher, lower, or equal probability of having a SIL on follow-up as women who had an AGUS diagnosis and no history of cervical dysplasia. If there were significant differences between these groups, depending on risk, women possibly could be triaged to different follow-up protocols. A one-sided Fisher exact test was used to measure statistical significance, because the initial hypothesis was that high risk women would have a higher probability of clinically significant disease on follow-up. RESULTS Of the women who had an AGUS diagnosis, 64 (68.8%) were low risk and 29 (31.2%) were high risk. The mean ages of low- and high-risk women were 37.8 and 36.2, respectively. The number and type of follow-up specimens per woman are listed in Table 1. The percentage of low- and high-risk women who had follow-up was 78.5 and 86.2%, respectively. The mean number of months to the first and last follow-up for the low-risk women was 5.4 and 33.2, respectively. The mean number of months to the first and last follow-up for the high-risk women was 6.1 and 26.7, respectively. Follow-up of the low-risk women who had a diagnosis of AGUS is shown in Table 2. Of those with follow-up, 12.0% (6) had a diagnosis of SIL and 6.0% (3) had a diagnosis of high grade SIL [all 3 had either severe dysplasia or squamous carcinoma in situ (CIN III)]. Of the low-risk women classified as having benign follow-up based on Pap smears alone (i.e., no surgical biopsy follow-up), the mean number of subsequent Pap smears and follow-up time were 2.9 and 31.9 months, respectively. Follow-up of the high-risk women who had a diagnosis of AGUS is shown in Table 3. Of those with follow-up, 32.0% (8) had a diagnosis of SIL or adenocarcinoma in situ and 24.0% (6) had a diagnosis of a high-grade SIL [4 cervical intraepithelial neoplasia II (CIN II) and 1 CIN III] or adenocarcinoma in situ [1]. Of the high-risk women classified as having benign follow-up based on Pap smears alone (i.e., no surgical biopsy
462
RAAB, BISHOP, AND ZALESKI
TABLE 2 Follow-up of Low-Risk Women Who Had a Diagnosis of Atypical Glandular Cells of Undetermined Significance (n 5 50) Most severe follow-up Benign
LSIL a
HSIL
ASCUS
AGUS
20 21 41
0 3 3
0 3 3
2 0 2
1 0 1
Cytology alone b Biopsy alone or biopsy and cytology Total
a LSIL, low-grade squamous intraepithelial lesion; HSIL, high-grade squamous intraepithelial lesion; ASCUS, atypical squamous cells of undetermined significance; AGUS, atypical glandular cells of undetermined significance. b Cytology alone: The worst follow-up diagnosis was established on cytology alone (i.e., either no surgical pathology follow-up was obtained or the surgical pathology follow-up was negative). Biopsy alone or biopsy and cytology: The worst follow-up diagnosis was established on the basis of a surgical pathology procedure or on a combination of a cytology and a surgical pathology procedure.
follow-up), the mean number of subsequent Pap smears and follow-up time were 2.7 and 25.1 months, respectively. The difference between the high- and low-risk groups in the probability of having or developing cervical dysplasia (based on follow-up) was statistically significant (P 5 0.04). In the low-risk women, the mean and median time interval before the worst diagnosis was established were 374 and 197 days, respectively. In 5 of 7 women who had a subsequent follow-up Pap smear (rather than a surgical procedure), a nonbenign diagnosis was made; in 5 of the 7 women, the next Pap smear diagnosis was the worst diagnosis. Assuming that the 41 women with benign follow-up truly had no disease (including those with benign Pap smears and no surgical follow-up), 2 of 48 (4.2%) women had a false-negative diagnosis on the subsequent follow-up smear. In high-risk women, the mean and median time interval before the worst diagnosis was established were 150 and 98 days, respectively. In 7 of 8 women who had a subsequent follow-up Pap smear (rather than a surgical procedure), a nonbenign diagnosis was made; in 4 of the 8 women, the next Pap smear diagnosis was the worst diagnosis. Assuming that the 15 women with benign follow-up truly had no disease (including those with benign Pap smears and no surgical follow-up), 1 of 23 (4.3%) women had a false negative diagnosis on the subsequent follow-up smear.
