1 Cervical and Vaginal Cytology
1 Cervical and Vaginal Cytology
1 Cervical and Vaginal Cytology
The 20th century witnessed a remarkable decline in the cancer was the most common cause of cancer deaths in
mortality from cervical cancer in many developed coun- women in the United States.1 Today, it is not even one of
tries. This achievement is directly attributable to the the top ten.2
implementation of the Papanicolaou (Pap) test. In the The incidence of cervical cancer in the United
1930s, before Pap screening was introduced, cervical States is approximately 11,000 cases, with 3670 deaths.2
1
2 Cervical and Vaginal Cytology
Worldwide, however, the cervical cancer incidence (over of negative smears she had had, by 48% with just one
500,000 cases annually) and mortality rates (288,000 negative smear, 69% with two to four negative smears,
deaths per year) are second only to those for breast and 100% with five or more smears.17
cancer.3 Screening programs, unfortunately, are rudimen- Screening guidelines differ around the world. Even in
tary or nonexistent in many parts of the world. Fewer than the United States, the recommendations of different orga-
5% of women in developing countries have ever had a Pap nizations vary in some of their details.18-20 The American
test.4 In contrast, 89% of women in the United States report Cancer Society (ACS) recommends the following:
having had a Pap test in the preceding 3 years. • Cervical cancer screening should begin approxi-
mately 3 years after a woman begins having vaginal
intercourse, but no later than 21 years of age.
The History of the PAP Test • Until age 30, cervical screening should be carried
out every year with conventional Pap tests or every
The Pap test is considered by many to be the most cost- 2 years using liquid-based Pap tests.
effective cancer reduction program ever devised.1 Credit • At or after age 30, a woman who has had three nor-
for its conception and development goes to George N. mal test results in a row may be screened every
Papanicolaou, an anatomist and Greek immigrant to the 2 to 3 years with a Pap test (smear or liquid-based)
United States. In 1928 he reported that malignant cells or every 3 years with a Pap plus human papilloma-
from the cervix can be identified in vaginal smears.5 virus (HPV) test.
Later, in collaboration with the gynecologist Herbert • A woman 70 years of age and older who has had
Traut, who provided him with a large number of clinical three or more normal Pap test results and no abnor-
samples, Papanicolaou published detailed descriptions mal results in the previous 10 years may choose to
of preinvasive cervical lesions.6,7 Pathologists and phy- stop cervical cancer screening.
sicians initially greeted this technique with skepticism, • A woman who has had a total hysterectomy may
but by the late 1940s Papanicolaou’s observations had choose to stop cervical cancer screening. (Exceptions
been confirmed by others. The Canadian gynecologist are women with a history of CIN 2,3, cervical cancer,
J. Ernest Ayre suggested taking samples directly from the or in utero diethylstilbestrol [DES] exposure.)
cervix with a wooden spatula rather than from the vagina Women with a history of cervical cancer, in utero DES
with a pipette as originally described by Papanicolaou.8 exposure, and who are immunocompromised (organ
Eventually, cytologic smears were embraced as an ideal transplantation, chemotherapy, chronic corticosteroid
screening test for preinvasive lesions, which, if treated, treatment, or positive for human immunodeficiency
would be prevented from developing into invasive virus [HIV]) may benefit from more frequent screening.19
cancer. Adherence to these guidelines is critical for cervical can-
The first cervical cancer screening clinics were estab- cer prevention. In the United States, more than 50% of
lished in the 1940s.9 The Pap test was never evaluated women who develop cervical cancer have not had a Pap
in a controlled, prospective study, but several pieces of test in the 3 years before their cancer diagnosis.21
evidence link it to the prevention of cervical cancer. The recent development of two prophylactic HPV
First, the mortality rate from cervical cancer fell dramat- vaccines provides a new opportunity for cervical can-
ically after screening was introduced, by 72% in British cer prevention.3 Both vaccines consist of empty protein
Columbia10 and 70% in Kentucky.11 Second, there was a shells called virus-like particles that are made up of the
direct correlation between the intensity of screening and major HPV capsid protein L1. They contain no DNA and
the decrease in mortality. Among Scandinavian countries, are not infectious. One of the vaccines, Gardasil (Merck
the death rate fell by 80% in Iceland, where screening was & Co., Inc.), is a quadrivalent vaccine against HPV
greatest; in Norway, where screening was lowest, the types 6, 11, 16, and 18. The other is the bivalent vaccine
death rate fell by only 10%.12 A similar correlation was Cervarix (GlaxoSmithKline) that protects against HPV 16
observed in high and low screening regions of Scotland13 and 18. They have shown extraordinary efficacy in pre-
and Canada.14 In the United States, the decrease in venting type-specific histologic CIN 2,3 lesions, with no
deaths from cervical cancer was proportional to the difference in serious adverse effects compared to pla-
screening rates in various states.15 Finally, women who cebo.22 The vaccines are administered in three doses to
do not develop invasive cancer are more likely to have females ages 9 to 26 years before the initiation of sexual
had a Pap test than women with cancer. In a Canadian activity. Continued Pap screening will remain important
study, the relative risk for women who had not had a for many decades, however, because these vaccines do
Pap test for 5 years was 2.7,16 and screening history was not protect against 30% of cervical cancers (i.e., those
a highly significant risk factor independent of other not related to HPV 16 or 18); the duration of protection
factors such as age, income, education, sexual history, is unknown; they are not effective in treating prevalent
and smoking. In Denmark, a woman’s risk of developing HPV infections; and the cost of the vaccines might limit
cervical cancer decreased in proportion to the number their use in some populations.3
Sampling and Preparation Methods 3
As seen in the aforementioned ACS recommenda- plastic spatula is recommended because wooden fibers
tions, the combination of a Pap test plus HPV test is trap diagnostic material. The spatula is rotated at least
included as an option for screening women 30 years of 360 degrees. The sample can be smeared on one half of
age or older. The rationale is to combine the superior sen- a slide and spray fixed (the other half should be covered
sitivity of HPV testing with the superior specificity of the to avoid coating it with fixative before the endocervi-
Pap test. This recommendation is controversial because cal sample is applied). Alternatively, one may set aside
it increases screening costs. Moreover, questions remain the spatula sample momentarily while the endocervical
regarding the ideal management of women with discrep- brush sample is obtained.
ant results (e.g., HPV test positive and Pap negative). The After the brush is inserted in the endocervical canal,
search for the best screening algorithm will undoubtedly some bristles should still be visible. If inserted too
continue, particularly as molecular diagnostic methods far, there may be inadvertent sampling of the lower
become more readily available. uterine segment (LUS), which causes diagnostic diffi-
culties because its epithelium resembles a high-grade
squamous intraepithelial lesion (HSIL) and adenocarci-
Sampling and Preparation noma in situ (AIS). The brush should be rotated gently
Methods only one-quarter turn. A larger rotation is unnecessary
because the circumferential bristles are in contact with
To obtain an ideal Pap specimen, the following guidelines the entire surface the moment the brush is inserted.
have been established by the Clinical and Laboratory The spatula sample, if not already applied and fixed,
Standards Institute.23 should be applied to the slide, then the brush sample
rolled over the slide, followed by immediate fixation. The
Patient instructions: two samples can be placed in quick succession on two
separate halves of the slide, or the endocervical sam-
• Schedule the examination 2 weeks after the first
ple can be rolled directly over the spatula sample, both
day of the last menstrual period. (It is preferable to
covering the entire slide. Immediate fixation (within
avoid examination during menses because blood
seconds) is critical to prevent air-drying artifact, which
may obscure significant findings.)
distorts the cells and hinders interpretation.
