Clinical Implications of Atypical Glandular Cells of
Undetermined Significance, Favor Endometrial Origin
David C. Chhieng, M.D.1
Paul Elgert, C.T.2
Jean-Marc Cohen, M.D.3
Joan F. Cangiarella, M.D.4
1Department of Pathology, University of Alabama
at Birmingham, Birmingham, Alabama.
2Cytopathology Laboratory, New York University
Medical Center, New York, New York.
3Department of Pathology, Beth Israel Medical
Center, New York, New York.
4Department of Pathology, New York University
Medical Center, New York, New York.
Presented at the 48th annual meeting of the Amer-
ican Society of Cytopathology, Philadelphia, Penn-
sylvania, November, 7–11, 2000.
Address for reprints: David C. Chhieng, M.D., De-
partment of Pathology, University of Alabama at
Birmingham, KB 627 619 19 Street S, Birmingham
AL 35249-6823; Fax: 205-975-7284; Email:
Received December 28, 2000; revision received
June 14, 2001; accepted June 28, 2001.
BACKGROUND. The Bethesda System recommends qualifying atypical glandular
cells with regard to their possible origin: endocervical versus endometrial. This
study was undertaken to determine the clinical significance of atypical glandular
cells of undetermined significance that favor an endometrial origin (AGUS-EM).
METHODS. A computer search identified 62 cervicovaginal smears (5.25% of all
smears classified as AGUS) with a diagnosis of AGUS-EM in the files of Shared
Cytopathology Laboratory of New York University Medical Center/Bellevue Hos-
pital Medical Center between January 1995 and December 1999. The patients
ranged in age from 29 years to 88 years (mean age, 53 years). Thirty-four patients
were postmenopausal (55%), and 5 patients were on hormonal replacement ther-
apy. Follow-up was available for 56 patients (90%); 45 patients (73%) underwent
biopsy, and 11 patients (17%) had repeat cervicovaginal smears. Six patients were
lost to follow-up.
RESULTS. Among patients who underwent biopsy, 14 patients (31%) had a clinically
significant uterine lesions, including 6 (13%) endometrial adenocarcinomas, 5
(11%) endometrial hyperplasias, and 3 (7%) squamous lesions (2 high-grade squa-
mous intraepithelial lesions and 1 squamous cell carcinoma). Ten of 11 patients
with significant endometrial pathology findings were postmenopausal. The re-
maining 31 patients had benign pathology results, which included chronic cervi-
citis, endometritis, endometrial polyps, microglandular hyperplasia, and tubal
metaplasia. Among the patients with repeat cervicovaginal smears, one patient had
atypical squamous cells of undetermined significance; the remaining patients were
within normal limits.
CONCLUSIONS. Approximately one-third of women with a diagnosis of AGUS-EM
had a significant uterine lesion on subsequent biopsy; the majority of these lesions
were endometrial in origin. Patients with a diagnosis of AGUS-EM on cervicovagi-
nal smears should be followed closely, and endometrial curettage or biopsy should
be included in their initial work-up. Cancer (Cancer Cytopathol) 2001;93:351–6.
© 2001 American Cancer Society.
KEYWORDS: cervicovaginal smears, atypical, glandular cells, endometrial cells,
1988 to denote cytologic changes in glandular cells that exceed those
typical of reactive changes but are quantitatively and/or qualitatively
not diagnostic of adenocarcinoma.1These changes may represent a
florid response to inflammation, a precursor neoplastic lesion, or a
nonrepresentative sampling of a neoplastic lesion. The Bethesda sys-
tem also recommends subclassifying the atypical glandular cells ac-
cording to their possible anatomic origin: endocervical versus endo-
metrial.2There have been a number of studies addressing the
he diagnostic category atypical glandular cells of undetermined
significance (AGUS) was introduced by the Bethesda System in
© 2001 American Cancer Society
implications and management of patients with AGUS
in general as well as patients with AGUS, favor endo-
cervical origin in particular.3–17Little is known about
the clinical significance of AGUS, favor endometrial
origin (AGUS-EM). In this retrospective study, we de-
termined the rate of a diagnosis of AGUS-EM and the
incidence of clinically significant lesions in women
with this cytologic diagnosis on subsequent follow-up.