DISCUSSION The Pap smear diagnosis of AGUS, like all Bethesda System diagnostic categories [22, 23], confers a relative risk of having or developing a clinically significant lesion. The literature is controversial regarding the percentage of women with AGUS who have a clinically significant lesion on follow-up, with atypical follow-up rates ranging from 14 to 83% [1–11, 13]. Knowing the relative risk is important from a practical standpoint, because this risk affects subsequent follow-up protocols. Because the risk associated with an AGUS diagnosis is sometimes reported to be high, aggressive follow-up protocols (e.g., colposcopy with endocervical curettage) have been advocated [14]. The data from this study show that women who have a previous history of cervical dysplasia have a higher probability of having or developing a clinically significant disease than women who have no history of cervical disease. Women who have a history of cervical dysplasia also have a higher risk of having or developing high-grade SIL or adenocarcinoma in situ, abnormalities generally treated more aggressively than other abnormalities such as low-grade SIL and ASCUS. In fact, in women with AGUS who have no history of cervical dysplasia, the risk of having or developing a high-grade SIL is slightly lower than the risk associated with the diagnosis of
TABLE 3 Follow-up of High-Risk Women (n 5 25) Most severe follow-up
Cytology alone b Biopsy alone or biopsy and cytology Total
Benign
LSIL a
HSIL
AIS
ASCUS
AGUS
9 6 15
0 2 2
0 5 5
0 1 1
1 0 1
1 0 1
a LSIL, low-grade squamous intraepithelial lesion; HSIL, high-grade squamous intraepithelial lesion; AIS, adenocarcinoma in situ; ASCUS, atypical squamous cells of undetermined significance; AGUS, atypical glandular cells of undetermined significance. b Cytology alone: The worst follow-up diagnosis was established on cytology alone (i.e., either no surgical pathology follow-up was obtained or the surgical pathology follow-up was negative). Biopsy alone or biopsy and cytology: The worst follow-up diagnosis was established on the basis of a surgical pathology procedure or on a combination of a cytology and a surgical pathology procedure.
CERVICAL DYSPLASIA HISTORY AND AGUS
ASCUS [24, 25]. Clearly, additional studies are needed to determine if these women with AGUS should receive the same follow-up as women with ASCUS, which is itself controversial [25]. Some have argued that women who have an ASCUS diagnosis be followed with a repeat Pap smear rather than undergo colposcopy [25]. Based on these data, it appears that women with AGUS may be subclassified into different risk groups; those women who have a history of cervical dysplasia theoretically could have more aggressive follow-up. One difficulty in interpreting AGUS studies is that the Bethesda system recommendations for AGUS subclassification are not always presented. The Bethesda Committee recommended that AGUS be subclassified into favor endocervical, favor endometrial, and not otherwise specified categories [22]. The AGUS, favor endometrial, category was specifically excluded from this study, because some studies have shown the follow-up of these lesions to be different from the follow-up of the other categories (i.e, squamous dysplasias rarely are seen in the AGUS, favor endometrial, category) [3]. An attempt to specifically subclassify AGUS, favor endocervical, and AGUS, not otherwise specified, by cytologic means into different risk groups, thereby allowing for different follow-up protocols, was not successful [21]. Women who have AGUS, favor endometrial, may be older and have a risk of having endometrial hyperplasia or carcinoma [9]. Zweizig et al. showed no statistical association between several variables, including age, and outcome, but nonetheless recommended endometrial biopsy in the older population [9]. Zweizig et al. reported that other studies [5, 7, 8] showed that endometrial disease occurred in an older population, although none of these studies tested for statistical significance of age as a factor, and the Kennedy et al. study [8] showed only one endometrial adenocarcinoma. Our study did not examine age as a factor, although it is interesting to note that follow-up in our study showed no clinically significant endometrial disease. This may be related to our study inclusion criteria. In 1992, follow-up data for women with AGUS were scarce, and this may have led to less aggressive follow-up in the University of Iowa AGUS population. Of the 19 women who had nonbenign follow-up, a Pap smear was the subsequent most immediate means of follow-up in 16 (only 4 women had a biopsy instead of a repeat smear). In only 3 (18.8%) of these 16 women was the subsequent Pap smear interpreted as benign and only one of these 3 women had a SIL on extended follow-up (one had ASCUS and one had AGUS). Thus, in the majority of women who had AGUS, repeat Pap smears were able to establish if a lesion was present. In other words, a false negative was unlikely if the Pap smear was repeated, and in most cases, the true nature of the lesion declared itself. This was true for all the women, regardless of their cervical dysplasia history. This raises the possibility that women with AGUS could be followed by repeat smears, with those having atypical diagnoses on the second smear being referred for more aggressive follow-up. Clearly, this conclusion is contrary to the
463