• Do not use vaginal medication, vaginal con-
The broomlike brush (“broom”) has a flat array of
traceptives, or douches for 48 hours before the
plastic strips contoured to conform to the cervix, with
appointment.
longer strips in the middle. This design allows simultane-
• Intercourse is not recommended the night before
ous sampling of the endocervix and ectocervix. The long
the appointment.
middle strips are inserted into the os until the shorter
outer strips bend against the ectocervix. The broom is
Specimen collection: rotated three to five times. To transfer the material, each
• Specimens should be obtained after a nonlubri- side of the broom is stroked once across the slide in a
cated speculum (moistened only with warm water painting motion.
if needed) is inserted. The cotton swab moistened with saline is no longer
• Excess mucus or other discharge should be recommended because its fibers trap cells, reducing the
removed gently with ring forceps holding a folded efficiency of cell transfer onto slides.
gauze pad. There are two options for smear fixation. Coating
• The sample should be obtained before the application fixatives contain alcohol and polyethylene glycol and
of acetic acid or Lugol iodine. are applied by pump sprays, by droppers from dropper
• An optimal sample includes cells from the ectocervix bottles, or by pouring from an individual envelope
and endocervix. included as part of a slide preparation kit. Alternatively,
the smear can be immersed directly into a container
filled with 95% ethanol.
Samples for liquid-based cytology (LBC) are obtained
Recent studies have challenged the prohibition as described except that, instead of smearing the cells on
against a lubricated speculum and suggest that water- a slide, the collection device is rinsed in a vial containing
based lubricants may be acceptable.24 a liquid fixative. In the United States, the LBC Pap test is
more common than the smear.
Conventional Smears
Conventional smears are often obtained using the com-
Liquid-Based Cytology
bination of a spatula and brush. The spatula is used first. An important landmark in the history of the Pap
Although a wooden or plastic spatula is acceptable, the test occurred in 1996 when the U.S. Food and Drug
4 Cervical and Vaginal Cytology
Administration (FDA) approved the ThinPrep™ (Hologic, increases in HSIL detection after the implementation of
Marlborough, Mass.) as an alternative to the conven- the ThinPrep.30–35
tional cervicovaginal smear. This was followed 3 years The ThinPrep is equivalent to the conventional
later by approval of the AutoCyte Prep™ (now known smear in the detection of endocervical AIS.36 Data also
as SurePath™; BD TriPath, Burlington, NC). The newest show comparable results between ThinPrep slides and
LBC is the MonoPrep™ (MonoGen, Inc., Lincolnshire, conventional smears for the detection of endometrial
Ill.), which was approved in 2006. LBCs were an impor- pathology.37
tant step in the development of automated Pap screening The ThinPrep collection vial has been approved by
devices—an improved preparation was needed to min- the FDA for direct testing for HPV, which is particularly
imize cell overlap so that automated screeners would useful for managing women whose Pap tests show atypi-
perform better in identifying abnormal cells. But LBC cal squamous cells (ASC).38,39
performed so well in clinical trials against conventional
smears that it found a market independent of auto-
SurePath Pap Test
mated screening. Although there are exceptions,25 the
great majority of peer-reviewed studies, some of them TriPath Imaging (acquired by Becton Dickinson in 2006)
detailed in this chapter, show an increased detection developed the SurePath Pap test (formerly AutoCyte Prep
of low-grade squamous intraepithelial lesions (LSILs) and CytoRich) for samples collected in an ethanol-based
or HSILs with LBC.26 The debate over increased disease transport medium. The process is shown in Figure 1.1B.
detection with LBC continues, however, and the stud- In contrast to the ThinPrep and MonoPrep methods, the
ies comparing LBC to smears have come under criticism practitioner snips off the tip of the collection device and
for allegedly sacrificing methodologic purity in their includes it in the sample vial. The equipment to pre-
design.26 Nevertheless, LBC offers several clear advan- pare slides includes a Hettich centrifuge and a PrepStain
tages over conventional smears: the opportunity to pre- robotic sample processer with computer and monitor.
pare duplicate slides and even cell block preparations The PrepMate™ is an optional accessory that automates
from the residual sample;27,28 the option of “out-of-vial” mixing the sample and dispensing it onto the density
aliquoting for HPV, chlamydia, and gonorrhea testing; reagent. Red blood cells and some leukocytes are elimi-
an improved substrate for automated screening devices; nated by density centrifugation. In addition to preparing
and a thinner cell preparation that most pathologists and an evenly distributed deposit of cells in a circle 13 mm in
cytotechnologists find less tiring to review than smears. diameter, the method incorporates a final staining step
that discretely stains each individual slide.
A multicenter, split-sample clinical trial showed a
ThinPrep Pap Test
7.2% increase in the detection of LSILs and more serious
The practitioner obtains the ThinPrep Pap sample with lesions and a significant decrease in the percentage of
either a broom-type device or a plastic spatula/endocer- unsatisfactory specimens.40
vical brush combination. The sampling device is swirled
or rinsed in a methanol-based preservative solution
MonoPrep Pap Test
(PreservCyt) for transport to the cytology laboratory and
then discarded. Red blood cells are lysed by the trans- The practitioner obtains the MonoPrep sample with
port medium. The vials are placed one at a time on the standard collection devices that are swirled or rinsed
ThinPrep 2000 instrument. The entire procedure (Fig. in a preservative-filled collection vial, after which the
1.1A) takes about 70 seconds per slide and results in a sampling device is discarded. As with the ThinPrep, red
thin deposit of cells in a circle 20 mm in diameter (con- blood cells are lysed by the transport medium. The vials
trast with cytospin: diameter = 6 mm). A batch-process- are delivered to the laboratory where slides are prepared
ing version (the ThinPrep 3000) is also available. It uses using the MonoPrep Processor, a fully automated, batch-
the same consumables (filters and solutions) but allows processing instrument capable of processing 40 samples
automated processing of 80 samples at one time. In most per hour, with a throughput capacity of 324 samples per
cases, only a fraction of the sample is used to prepare the 8-hour run. The process is shown in Figure 1.1C. In a
slide used for diagnosis. If needed, the residual sample is split-sample clinical trial similar in design to the ThinPrep
available for additional ThinPrep slide preparation, cell and SurePath trials, slides prepared by the MonoPrep
block preparation, or molecular diagnostic testing (e.g., method showed a 26% increase in the detection of LSILs
high risk HPV, chlamydia, gonorrhea). and more serious lesions, with no significant difference
A multicenter, split-sample study found that the in relative specificity.41 MonoPrep provided a signifi-
ThinPrep detected 18% more cases of LSILs and more cant reduction in unsatisfactory slides, and there was no
serious lesions as compared to conventional smears, difference in the presentation of endocervical or trans-
with no significant difference in the detection of organ- formation zone component or the detection of benign
isms.29 A number of studies have shown significant conditions.