We also suggested guidelines for the clinical manage-
ment of patients with this diagnosis.
MATERIALS AND METHODS
From January 1995 to December 1999, 211,220 cervi-
covaginal smears were screened at the Shared Cyto-
pathology Laboratory at New York University Medical
Center/Bellevue Hospital Medical Center. All cervico-
vaginal smears were collected using an endocervical
brush, an Ayre spatula, or a cervical brush. One or two
slides per patient were prepared, depending on the
preferences of the health care providers who obtained
the smears. All smears were fixed immediately with
nonaerosal spray fixative and sent to the cytopathol-
Patients with a cytologic diagnosis of AGUS-EM,
were identified during the study period. Patients with
benign appearing endometrial cells were excluded.
The diagnoses were made by staff pathologists using
the criteria listed in Table 1.
Information regarding previous gynecologic his-
tory and hormone replacement therapy was obtained.
The methods and results of patient evaluation, which
included repeat cervicovaginal smears, colposcopic
examination along with endocervical curettage and/or
cervical biopsy, and endometrial biopsy, were re-
viewed and recorded. Follow-up smears were ob-
tained 6–12 months after the initial abnormal smear;
all biopsies and curettages were performed within 1
During the 5-year study period, 1179 cervicovaginal
smears (0.56%) had a cytologic diagnosis of AGUS.
Among these smears, 62 (5.25%) had a diagnosis of
AGUS-EM. The mean age of this group of patients was
53 years, with a range of 29–88 years. Thirty-four
patients (55%) were postmenopausal. Five of 34 pa-
tients (15%) were receiving hormone replacement
therapy at the time of the AGUS diagnosis. Seven
patients presented with abnormal vaginal bleeding:
Six patients were postmenopausal, and one patient
was premenopausal. One patient had a history of en-
dometrial adenocarcinoma and was status posthyster-
ectomy. An additional patient presented with a history
of an ovarian mucinous neoplasm of low malignant
Follow-up information was available for 56 pa-
tients (90%): Forty-five patients (73%) had histologic
follow-up, and 11 patients (17%) had repeat cervico-
vaginal smears. Six patients were lost to follow-up.
The types of procedures performed for patients who
underwent histologic evaluation are shown in Table 2.
All but six patients underwent endometrial biopsy.
The patient who was status posthysterectomy for en-
dometrial carcinoma had biopsy of the vaginal cuff.
Fourteen of 45 patients (31%) with subsequent
histologic follow-up had a clinically significant uterine
lesion that was defined as a preneoplastic or neoplas-
tic, glandular or squamous lesion. Eleven patients
(24%) had a glandular lesion, including 6 endometrial
adenocarcinomas (Fig. 1) and 5 endometrial hyperpla-
sias. Three patients (7%) had squamous lesions, in-
cluding two high-grade squamous intraepithelial le-
sions (Fig. 2) and one invasive squamous cell
carcinoma. Table 3 shows the results of histologic
diagnoses in relation to the patients’ hormonal status.
Ten of 11 patients with significant endometrial pathol-
ogy and all patients with squamous lesions were post-
menopausal. In addition, four of five patients who
presented with abnormal vaginal bleeding and under-
went histologic evaluation had significant endometrial
lesions. The remaining 31 patients with biopsies had
Criteria of Atypical Glandular Cells, Favor Endometrial Origin
Arranged in loose and tight clusters
?/? Irregular nuclear contours
?/?: With or without.
Methods No. of patients (%)
EMB and ECC
Cervical biopsy and ECC
Cervical biopsy, ECC, and EMB
EMB: endometrial biopsy; ECC: endocervical curettage.
352 CANCER (CANCER CYTOPATHOLOGY) December 25, 2001 / Volume 93 / Number 6
benign pathology that included chronic cervicitis with
squamous metaplasia, endometritis (Fig. 3), endome-
trial polyps, mircoglandular hyperplasia (Fig. 4), and
tubal metaplasia. None of them demonstrated any
significant endocervical lesions. Among the patients
with repeat cervicovaginal smears, one patient had a
diagnosis of atypical squamous cells of undetermined
significance. The repeat smears of the remaining 10
patients were interpreted as within normal limits or as
benign cellular changes.