general opinion in the literature [1] and needs further investigation. The 1992 University of Iowa frequency of AGUS, favor endocervical origin, and AGUS, not otherwise specified (0.51%), falls within the general AGUS frequency range reported in the literature (range: 0.17–1.8%, mean: 0.48%) [1–11]. However, only 25.3% (19 of 75 women with followup) of University of Iowa women had a clinically significant diagnosis, which is on the low end of the range reported by Veljovich et al. [1] (range: 13.9 –54.0%; mean: 43.9% [5–11]). Many of these studies used different inclusion criteria which affected overall AGUS follow-up results [5–11]. In our study, AGUS, favor endometrial origin, cases were specifically excluded because these cases have a different cytologic appearance from other AGUS cases and at the University of Iowa, follow-up for these AGUS women may include an endometrial curettage [14]. Women with AGUS, favor endocervical origin, or AGUS, not otherwise specified, generally do not have an endometrial curettage (although this is somewhat dependent on age and other clinical factors) [14]. AGUS cases that have adenocarcinoma or hyperplasia follow-up generally fall within the AGUS, favor endometrial origin, category [3], whereas the cases with adenocarcinoma in situ follow-up generally fall within the AGUS, favor endocervical origin, and AGUS, not otherwise specified, categories [3, 21, 24]. If the cases with follow-up of adenocarcinoma or hyperplasia are excluded from these other reports (assuming these cases were diagnosed or could have been diagnosed as AGUS, favor endometrial origin), our data appear even more in line with the other reported results (e.g., Veljovich et al. [1] reported that 25.1% of AGUS cases had follow-up of SIL or adenocarcinoma in situ). Including AGUS, favor endometrial origin, cases may lead to different overall follow-up results and the consequent conclusion that all AGUS women should have aggressive follow-up. It is uncertain why some laboratories do not use this AGUS subclassification system or separately report these categories in published reports. Some laboratories use the Bethesda Committee-recommended classification system of probably neoplastic and favor reactive [21, 22]. This designation was intended to subclassify AGUS, favor endocervical origin, cases and the Bethesda Committee made no recommendation regarding this designation for AGUS, favor endometrial origin, cases [21, 22]. The favor reactive and probably neoplastic categories generally are not used at the University of Iowa because of the low specificity and sensitivity of these categories [21]. The data from this study cannot be used to comment on this Bethesda recommendation, but it is important to know how laboratories use the AGUS categories before any study may be effectively interpreted. The ability to compare AGUS data across laboratories is controversial [21]. One study showed that when four expert cytopathologists reviewed 100 cases originally called AGUS, favor endocervical origin, or AGUS, not otherwise specified, all four cytopathologists concurred with the AGUS diagnosis
464
RAAB, BISHOP, AND ZALESKI
in only 15% of cases and all four cytopathologists used a different Bethesda System diagnosis (when reclassifying the AGUS cases) from each other in 45% of cases [21]. Interobserver variability using the AGUS category is significantly higher than the interobserver variability of the SIL categories [21, 26]. Some cytopathologists use the AGUS category more frequently than others, and it is possible that University of Iowa adenocarcinoma cases were missed or classified in other Bethesda categories (e.g., AGUS, favor endometrial origin). Similar to the ASCUS category, the AGUS category includes a spectrum of benign and neoplastic lesions. Confirming previously reported data, this study showed that women with AGUS with clinically significant follow-up are more likely to have squamous rather than glandular abnormalities [1–11]. However, in contrast to previously published reports, women who have a diagnoses of AGUS, favor endocervical origin, or AGUS, not otherwise specified, may be stratified into two risk groups on the basis of previous clinical history. Additional studies are necessary to determine the most appropriate follow-up protocols. Also, the initial follow-up smear in these AGUS women was highly unlikely to be a false negative. REFERENCES 1. 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 179:382–390, 1998 2. Bose S, Kannan V, Kline T: Abnormal endocervical cells: Really abnormal? Really endocervical? Am J Clin Pathol 101:708 –713, 1994 3. 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 104:574 –582, 1995 4. DiTomasso JP, Ramzy I, Mody DR: Glandular lesions of the cervix: Validity of cytologic criteria used to differentiate reactive changes, glandular intraepithelial lesions and adenocarcinoma. Acta Cytol 40:1127– 1135, 1996 5. Goff BA, Atanasoff P, Brown E, Muntz HG, Bell DA, Rice LW: Endocervical glandular atypia in Papanicolaou smears. Obstet Gynecol 79:101– 104, 1992 6. Taylor RR, Guerrieri JP, Nash JD, Henry MR, O’Connor DM: Atypical cervical cytology: Colposcopic follow-up using the Bethesda system. J Reprod Med 38:443– 447, 1993 7. Nasu I, Meurer W, Fu YS: Endocervical glandular atypia and adenocarcinoma: A correlation of cytology and histology. Int J Gynecol Pathol 12:208 –218, 1993 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 63:14 –18, 1996 9. Zweizig S, Noller K, Reale F, Collis S, Resseguie L: Neoplasia associated
10.