Sampling and Preparation Methods 5
1 2 3 4 5
Vortexing Disaggregation Transfer Sedimentation Cell deposition
to sedimentation 2 and staining
tube
B
1 2 3 4
High-speed Turbidity Aspiration Cell deposition
mixing check
C
Figure 1.1 Liquid-based slide preparation methods. A, ThinPrep method: 1. The sample vial sits on a stage and a hollow plastic
cylinder with a 20-mm diameter polycarbonate filter bonded to its lower surface is inserted into the vial. A rotor spins the cylinder for
a few seconds, homogeneously dispersing the cells. 2. A vacuum is applied to the cylinder, trapping cells on the filter. The instrument
monitors cell density. 3. With continued application of vacuum, the cylinder (with cells attached to the filter) is inverted 180 degrees,
and the filter is pressed against a glass slide. The slide is immediately dropped into an alcohol bath. B, SurePath method: 1. The
sample is quickly vortexed. 2. A proprietary device, the Cyringe™, disaggregates large clusters by syringing the sample through a
small orifice. 3. The sample is poured into a centrifuge tube filled with a density gradient reagent. 4. Sedimentation is performed in
a centrifuge. A pellet is obtained and resuspended, and the sedimentation is repeated. 5. The tubes are transferred to the PrepStain
instrument, where a robotic arm transfers the fluid into a cylinder. Cells settle by gravity onto a cationic polyelectrolyte-coated slide.
The same robotic arm also dispenses sequential stains to individual cylinders. C, MonoPrep method: 1. An integrated stirrer mixes
the specimen briefly to disperse mucus and aggregates. 2. The specimen is aspirated into the hollow stirrer and dual-flow technol-
ogy captures a representative sample on a frit-backed filter. 3. The filter is pressed against the slide to transfer the cells onto a 20-mm
diameter circular area. 4. After cell transfer, the instrument applies a premeasured amount of alcohol fixative directly onto the slide.
6 Cervical and Vaginal Cytology
B
Figure 1.2 Automated cytology screening devices. A, FocalPoint Slide Profiler. The FocalPoint consists of an imaging system and
accompanying computer workstation with monitor and keyboard. After imaging is completed, the instrument prints a score for each
slide. Depending on the score, the slide is either reported as negative and archived without further review, or it is triaged for man-
ual review. B, ThinPrep Imaging System. The ThinPrep imager consists of two components, a table-top imager and an electronically
linked customized review microscope. Slides are imaged on the imager and brought to the microscope for location-guided review.
for ThinPrep slides showed a significantly higher detec- interpreted as showing ASC-US; the second reviewer
tion rate of histologic HSIL (CIN 2,3) with the TIS.57 agreed with only 43% of cases. The greatest disagree-
Because 22 FOV represent approximately 25% of the ment with biopsies involved those originally interpreted
ThinPrep cell spot,58 implementation of the TIS comes as LSIL; the second reviewer concurred in only 43% of
with a significant productivity enhancement, and in cases.79
some laboratories the productivity of cytotechnologists A graphic demonstration of the relative reproduc-
has as much as doubled.56,59,60 ibility of various cytologic findings is available on the
Implementing the TIS requires adopting the propri- Bethesda System Web Atlas, which contains the results
etary ThinPrep Pap stain, to which some adjustment is of the Bethesda Interobserver Reproducibility Project.
necessary because it yields darker nuclear staining of A large number of images were reviewed by hundreds
metaplastic and endocervical cell clusters than most of observers, who were asked to place the images into
traditional Pap stains. Like FPSP, TIS does not eliminate one of the Bethesda System categories. The results are
false-negatives, which are still encountered, albeit less displayed for each image as a histogram.80
frequently than in the absence of imaging.56 A number
of postapproval studies have shown significant increases
in the detection of LSIL and HSIL after implementation Diagnostic Terminology
of the TIS.61–63 and Reporting Systems
Papanicolaou devised a numerical system for reporting
Accuracy and cervical smears, which was originally intended to convey
Reproducibility his degree of suspicion that the patient had cancer: class I,
absence of atypical or abnormal cells; class II, atypi-
The sensitivity of cytology for detecting preinvasive cal but no evidence of malignancy; class III, suggestive
squamous and glandular lesions is difficult to establish, of but not conclusive for malignancy; class IV, strongly
but it is clearly far from perfect. Most studies of preinva- suggestive of malignancy; and class V, conclusive for
sive lesions suffer from verification bias (i.e., cases are malignancy. Over time, however, the Papanicolaou class
referred for biopsy on the basis of an abnormal smear, system underwent many modifications and was not
and women with negative Pap tests are not biopsied). used in a uniform fashion.81 It persisted in many labora-
The few relatively unbiased studies show that the mean tories well into the 1980s, however. In other laboratories it
sensitivity of the Pap test is 47% (range 30% to 80%), and was replaced (or supplemented) by descriptive terms
the mean specificity is 95% (range 86% to 100%).64 borrowed from histologic classifications of squamous
The sensitivity of cytology is less than ideal for inva- lesions. Squamous cancer precursors were originally
sive cancers as well, and estimates range widely (16% to divided into carcinoma in situ, which was a high-risk
82%). Many women with cervical cancer have a history lesion of immature, undifferentiated atypical cells, and
of one or more negative smears.65–76 The relative contri- dysplasia (subdivided into mild, moderate, and severe),
butions of sampling and laboratory error vary from one considered to be a low-risk lesion composed of more
study to another and likely depend on how carefully mature squamous cells. In the 1960s, Richart challenged
retrospective rescreening is performed. the duality of dysplasia/carcinoma in situ and proposed
False-positive diagnoses of cervical cancer occur in a new term, cervical intraepithelial neoplasia (CIN).
10% to 15% of cases.77,78 The chief culprits are the atro- CIN was graded from 1 to 3, but Richart believed that
phic smear with benign squamous atypia in a granular, CIN 1 (mild dysplasia) had a strong propensity to prog-
pseudonecrotic background; reparative changes; and ress to CIN 3 and cancer. The high rate of progression
keratinizing HSILs. found in his study most likely related to stringent entry
The interobserver reproducibility of cytologic inter- criteria; for inclusion, CIN 1 had to be confirmed on three
pretations is less than perfect. In a large study of women, consecutive Paps.82 Richart’s data showed a higher pro-
most of whom had mild cytologic abnormalities, the gression rate for mild dysplasia than most other natural
unweighted κ statistic for four categories of diagnosis— history studies.83 The CIN concept was highly influential,
negative, atypical, LSIL, and HSIL—was 0.46, indicating however, and for many years squamous precursors were
moderate reproducibility.79 (Roughly, a κ of 0 or less rep- treated as much on the basis of their size and location as
resents poor agreement, 0 to 0.2 slight agreement, 0.2 to on their grade. This situation remained for two decades.