Among all patients with an AGUS diagnosis, 62%
underwent histologic evaluation. Twenty-nine percent
had a significant uterine lesion on subsequent follow-
up: Twenty-two percent were squamous lesions, 2%
were endocervical lesions, and 5% were endometrial
FIGURE 3. A large, three-dimensional group of glandular cells with mild
nuclear crowding and overlapping as well as mild nuclear hyperchromasia.
Subsequent biopsy revealed marked chronic endometritis (Papanicolaou stain;
original magnification, ?600).
FIGURE 4. A small cluster of glandular cells with scant cytoplasm and mild
nuclear hyperchromasia and pleomorphism. Subsequent biopsy revealed mi-
croglandular hyperplasia; the patient was on hormone replacement therapy
(Papanicolaou stain; original magnification, ?600).
FIGURE 1. Three-dimensional cluster of glandular cells with hyperchromatic
nuclei and scant-to-moderate, sometimes vacuolated cytoplasm. Endometrial
biopsy revealed a well-differentiated endometrial adenocarcinoma (Papanico-
laou stain; original magnification, ?600).
FIGURE 2. Haphazardly arranged glandular cells with coarse chromatin and
an increased nuclear-to-cytoplasmic ratio. It was determined that the patient
had a high-grade squamous intraepithelial lesion that involved the endocervical
gland (Papanicolaou stain; original magnification, ?600).
Results of Histologic Follow Up and Menopausal Status
Histologic diagnosis PostmenopausalPremenopausalTotal
Significant uterine lesions
SCC or HGSIL
SCC: squamous cell carcinoma; HGSIL: high-grade squamous intraepithelial lesion.
Significance of AGUS, Favor Endometrial Origin/Chhieng et al. 353
Since the introduction of the diagnostic category
AGUS by the Bethesda System in 1988, there have
been a number of studies addressing the incidence
and the clinical significance of AGUS.3–17The reported
rates of AGUS among the general patient population
ranged from 0.17% to 1.83%. Most laboratories report
an AGUS rate ? 1%.3–17The AGUS rate of 0.56% ob-
served among the general patient population in our
laboratory was in agreement with the current litera-
Few studies have investigated the incidence of
atypical endometrial cells in cervicovaginal smears.
One study was carried out by Cherkis et al.18before
the introduction of the Bethesda system. Those au-
thors identified 177 women with atypical endometrial
cells from a screening pool of approximately 300,000
cervicovaginal smears, resulting in a frequency of 1 in
1700 or 0.06%. In another study, the incidence of
atypical endometrial cells was 6.2% (5 of 81 smears)
among all cervicovaginal smears with a diagnosis of
AGUS.19In a more recent study, Soofer and Sidaway17
reported that the incidence of atypical endometrial
cells was 12% among all AGUS smears and 0.01%
among all cervicovaginal smears screened during the
study period. In the current study, atypical endome-
trial cells accounted for 5.25% of smears with a diag-
nosis of AGUS or 0.03% of all smears screened in our
laboratory during the study period, findings consistent
with those in the literature. It is interesting to note that
many studies did not subclassify the origin of the
atypical glandular cells according to their possible an-
Thirty-one percent of our patients with histologic
follow-up had a significant uterine lesion; 11 lesions
(24%) were of endometrial origin, and 3 lesions (7%)
were of squamous origin. These findings are similar to
those reported by Cherkis et al.18: 32 of 134 patients
(31.1%) with atypical endometrial cells had a signifi-
cant uterine lesion. Duska et al.19also reported that 2
of 5 patients (40%) with atypical endometrial cells had
a significant uterine lesion on subsequent follow-up.