11.
12. 13.
14. 15.
16.
17.
18.
19.
20.
21.
22. 23. 24.
25.
26.
with atypical glandular cells of undetermined significance on cervical cytology. Gynecol Oncol 65:314 –318, 1997 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 177:1188 –1195, 1997 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 91:278 –282, 1998 U.S. Bureau of the Census: Current Population Reports, Series P23, No. 187, 1994 Currie MM, Cason Z, Baliga M, Lemos LB: The significance of atypical glandular cells on Papanicolaou smears: A 8-year follow-up study. Acta Cytol 38:810, 1994 (abstract) Cox JT: ASCP practice guidelines: Management of glandular abnormalities in the cervical smear. J Lower Genital Tract Dis 1:41– 45, 1997 Raab SS, Snider TE, Potts SA, McDaniel H, Robinson RA, Nelson DL, Thomas PA: Atypical glandular cells of undetermined significance: Diagnostic accuracy and interobserver variability using select cytologic criteria. Am J Clin Pathol 107:299 –307, 1997 van Aspert-van Erp AJM, van’t Hof-Grootenboer AB, Brugal G, Vooijs GP: Endocervical columnar cell intraepithelial neoplasia. I. Discriminating cytomorphologic criteria. Acta Cytol 39:1199 –1215, 1995 van Aspert-van Erp AJM, van’t Hof-Grootenboer AB, Brugal G, Vooijs GP: Endocervical columnar cell intraepithelial neoplasia. II. Grades of expression of cytomorphologic criteria. Acta Cytol 39:1216 –1232, 1995 Novotny DB, Maygarden SJ, Johnson DE, Frable WJ: Tubal metaplasia: A frequent potential pitfall in the cytologic diagnosis of endocervical glandular dysplasia on cervical smears. Acta Cytol 36:1–10, 1992 Pisharodi LR, Ramirez N, Gudlaugsson E, Lipinski D, Bedrossian C: The spectrum and significance of atypical glandular cells of undetermined significance (AGUS) on Papanicolaou smears. Acta Cytol 38:794 –795, 1994 Babkowski RC, Wilbur DC, Rutkowski MA, Facik MS, Bonfiglio TA: The effects of endocervical canal topography, tubal metaplasia, and high canal sampling on the cytologic presentation of nonneoplastic endocervical cells. Am J Clin Pathol 105:403– 410, 1996 Raab SS, Geisinger KR, Silverman JF, Thomas PA, Stanley MW: Interobserver variability of a Pap smear diagnosis of atypical glandular cells of undetermined significance (AGUS). Am J Clin Pathol 110:653– 659, 1998 The Bethesda Committee: The Bethesda system for reporting cervical/ vaginal cytologic diagnoses. Acta Cytol 37:115–124, 1993 The Bethesda Committee: The Bethesda System for Reporting Cervical/ Vaginal Diagnoses. New York, Springer-Verlag, 1994 Raab SS, Bishop NS, Zaleski MS: Long term outcomes and relative risk of women who have a diagnosis of atypical squamous cells of undetermined significance. Am J Clin Pathol 112:57– 62, 1999. Raab SS, Steiner AL, Hornberger J: The cost effectiveness of treating women with a cervical vaginal smear diagnosis of atypical squamous cells of undetermined significance. Am J Obstet Gynecol 179:411– 420, 1998 Sherman ME, Schiffman MH, Lorincz AT, Manos MM, Scott DR, Kurman RJ, Kiviat NB, Stoler M, Glass AG, Rush BB: Toward objective quality assurance in cervical cytopathology: Correlation of cytopathologic diagnoses with detection of high-risk human papillomavirus types. Am J Clin Pathol 102:182–187, 1994