0.4 fair agreement, 0.4 to 0.6 moderate agreement, 0.6 to In 1989 the Bethesda System was introduced to stan-
0.8 very good agreement, and 0.8 to 1.0 almost perfect dardize the reporting of cervical cytology results and
agreement.) In the same study, the reproducibility of his- incorporate new insights gained from the discovery of
tologic interpretations of cervical biopsies, also for four HPV.84 The name for a squamous cancer precursor was
categories of diagnosis, was identical (0.46). The greatest changed to squamous intraepithelial lesion (SIL), sub-
disagreement with Pap tests involved those originally divided into only two grades (low and high) based on the
The Bethesda System 9
evolving understanding of the biology of HPV. In this sys- Table 1.1 The 2001 Bethesda System Categories
tem, LSIL encompasses CIN 1, and HSIL encompasses for Specimen Adequacy
CIN 2 and 3. This was a shift away from the CIN concept,
Satisfactory for Evaluation
one based on a reevaluation of the existing evidence,
A satisfactory squamous component must be present (see
which demonstrated that most LSILs are, in fact, tran- text).
sient HPV infections that carry little risk for oncogenesis, Note the presence or absence of endocervical or
whereas most HSILs are associated with viral persistence transformation zone component.
and a significant potential for progression to invasive Obscuring elements (inflammation, blood, drying artifact,
cancer. other) may be mentioned if 50% to 75% of epithelial cells
The first Bethesda System workshop in 1988 was fol- are obscured.
lowed by two others in 1991 and 2001, which made modi- Unsatisfactory for Evaluation
fications to the original framework and terminology. The Specimen rejected or not processed because (specify
reason). Reasons may include:
2001 workshop broadened participation by using a dedi-
• lack of patient identification.
cated Web site on the Internet, and an electronic bulle-
• unacceptable specimen (e.g., slide broken beyond
tin board received more than 1000 comments regarding repair).
draft recommendations. The 2001 Bethesda System, like or:
its predecessors, recommends a specific format for the Specimen processed and examined, but unsatisfactory for
cytology report, starting with an explicit statement on evaluation of an epithelial abnormality because (specify
the adequacy of the specimen, followed by a general cat- reason). Reasons may include:
egorization and an interpretation or result.85,86 • insufficient squamous component (see text).
• obscuring elements cover more than 75% of epithelial
cells
A spot-counting method is used to evaluate LBCs metaplastic cells, either isolated or in groups, are present.
with borderline squamous cellularity. A minimum of The data on the endocervical component as a measure
10 fields are counted along a diameter that includes the of adequacy are contradictory.94 The importance of
center of the slide (Fig. 1.3A). If the cell circle has blank endocervical cells was first suggested by cross-sectional
spots, these should be represented in the fields counted studies, which showed that smears are more likely to
(Fig. 1.3B). The average number of squamous cells is contain SIL when endocervical cells are present.95–97 Data
then compared against tables that take into account the from retrospective case-control studies, however, do not
objective, the eyepiece field number, and the diameter support this; investigators have found no association
of the circle that contains cellular material.86 For exam- between false-negative Pap samples and the absence of
ple, with an FN20 eyepiece, and a ×40 objective, the sam- endocervical cells.98,99 Retrospective cohort studies have
ple is adequate if the average number of cells counted is shown that women whose initial smears lack endocervi-
greater than 3.1 for a ThinPrep slide. cal cells do not develop more lesions on follow-up than
Additional slides can usually be generated from the women whose smears do have an endocervical compo-
residual vial of an LBC sample. In some laboratories, an nent,100–102 implying that an endocervical component
additional slide is prepared when the initial slide has is not essential. Currently, a smear without endocervi-
insufficient cellularity. The addition of a washing step cal cells is not considered unsatisfactory, although the
with 10% glacial acetic acid increases the percentage of absence of an endocervical or transformation zone
satisfactory ThinPrep Pap samples, uncovering occa- component is mentioned as a “quality indicator.” This
sional cases of SIL and invasive cancer.92,93 is not to imply that a repeat Pap is necessary. Physicians
The cellularity of the squamous cell component are expected to use their judgment and to consider
is estimated; laboratories are not expected to count repeating the Pap if the patient is at high risk for cervical
individual cells. Squamous cellularity is sometimes cancer.
particularly difficult to estimate, for example, when
there is marked cell clustering or cytolysis. In certain
clinical settings, particularly in women with atrophy,
General Categorization
a lower number may be adequate. In these situations, The general categorization is an optional component of
cytologists are expected to use their judgment when the 2001 Bethesda System.
evaluating adequacy.86
In the 2001 Bethesda System, the presence or absence Three categories:
of an endocervical or transformation zone component is
noted on the report. An endocervical component is con- • negative for intraepithelial lesion or malignancy
sidered present if 10 or more endocervical or squamous • epithelial cell abnormality
• other
Table 1.2 The 2001 Bethesda System In the United States, a pathologist is required to review
cases that show reactive or reparative changes and any
Specimen Adequacy (see Table 1.1)
abnormality at the level of ASC-US or higher. This rep-
General Categorization (Optional)
resents about 10% to 20% of the total Pap volume in
Negative for intraepithelial lesion or malignancy (NILM)
most laboratories.
Epithelial cell abnormality
Other
Interpretation/results Squamous Cells
NILM The ectocervix is lined by a stratified squamous epi-
Organisms
Trichomonas vaginalis thelium that matures under the influence of estrogen.
Fungal organisms morphologically consistent with The most mature squamous cells are called superficial
Candida species cells. They have a small, pyknotic nucleus that is 5 to 6
Shift in flora suggestive of bacterial vaginosis μm in diameter. Intermediate cells have a larger nucleus
Bacteria morphologically consistent with Actinomyces measuring 8 μm in diameter, which is not pyknotic but
species
Cellular changes consistent with herpes instead has a finely granular texture. Intermediate cells
simplex virus are occasionally binucleated and even multinucleated.
Other non-neoplastic findings (optional to report; list not Both superficial and intermediate cells are large poly
comprehensive) gonal cells with transparent pink or green cytoplasm
Reactive cellular changes associated with: inflammation (Fig. 1.4). Superficial and intermediate cells are the pre-
(includes typical repair); radiation; intrauterine dominant cells in cytologic samples from women of
contraceptive device (IUD)
Glandular cells status post hysterectomy reproductive age.
Atrophy Immature squamous cells are called parabasal cells
Epithelial cell abnormalities and basal cells. Because a Pap test does not usually
Squamous cell scrape off the entire thickness of the epithelium but only
Atypical squamous cells (ASC) the upper few layers, immature cells near the base of a
- of undetermined significance (ASC-US) mature epithelium are not sampled. An immature epi-
- cannot exclude HSIL (ASC-H)
Low-grade squamous intraepithelial lesion (LSIL) thelium, however, is composed throughout its thickness
High-grade squamous intraepithelial lesion (HSIL) by parabasal-type cells or basal-type cells. Immature
Squamous cell carcinoma (SQC) epithelium is common at the transformation zone,
Glandular cell where it is called squamous metaplasia, and whenever
Atypical glandular cells (AGC); specify endocervical, there is squamous epithelial atrophy as a result of a low
endometrial, or not otherwise specified
AGC, favor neoplastic (specify endocervical or not estrogen state. Thus, parabasal and basal cells are typi-
otherwise specified) cally obtained from squamous metaplasia or atrophic
Endocervical adenocarcinoma in situ (AIS) epithelium.
Adenocarcinoma Squamous atrophy is encountered in a variety of
Other clinical settings associated with a low estrogen state.
Endometrial cells in a woman older than 40 years
of age
Automated Review and Ancillary Testing (Include as
Appropriate)
Educational Notes and Suggestions (Optional)
A B
Figure 1.6 Parabasal cells (postmenopausal smear). A, Atrophic epithelium is composed almost exclusively of parabasal cells,
often arranged in broad, flowing sheets. B, Transitional cell metaplasia. In this uncommon condition, the atrophic epithelium resem-
bles transitional cell epithelium by virtue of its longitudinal nuclear grooves. Nuclear membrane irregularities raise the possibility of
a high-grade squamous intraepithelial lesion (HSIL), but the chromatin is pale and finely textured.