In both studies, the significant uterine lesions were
exclusively of endometrial origin. Soofer and Sid-
away,17in a different study, identified 2 (18%) signif-
icant uterine lesions, 1 high-grade squamous intraepi-
thelial lesion (HGSIL), and 1 endometrial lesion in 11
patients with atypical endometrial cells. The incidence
of a significant endometrial lesion (either hyperplasia
or adenocarcinoma) on subsequent follow-up when
normal endometrial cells were identified in the pre-
ceding cervicovaginal smears ranged from 10.5% to
24.1%, with an average of 14.8%.20–23Therefore, the
incidence of a significant uterine lesion on subsequent
follow-up is higher for patients with atypical endome-
trial cells on their cervicovaginal smears compared
with patients with smears that show normal endome-
In a previous study, we reported that 75% of the
significant uterine lesions that were identified after a
diagnosis of AGUS were of squamous origin.3In the
current study, the majority of significant uterine le-
sions (79%) were endometrial in origin after a diagno-
sis of atypical endometrial cells. This may be ex-
plained by the fact that patients with atypical
endometrial cells tended to be older (average age, 53
years vs. 45 years in a previous study3) and were more
likely to be postmenopausal (55% vs. 40% in a previ-
ous study3) compared with patients who were diag-
nosed with AGUS. Similar observations were made by
other authors.18,19Cherkis et al.18reported that all of
the significant uterine lesions after a diagnosis of atyp-
ical endometrial cells were of endometrial origin.
Therefore, our findings and findings from other re-
ports justify the subclassification of AGUS-EM.
Twenty-seven of 45 postmenopausal women
(60%) with atypical endometrial cells had a significant
uterine lesion on subsequent follow-up, whereas 1 of
18 premenopausal women (6%) with atypical endo-
metrial cells had a significant uterine lesion on follow-
up. There was a 10-fold increase in the frequency of
significant uterine lesions in postmenopausal women
compared with premenopausal women. The differ-
ence was statistically significant (P ? 0.003; Fisher
exact test). Other authors also have observed an in-
creased incidence of endometrial adenocarcinoma as-
sociated with increasing age in patients with atypical
endometrial cells.18Patients age ? 59 years were more
likely to have an endometrial adenocarcinoma than
patients age ? 59 years, with a relative ratio of 2.3.
In the current study, among the five patients who
presented with vaginal bleeding and had histologic
follow-up, four women were postmenopausal and had
either endometrial hyperplasia or endometrial carci-
noma after a diagnosis of atypical endometrial cells in
cervicovaginal smears. The remaining patient was pre-
menopausal and had an endometrial polyp on histo-
logic evaluation. It is not surprising that symptomatic
postmenopausal women with atypical endometrial
cells were at risk of harboring a significant endome-
trial lesion. It is more important to note that more
than half (64%) of the patients with significant endo-
metrial lesions were asymptomatic. The cytologic
finding of atypical endometrial cells triggered a sub-
sequent work-up, resulting in the identification of the
occult endometrial lesions.
Three of our patients with atypical endometrial
354CANCER (CANCER CYTOPATHOLOGY) December 25, 2001 / Volume 93 / Number 6
cells had an HGSIL or invasive squamous cell carci-
noma on follow-up. Other authors also have reported
the finding of HGSIL on follow-up of a patient with
AGUS-EM.17HGSIL often is characterized by the pres-
ence of three-dimensional, hyperchromatic, crowded
groups that can mimic endometrial cells.24In a recent
study, Zhou et al.25reported that small cell carcinoma
of the cervix can form ball-like clusters of cells with
scant cytoplasm resembling endometrial cells. The
finding of dysplastic squamous cells should provide
evidence that the hyperchromatic, crowded groups
most probably are of squamous origin.24In addition,
the identification of a squamous cervical lesion does
not necessarily preclude the possibility of a concur-
rent endometrial lesion.
Various investigators have attempted to qualify
the diagnosis of AGUS to favor a reactive/benign pro-
cess or a premalignant/malignant condition.11,26–31
The majority of these studies focused on atypical glan-
dular cells of endocervical origin, particularly the cri-
teria for separating endocervical adenocarcinoma in
situ from benign lesions. There was only one study
that attempted to subclassify atypical endometrial
cells into three categories based on the nuclear size
and hyperchromasia, chromatin pattern, and pres-
ence or absence of nucleoli.18Those authors reported
that 40% of patients with markedly atypical endome-
trial cells, 20% of patients with moderately atypical
endometrial cells, and none of the patients with mildly
atypical endometrial cells had an endometrial carci-
noma on follow-up, with a risk ratio of 1.7. Both the
Bethesda System and the International Academy of
Cytology Task Force did not recommend the subclas-
sification of atypical endometrial cells because of the
lack of well-defined criteria.2,32Therefore, it is our
routine practice not to qualify further the atypical
endometrial cells according to whether they favor a
reactive or neoplastic process.