The Normal PAP 13
A B
Figure 1.7 Parabasal cells (postmenopausal smear). A, Degenerated parabasal cells in atrophic smears have hypereosinophilic
cytoplasm and a pyknotic nucleus. Note the granular background, which is commonly seen in normal atrophic smears. B, Dark blue
blobs are seen in some atrophic smears. These featureless structures should not be interpreted as a significant abnormality.
Figure 1.8 Squamous metaplasia. Interlocking parabasal-type cells, as seen here, represent squamous metaplasia of the
endocervix.
histologic term describing mature squamous epithelium Parakeratosis, a benign reactive change also caused
overlying endocervical glands, is not recognized as such by chronic irritation, is characterized by small, heavily
on cytologic preparations. keratinized squamous cells with dense orangeophilic
Other normal changes of squamous cells are hyper- cytoplasm and small, pyknotic nuclei (Fig. 1.9B). When
keratosis and parakeratosis. Hyperkeratosis is a benign such densely keratinized cells show nuclear atypia in the
response of stratified squamous epithelium as a result form of enlargement and membrane irregularity with
of chronic mucosal irritation, as in uterine prolapse. hyperchromasia, they are called “dyskeratocytes” or
Anucleate, mature, polygonal squamous cells appear “atypical parakeratosis” and should be categorized as an
as isolated cells or plaques of tightly adherent cells (Fig. epithelial cell abnormality.
1.9A). Such cells are benign and should not be consid-
ered abnormal. This cytologic picture is mimicked by
contamination of the slide by squamous cells of the
Endocervical Cells
vulva or skin from the fingers of the persons handling The endocervix is lined by a mucin-producing columnar
the slide. cell that has an eccentrically placed nucleus with a finely
14 Cervical and Vaginal Cytology
A B
Figure 1.9 Keratosis. A, Hyperkeratosis. Anucleate squames are a protective response of the squamous epithelium. B, Parakeratosis.
Parakeratosis appears as plaques, as seen here, or as isolated cells.
A B
Figure 1.10 Endocervical cells. A, Normal endocervical cells are often arranged in cohesive sheets. Note the even spacing of the
nuclei, their pale, finely granular chromatin, and the honeycomb appearance imparted by the sharp cell membranes. B, Sometimes
they appear as strips or isolated cells. Abundant intracytoplasmic mucin results in a cup-shaped nucleus.
The Normal PAP 15
Figure 1.11 Tubal metaplasia. Ciliated endocervical cells are occasionally seen.
Figure 1.12 Endometrial cells. Spontaneously exfoliated The differential diagnosis includes a number of sig-
endometrial cells, as in menses, are small cells arranged in balls. nificant lesions that mimic endometrial cells and thus
Cytoplasm is scant. Nuclei around the perimeter appear to be are sometimes mistakenly interpreted as normal, par-
wrapping around adjacent cells (arrow), a characteristic but non- ticularly if the woman is in the first 12 days of her men-
specific feature.
strual cycle. Attention to certain cytologic details can
help avoid some if not all of these misattributions.
A minority of HSILs are composed of relatively small
Occasionally, endometrial cell clusters consist of an cells. Like endometrial cells, their nuclei are dark, and
obvious dual cell population with small, dark stromal they have scant cytoplasm (Fig. 1.13A). HSIL cells, even
cells (in the center) and larger glandular cells (around the when small, are usually bigger than endometrial cells,
edges). Most endometrial cell clusters, however, do not vary more in size, and have denser cytoplasm. HSIL
have this dual population. “Monocontoured clusters” like clusters are usually less well circumscribed and are not
that in Figure 1.12 may consist of glandular endometrial as spherical as endometrial cell balls. Some poorly dif-
cells, stromal endometrial cells, or a mix of both.106 ferentiated squamous cell carcinomas (SQCs) are com-
Shedding endometrial cells after day 12 (“out of phase”) posed of small dark cells that mimic endometrial cells
is associated with endometritis, endometrial polyps, and to perfection (Fig. 1.13B). In such cases, suspicious
intrauterine devices (IUDs). In a young woman, abnor- clinical findings (e.g., postcoital bleeding) might be the
mal shedding is almost never a result of endometrial only clue to the correct interpretation. Most AIS have a
adenocarcinoma.107,108 For this reason, endometrial cells columnar cell morphology, but a minority are made
do not need to be mentioned in the report for women up of smaller and rounder cells (Fig. 1.13C), particu-
under 40 years of age. Some laboratories do so anyway, to larly on LBC preparations. Careful examination for
document that the cells were identified and interpreted focal columnar differentiation and mitoses can be quite
as benign endometrial cells. Endometrial cells are notori- helpful. The rare small cell carcinoma of the cervix may
ous for their ability to cause diagnostic difficulty, because display crush artifact (Fig. 1.13D), which is rarely seen
a variety of neoplastic cells resemble endometrial cells. with endometrial cells.
16 Cervical and Vaginal Cytology
A B
C D
Figure 1.13 Mimics of exfoliated endometrial cells. A, High-grade squamous intraepithelial lesion (HSIL). The cells of some HSILs
are small but still larger than endometrial cells and usually arranged in flatter aggregates rather than spheres. B, Squamous cell car-
cinoma (SQC). Some poorly differentiated SQCs are indistinguishable from endometrial cells. The granular debris (tumor diathesis)
seen here can also be seen in normal menstrual Pap samples. C, Adenocarcinoma in situ (AIS). Some cases of AIS have an endome-
trioid appearance, but mitoses (arrows) are distinctly uncommon in exfoliated endometrial cells. D, Small cell carcinoma. The cells
resemble endometrial cells but are even darker. There is nuclear smearing, which is rarely seen with benign endometrial cells.
women.111 The cells are large, with abundant blue or pink Decidual cells are isolated cells with abundant gran-
cytoplasm. They have multiple nuclei that have a gran- ular cytoplasm, a large vesicular nucleus, and a promi-
ular chromatin texture and slightly irregular contours. nent nucleolus. They often show degenerative changes.
Trophoblastic cells can be distinguished from multinu-
cleated histiocytes because their nuclei are darker and
more irregular in contour (Fig. 1.15). They do not show
Inflammatory Cells
the prominent molding and ground-glass appearance Neutrophils are seen in all Pap samples and do not
of nuclei of herpes simplex infection. Immunostains for necessarily indicate infection, but they are present
human chorionic gonadotropin and human placental in increased numbers after injury or infection.
lactogen can be used to confirm their identity as tropho- Lymphocytes and plasma cells are rare, but occasionally—
blastic cells. The presence of syncytiotrophoblastic cells most often in older women—they are numerous
is not a reliable predictor of an impending abortion.111 (Figs. 1.16, 1.71A). This pattern is called follicular
cervicitis because biopsies show lymphoid follicle for-
mation. The lymphocytes of follicular cervicitis can be
confused with HSIL cells, endometrial cells, and lym-
phoma. Histiocytes are associated with a myriad of
conditions (e.g., menses, pregnancy, foreign bodies,
radiotherapy, and endometrial hyperplasia and carci-
noma) (Fig. 1.17), but by themselves are a nonspecific
finding of no clinical significance.