Ronnett et al.33have assessed the value of human
papillomavirus (HPV) DNA testing in the management
of patients with cervicovaginal smears classified as
AGUS.33HPV testing identified 92% of women with
histologically confirmed HGSIL, all patients with en-
docervical adenocarcinoma in situ, but none of the
patients with endometrial adenocarcinoma. There-
fore, HPV DNA testing did not provide useful informa-
tion for evaluating the status of the endometrium.
Currently, there is no clear consensus regarding
the management of patients with atypical endometrial
cells. Based on the findings of our study, we would like
to suggest the following guidelines: A complete endo-
metrial evaluation should be included as part of the
initial examination, particularly for postmenopausal
women. Colposcopic examination with endocervical
curettage should be considered part of the initial ex-
amination, because a small but significant number of
patients can have a high-grade squamous lesion.
Asymptomatic premenopausal patients may be fol-
lowed conservatively. In patients with an otherwise
negative initial histologic evaluation and persistent
glandular atypia on repeat cervicovaginal smears, the
clinician should consider more aggressive evalua-
tion.3,18,19It cannot be overemphasized that commu-
nication between clinicians and cytopathologists must
be established to convey the degree of concern about
possible neoplasia for a cytologic interpretation of
In summary, the incidence of atypical endome-
trial cells was low: 5.25% of all AGUS smears diag-
nosed in our laboratory. About one-third of these pa-
tients had a significant uterine lesion on follow-up,
with the majority of endometrial origin. Close fol-
low-up is indicated in patients with a cytologic diag-
nosis of AGUS-EM.
1.Anonymous. The 1988 Bethesda System for reporting cervi-
cal/vaginal cytological diagnoses. National Cancer Institute
Workshop. JAMA 1989;262(7):931–4.
2. Kurman R, Solomon D. The Bethesda System for reporting
cervical/vaginal cytologic diagnoses: definitions, criteria
and explanatory notes for terminology and specimen ade-
quacy. New York: Springer-Verlag, Inc., 1994.
3. Chhieng DC, Elgert PA, Cangiarella JF, Cohen JM. Clinical
significance of atypical glandular cells of undetermined sig-
nificance. A follow-up study from an academic medical
center. Acta Cytol 2000;44(4):557–66.
4. Eddy GL, Ural SH, Strumpf KB, Wojtowycz MA, Piraino PS,
Mazur MT. Incidence of atypical glandular cells of uncertain
significance in cervical cytology following introduction of
the Bethesda System. Gynecol Oncol 1997;67(1):51–5.
5. 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(5):
6.Gornall RJ, Singh N, Noble W, Boyd IE, Hitchcock A. Glan-
dular abnormalities on cervical smear: a study to compare
the accuracy of cytological diagnosis with underlying pa-
thology. Eur J Gynaecol Oncol 2000;21(1):49–52.
7.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) de-
tected on cervical cytology screening. Gynecol Oncol 1996;
8.Kim TJ, Kim HS, Park CT, Park IS, Hong SR, Park JS, et al.
Clinical evaluation of follow-up methods and results of
(AGUS) detected on cervicovaginal Pap smears. Gynecol On-
9.Manetta A, Keefe K, Lin F, Ahdoot D, Kaleb V. Atypical
glandular cells of undetermined significance in cervical cy-
tologic findings. Am J Obstet Gynecol 1999;180(4):883–8.
Significance of AGUS, Favor Endometrial Origin/Chhieng et al.355
10. Schindler S, Pooley RJ Jr., De Frias DV, Yu GH, Bedrossian
CW. Follow-up of atypical glandular cells in cervical-endo-
cervical smears. Ann Diagn Pathol 1998;2(5):312–7.
11. Lee KR, Manna EA, St John T. Atypical endocervical glandu-
lar cells: accuracy of cytologic diagnosis. Diagn Cytopathol
12. Nasu I, Meurer W, Fu YS. Endocervical glandular atypia and
adenocarcinoma: a correlation of cytology and histology. Int
J Gynecol Pathol 1993;12(3):208–18.