Lactobacilli
The vagina is colonized by gram-positive rod-shaped
bacteria of the genus Lactobacillus. They are benefi-
cial because they produce lactic acid, which reduces
the ambient acid-base balance (pH) and possibly
Figure 1.15 Syncytiotrophoblast. The nuclei of these multi- protects from infection by Candida and other patho-
nucleated cells are dark and coarsely granular, unlike those of gens. Lactobacilli metabolize the glycogen contained
histiocytes. within exfoliated squamous cells. The resulting cellular
Figure 1.16 Follicular cervicitis. This smear from a 61-year-old woman contains numerous lymphocytes in various stages of
maturation, including an occasional plasma cell (arrow). Most normal lymphocytes have a round nuclear contour, unlike the cells of a
high-grade squamous intraepithelial lesion (HSIL), to which they bear a superficial resemblance.
Organisms and Infections 19
Figure 1.18 Lactobacilli. These bacteria are part of the normal flora of the vagina. Note the bare nuclei of the intermediate cells,
which are subject to cytolysis by these organisms.
A B
C D
Figure 1.19 Artifacts and contaminants. A, “Cornflaking.” This refractile brown artifact results from bubbles of air trapped on super-
ficial squamous cells, resulting in obscuring of the nuclei. It can be reversed by returning the slide through xylene and alcohol to water,
then restaining and recoverslipping. B, “Cockleburrs.” This is the name given to radiate arrays of club-shaped orange bodies composed
of lipid, glycoprotein, and calcium, surrounded by histiocytes. They are most commonly associated with, but not limited to, pregnant
patients. They have no clinical significance. C, Trichome. These large star-shaped structures are derived from the arrow-wood plant. They
stain a pale yellow and have from three to eight legs. Trichomes are produced by many different plants and vary in color, size, and shape.
D, Carpet beetle parts. These arrow-shaped structures are contaminants from sources such as gauze pads and tampons.
Figure 1.20 Shift in flora suggestive of bacterial vaginosis. Numerous small bacteria cover the slide. In some but not all cases, these
bacteria adhere to squamous cells (“clue cells”), giving them the appearance of a shag rug, as seen here. Lactobacilli are absent.
Organisms and Infections 21
for the diagnosis. Requiring at least 20% clue cells may of HPV, but Trichomonas-related halos are smaller and
increase the specificity of the diagnosis.114 Neutrophils accompanied by only minimal nuclear atypia.
are often scarce. Patients and their sexual partners are treated with
This pattern is common and seen in about 50% of metronidazole.116
patients referred to a dysplasia clinic.115 Clinical cor-
relation is required for a definite diagnosis of bacte-
rial vaginosis because the cytologic pattern is neither
Candida
sufficient nor necessary for the diagnosis. Women who Candida albicans and C. glabrata are fungal species
are symptomatic are treated with metronidazole or that infect the vulva, vagina, and cervix. Patients may be
clindamycin. asymptomatic, or they may complain of burning, itch-
ing, and a thick, cheesy discharge.
Trichomonas Vaginalis Cytomorphology of Candida:
Trichomonas vaginalis is a primitive eukaryotic organ-
• pink
ism, a parasitic protozoan that causes trichomoniasis,
• yeast forms (3 to 7 µm diameter)
a sexually transmitted disease. Patients may experience
• long pseudohyphae and true hyphae
burning, itching, and a malodorous vaginal discharge,
• tangles and skewers of squamous cells around
but up to 50% are asymptomatic.116 Although regarded
pseudohyphae (“spaghetti and meatballs,” “shish
primarily as a disease of women, it also occurs in men,
kebabs”)
most of whom are asymptomatic.
Cytomorphology of Actinomyces:
• tangled clumps of bacteria (“cotton balls,” “dust
bunnies”)
Figure 1.21 Trichomonas vaginalis. This organism has an • long, filamentous organisms
indistinct, ghostly appearance, with a pale oval nucleus and faint • Figure 1.23
red granules.
22 Cervical and Vaginal Cytology
Figure 1.22 Candida. Pseudohyphae and yeast forms, some of them budding from pseudohyphae, are seen. Note the skewered
squamous cells.
Figure 1.23 Actinomyces spp. These bacterial colonies resemble dark cotton balls. The organisms are filamentous, shown here
protruding from the mass of bacteria.
If Actinomyces are seen on a Pap, removal of the IUD Cytomorphology of Herpes simplex
is not necessary, and treatment of asymptomatic women cytopathic changes:
is not recommended.117
• multinucleation
• molding of nuclei
Herpes Simplex • margination of chromatin
• ground-glass nuclei
Infection by the herpes simplex virus is identified by the • eosinophilic intranuclear inclusions
characteristic nuclear changes of infected epithelial cells.
Organisms and Infections 23
The nucleus has a homogeneous, glassy appearance promised.119 In patients who are immunocompetent, the
(“ground-glass”), and nuclear membranes are thick infection is transient and usually asymptomatic.
resulting from peripheral margination of chromatin
(Fig. 1.24A). Multinucleation is common, with molding Cytomorphology of cytomegalovirus
of nuclei. Eosinophilic intranuclear inclusions may be cytopathic changes:
present.
• mononuclear cells
• markedly enlarged
• basophilic intranuclear inclusion
Cytomegalovirus • small granular cytoplasmic inclusions
Exposure to and infection by cytomegalovirus (CMV) is
common in the general population, but clinical mani- Infected cells are enlarged, and the nuclei have a soli-
festations, such as mononucleosis, are relatively uncom- tary basophilic inclusion surrounded by a halo. Multiple
mon. The cytologic changes of cytomegalovirus infection small, granular cytoplasmic inclusions are also present
can be seen on cervical-vaginal preparations from women (Fig. 1.24B). The infected cells are endocervical or ecto-
who are immunocompetent and who are immunocom- cervical in origin.120
A B
Figure 1.25 Benign squamous cell changes. A, PM cells. Nuclear enlargement, with little in the way of nuclear membrane irreg-
ularity or hyperchromasia, is a common finding in intermediate squamous cells from perimenopausal women. Such bland nuclear
enlargement should not be mistaken for a significant atypia. B, A similar bland nuclear enlargement occurs in metaplastic cells.
Benign and Reactive Changes 25
nuclear membrane irregularity, these cells are unlikely to variation in nuclear size, and nucleoli are sometimes
represent a significant squamous lesion.126 The cause of prominent. Smooth nuclear membranes and finely tex-
nuclear enlargement in squamous cells from perimeno- tured chromatin are reassuring. In some cases, however,
pausal women is not known. the alterations in metaplastic squamous cells are more
Nonspecific perinuclear cytoplasmic clearing in super- marked and overlap with the features of HSIL. Such bor-
ficial and intermediate squamous cells is associated with derline cases are called atypical squamous metaplasia.
inflammatory conditions like Trichomonas infection,
but it can also be a slide preparation artifact. It is distin-
guished from koilocytosis by the small size of the halo and
Benign Endocervical Changes
the absence of increased cytoplasmic density outlining Reactive endocervical cells often show much greater
the cavity (Fig. 1.26A). Large cytoplasmic clearings occur increases in nuclear size than squamous cells. Some reac-
in squamous cells with abundant cytoplasmic glycogen. tive endocervical cell nuclei are four or five times larger
They are distinguished from LSIL cells because they have than normal, usually with an accompanying increase in
a normal intermediate cell nucleus (Fig. 1.26B). cytoplasm. The enlarged nuclei remain round or oval,
Squamous metaplastic cells are particularly prone to but they frequently have a large nucleolus (Fig. 1.27).
reactive changes. There can be nuclear enlargement and Such changes are not uncommon in pregnancy, where in
A B
Figure 1.26 Nonspecific halos. A, Small halos around the nuclei of squamous cells are nonspecific and do not represent human
papillomavirus (HPV)-related changes. B, Some normal squamous cells have abundant glycogen that mimics koilocytosis. Note the
normal nucleus.
A B
Figure 1.27 Reactive endocervical cells. A, A common finding, reactive endocervical cells are enlarged and have a prominent
nucleolus. B, Isolated cells can be as big as mature squamous cells and mimic a low-grade squamous intraepithelial lesion (LSIL), but
a prominent nucleolus is uncharacteristic of an LSIL.
26 Cervical and Vaginal Cytology
their extreme form they represent the Arias-Stella reac- invasive cancer. The differential diagnosis of reactive
tion.127 They are also seen in patients with endocervical endocervical cells is discussed in greater detail in the cor-
polyps and inflammation of any cause. responding sections that follow. Ultimately, the benign
Reactive endocervical cells are also seen in micro- nature of reactive endocervical cells is betrayed by the
glandular hyperplasia, a benign alteration of endocer- roundness of the nucleus, its fine chromatin granularity,
vical epithelium associated with oral contraceptive use. and the normal nuclear-to-cytoplasmic ratio.
Microglandular hyperplasia was originally described in
histologic material, where it was sometimes confused
with adenocarcinoma. Cytologic changes range from
Repair
entirely normal endocervical cells to marked nuclear Reparative changes result from injury to the cervical
enlargement, often with prominent nucleoli and cyto- epithelium and the proliferation of reserve cells, which
plasmic vacuolization (Fig. 1.28).128 Clinical correlation grow to reepithelialize a focus of ulceration.
is useful. Knowledge that the patient is pregnant or has
a visible endocervical polyp can alert the cytologist to Cytomorphology of repair:
the possibility of reactive changes and provide a ratio-
• cohesive, flat sheets
nal explanation for the alterations. In their most extreme
• streaming appearance
forms, however, reactive endocervical cells raise a dif-
• large nucleus with marked size variation
ferential diagnosis that includes LSIL, HSIL, AIS, and
• large nucleolus, sometimes irregular
• pale chromatin
• mitoses
Figure 1.30 Radiation effect. Radiation looks like a wild reparative reaction, with large cells, multinucleation, cytoplasmic vacuol-
ization, and a curious “two-tone” cytoplasmic staining pattern.
28 Cervical and Vaginal Cytology
Differential diagnosis of
intrauterine device effect:
Other Benign Changes
The cells of tubal metaplasia of the endocervix often
• adenocarcinoma
look like normal endocervical cells, except that they
• HSIL
have cilia. Sometimes they have a higher nuclear-to-
cytoplasmic ratio and slight hyperchromasia and may
The vacuolated cells of IUD effect are virtually indis- be mistaken for a significant squamous or glandular
tinguishable from the cells of an adenocarcinoma, lesion if a careful search is not made for cilia.132 Cilia
particularly those of endometrial origin. If the woman are reliable evidence that the cell they are attached to is
has an IUD, these changes are most likely benign, but benign because ciliated adenocarcinomas of the endo-
clinical correlation and a repeat Pap after removal of the cervix are uncommon.133,134 Endometriosis of the cervix
IUD might be considered. The small IUD cells resemble resembles abraded endometrium (see “Abraded endo-
HSIL cells except that they have a nucleolus.130 metrium and lower uterine segment” above).
ASC-US, atypical squamous cells of undetermined significance; HSIL, high-grade squamous intraepithelial lesion; LSIL, low-grade squamous
intraepithelial lesion. From Melnikow J, Nuovo J, Willan AR, et al.: Natural history of cervical squamous intraepithelial lesions: A meta-analysis.
Obstet Gynecol 1998;92(4 Pt 2):727-735.
30 Cervical and Vaginal Cytology
Inhibits apoptosis
by binding p53
L2
E6
Host Cell
HPV E7
L1 genome
E1 S G2
E7 pRb
E2 Binds to
E5 G1 M
E4 pRb to E7 pRb
induce
S-phase
entry
Viral capsid
protein
Figure 1.33 The human papillomavirus (HPV) genome and its effects on the host cell. The HPV genome has early (E) and late
(L) genes. The E6 and E7 genes are most responsible for the transforming effects of integrated HPV DNA on the host cell. Inset:
Detection of HPV by in situ hybridization. The dark brown signal is centered on the nucleus of infected cells. (Courtesy of Miu-Fun
Chau, DakoCytomation, Carpinteria, Calif.)
the epithelium. HPV infection is established in the basal HPV 16 is the prototype of the high-risk viruses and
layers of the epithelium, where the HPV genome is main- the one most commonly detected in cervical cancers.
tained, with expression of the E genes. As the epithelium A variety of molecular techniques—the polymerase
matures toward the surface, gene amplification and viral chain reaction, in situ hybridization (Fig. 1.33 inset), and
assembly occur, with expression of L1 and L2, with even- hybrid capture—can be used to detect HPV within cer-
tual viral release. L1 is the major viral capsid protein and is vical lesions. The Hybrid Capture 2™ test, which was
the principal component of the HPV vaccines. The E6 and evaluated in the multicenter ASCUS/LSIL Triage Study
E7 gene products play the most significant part in cervical (ALTS) trial sponsored by the National Cancer Institute,
oncogenesis. They have a number of cellular targets, with uses a cocktail of probes to the 13 high-risk HPV types
a multitude of effects that lead to malignant transforma- listed, which account for nearly 90% of HPVs detected in
tion.149 The two most important appear to be (1) the bind- HSIL and invasive cancers.38
ing of E6 to p53, which results in the blocking of apoptosis, The risk of HPV infection per sexual contact is not
and (2) the binding of E7 to the retinoblastoma tumor sup- known but is probably fairly high. Most women, if they
pression protein pRB, which abolishes cell-cycle arrest are sexually active, are infected with one or more HPV
and leads to unscheduled cellular proliferation.149,150 types at some point in their lives. For unclear reasons,
More than 100 types of HPV have been isolated, of which the virus has a strong predilection for the transfor-
more than 40 infect the female genital tract. Only a minor- mation zone. Serology is not an accurate measure of
ity cause cervical cancer. The genital HPVs are divided into infection, because only 50% to 60% of infected women
low-risk and high-risk types based on the frequency of their have circulating antibodies to HPV.151 Clearly, only a
association with invasive cervical cancer. By definition, an minority of HPV infections persist and lead to cancer.
HPV is low risk if it has never been isolated from a cervical Cellular immune responses play a role in clearing infec-
carcinoma and high risk if it ever has been. Persistent infec- tion, but how they work is still poorly understood.
tion with any one of about 15 high-risk (carcinogenic) types
accounts for virtually all cervical cancers.149
Grading Squamous Intraepithelial Lesions
Examples of low-risk and high-risk human The Bethesda System recommends a low-grade/high-
papillomaviruses: grade approach to grading SIL. This is based on the
evidence that most LSILs are transient infections that
• low-risk: 6, 11, 42, 43, 44, 53, 54, 57, and 66
carry little risk for oncogenesis, whereas most HSILs
• high-risk: 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59,
are associated with viral persistence and a significant
and 68
potential for progression to invasive cancer.