13. Goff BA, Atanasoff P, Brown E, Muntz HG, Bell DA, Rice LW.
Endocervical glandular atypia in Papanicolaou smears. Ob-
stet Gynecol 1992;79(1):101–4.
14. Bose S, Kannan V, Kline TS. Abnormal endocervical cells.
Really abnormal? Really endocervical? Am J Clin Pathol
15. Cenci M, Mancini R, Nofroni I, Vecchione A. Endocervical
atypical glandular cells of undetermined significance. II.
Morphometric and cytologic analysis of nuclear features
useful in characterizing differently correlated subgroups.
Acta Cytol 2000;44(3):327–31.
16. Cenci M, Mancini R, Nofroni I, Vecchione A. Endocervical
atypical glandular cells of undetermined significance. I.
Morphometric and cytologic characterization of cases that
“cannot rule out adenocarcinoma in situ.” Acta Cytol 2000;
17. Soofer SB, Sidawy MK. Atypical glandular cells of undeter-
mined significance: clinically significant lesions and means
of patient follow-up. Cancer 2000;90(4):207–14.
18. Cherkis RC, Patten SF Jr., Dickinson JC, Dekanich AS. Sig-
nificance of atypical endometrial cells detected by cervical
cytology. Obstet Gynecol 1987;69(5):786–9.
19. Duska LR, Flynn CF, Chen A, Whall-Strojwas D, Goodman A.
Clinical evaluation of atypical glandular cells of undeter-
mined significance on cervical cytology. Obstet Gynecol
20. Gondos B, King EB. Significance of endometrial cells in
cervicovaginal smears. Ann Clin Lab Sci 1977;7(6):486–90.
21. Cherkis RC, Patten SF Jr., Andrews TJ, Dickinson JC, Patten
FW. Significance of normal endometrial cells detected by
cervical cytology. Obstet Gynecol 1988;71(2):242–4.
22. Kerpsack JT, Finan MA, Kline RC. Correlation between en-
dometrial cells on Papanicolaou smear and endometrial
carcinoma. Southern Med J 1998;91(8):749–52.
23. Gomez-Fernandez CR, Ganjei-Azar P, Capote-Dishaw J,
Averette HE, Nadji M. Reporting normal endometrial cells in
Pap smears: an outcome appraisal. Gynecol Oncol 1999;
24. DeMay R. The art and science of cytopathology, vol I. Chi-
cago: ASCP Press, 1995.
25. Zhou C, Hayes MM, Clement PB, Thomson TA. Small cell
carcinoma of the uterine cervix: cytologic findings in 13
cases. Cancer 1998;84(5):281–8.
26. DiTomasso JP, Ramzy I, Mody DR. Glandular lesions of the
cervix. Validity of cytologic criteria used to differentiate re-
active changes, glandular intraepithelial lesions and adeno-
carcinoma. Acta Cytol 1996;40(6):112–-35.
27. 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 be-
nign lesions. Am J Clin Pathol 1995;104(5):574–82.
28. 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(2):
29. Burja IT, Thompson SK, Sawyer WL Jr., Shurbaji MS. Atyp-
ical glandular cells of undetermined significance on cervical
smears. A study with cytohistologic correlation. Acta Cytol
30. Siziopikou KP, Wang HH, Abu-Jawdeh G. Cytologic features
of neoplastic lesions in endocervical glands. Diagn Cyto-
31. Roberts JM, Thurloe JK, Bowditch RC, Laverty CR. Subdivid-
ing atypical glandular cells of undetermined significance
according to the Australian modified Bethesda, system:
analysis of outcomes. Cancer 2000;90(2):87–95.
32. 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 cy-
tology towards the 21st century: an international expert
conference and tutorial. Acta Cytol 1998;42(1):16–24.
33. Ronnett BM, Manos MM, Ransley JE, Fetterman BJ, Kinney
WK, Hurley LB, et al. Atypical glandular cells of undeter-
mined significance (AGUS): cytopathologic features, his-
topathologic results, and human papillomavirus DNA detec-
tion. Hum Pathol 1999;30(7):816–25.
356 CANCER (CANCER CYTOPATHOLOGY) December 25, 2001 / Volume 93 / Number 6