Squamous Abnormalities 31
B
Figure 1.34 Low-grade squamous intraepithelial lesions (LSIL). A, LSIL. Classic koilocytes, as seen here, have a large cytoplasmic
cavity with a sharply defined inner edge and are frequently binucleated. Nuclear enlargement may not be as marked as in the non-
koilocytic LSILs. B, Nonkoilocytic LSIL. Nuclei are significantly enlarged and show mild hyperchromasia and nuclear contour irregular-
ity. No definite koilocytes are seen. This pattern was once called mild dysplasia or CIN 1.
colposcopy. At the 24-month follow-up, those with a refer the woman to colposcopy only if the HPV test is
Pap showing ASC-US or greater should be referred for positive or any one of the Paps is ASC-US or greater.39
colposcopy.
As with adolescents, postmenopausal women with
High-Grade Squamous Intraepithelial
an LSIL Pap can be managed less aggressively than
Lesion
premenopausal women. Although immediate colpos-
copy is an option, it is acceptable instead to repeat Pap HSIL is an intraepithelial lesion that is encountered in
testing at 6 and 12 months or perform an HPV test and about 0.5% of all Pap samples. Virtually all women (97%)
Squamous Abnormalities 33
Figure 1.36 Management guidelines for women with a Pap showing a low-grade squamous intraepithelial lesion (LSIL).
Management may vary if the woman is an adolescent, postmenopausal, or pregnant. (Reprinted with the permission of ASCCP ©
American Society for Colposcopy and Cervical Pathology 2008.)
34 Cervical and Vaginal Cytology
• exfoliated endometrial cells Distinguishing HSIL from its many mimics is an impor-
• follicular cervicitis tant skill of the cytologist. As with histologic sections, one
• histiocytes of the most frequent cytologic mimics is squamous meta-
• IUD effect plasia. Squamous metaplastic cells commonly show only
• endocervical polyp atypia mild nuclear enlargement, nuclear membrane irregu-
• AIS larity, and even chromatin coarsening. These changes
• SQC rarely rise to the level of atypia seen in HSIL. In post-
• atypical squamous cells—cannot exclude HSIL menopausal women, sheets of atrophic squamous epi-
• ASC-US associated with atrophy (see Fig. thelium mimic the syncytium-like clusters of HSIL (see
1.50B-D) Fig. 1.6A). Although atrophic squamous cells have a
high nuclear-to-cytoplasmic ratio, their nuclei are usually
Squamous Abnormalities 35
regular, with finely textured chromatin. Transitional cell pausal women, are smaller than HSIL cells, their chromatin
metaplasia, associated with Pap samples from older is even more coarsely textured, and there are often admixed
women, is likely to raise the possibility of HSIL because of plasma cells, dendritic cells (with a larger, pale nucleus), and
the irregularity of the nuclear outlines and prominence of tingible-body macrophages (see Fig. 1.16). Histiocytes are
nuclear grooves (see Fig. 1.6B). The absence of hyperchro- roughly the same size as HSIL cells, and many have irregu-
masia and the abundance of coffee-bean shaped nuclei is lar nuclear contours, but their chromatin is finely textured;
a clue to the benign metaplastic nature of these cells. HSIL they often have abundant fluffy cytoplasm (see Fig. 1.17).
cells, even those of the small-cell type,157 are usually big- The small cells of IUD effect are usually few in number and
ger than endometrial cells (see Fig. 1.12), vary more in size, have a more prominent nucleolus than is commonly seen
and have denser cytoplasm. HSIL clusters are usually less with HSIL (see Fig. 1.31). Occasional inflamed endocervical
well circumscribed and not as spherical as endometrial cell polyps are lined by a single layer of highly atypical, hyper-
clusters. Lymphoid cells, commonly seen in postmeno- chromatic endocervical cells that are easily overinterpreted
36 Cervical and Vaginal Cytology
A B
Figure 1.42 Endocervical polyp atypia mimicking HSIL. A, The slide contains scattered isolated cells with dark nuclei. B, The
surface of the endocervical polyp reveals a single layer of reactive endocervical cells.
as HSIL. Their true nature is often clarified only after histo- In adolescents with an HSIL Pap, colposcopy is the
logic correlation (Fig. 1.42). recommended management. (The “see-and-treat”
The neoplastic cells of AIS share many of the nuclear approach is unacceptable.) If colposcopy confirms CIN
features of HSIL. Clusters of neoplastic cells are more 2,3, either treatment or observation for up to 2 years
likely to represent HSIL rather than AIS, unless there is is acceptable, provided that colposcopy was satisfac-
clear columnar differentiation in the form of feathering or tory.156 If colposcopy is negative (no lesion or only CIN
rosette formation. SQC should be considered whenever 1 is confirmed by biopsy), observation with colpos-
the cytologic criteria for HSIL are fulfilled, but in addition copy and Pap testing at 6-month intervals is recom-
one finds prominent nucleoli or necrotic debris. mended, provided that colposcopy is satisfactory and
In some cases, uncertainty remains regarding the true endocervical sampling is negative.39 If HSIL cytology
nature of the cells examined. Cells with the features of persists for 24 months without histologic confirmation,
squamous metaplasia sometimes show a degree of nuclear a diagnostic excisional procedure is recommended.
atypia that makes it impossible to exclude an HSIL. These A diagnostic excisional procedure is recommended if
“atypical squamous metaplasias” are reported as “ASC, colposcopy is unsatisfactory or CIN of any grade is found
cannot exclude HSIL” (ASC-H). Another diagnostically dif- on endocervical assessment.
ficult pattern is the marked squamous atypia associated In pregnant women with an HSIL Pap, it is recom-
with a deeply atrophic Pap. Atrophic cervical epithelium mended that colposcopy be performed by a physician
sometimes displays a marked squamous atypia that is experienced with this technique in patients who are
impossible to distinguish from HSIL. The recommended pregnant. Biopsy of lesions suspicious for CIN 2,3 or
approach is to call such cases ASC-US. cancer is preferred, and biopsy of other lesions is accept-
The recommended management of a woman with able. Endocervical curettage is unacceptable. If invasive
an HSIL Pap is illustrated in Figure 1.43. Management is cancer is suspected, a diagnostic excisional procedure
more aggressive than it is for an LSIL Pap, based on the is acceptable. If CIN 2,3 has not been diagnosed histo-
conviction that cytologic HSIL has a higher risk of pro- logically, reevaluation with colposcopy and Pap testing
gression to invasive cancer. With the exception of ado- is recommended no sooner than 6 weeks postpartum.39
lescents and those who are pregnant, an immediate loop
electrosurgical excision (the “see-and-treat” approach)
Problems in the Diagnosis of Squamous
is acceptable as the initial treatment if the woman has
Intraepithelial Lesions
an HSIL Pap, but not LSIL. An alternative to loop is col-
poscopy with endocervical assessment (evaluating the Avoiding Overdiagnosis of Low-Grade Squamous
canal using the colposcope or tissue sampling). If col- Intraepithelial Lesions. Care must be taken not to
poscopy confirms CIN 2,3, the lesion is surgically excised overinterpret nonspecific halos (see Fig. 1.26A and B)
or ablated.156 If colposcopy is negative (no lesion or or the minimal nuclear changes of benign cells like the
only CIN 1), either a diagnostic excisional procedure or PM cells of perimenopausal women (see Fig. 1.25A).126
observation with colposcopy and Pap testing at 6-month Without hyperchromasia or nuclear membrane irreg-
intervals is acceptable, provided that colposcopy is satis- ularity, such cells are best called negative. Cellular
factory and endocervical sampling is negative.39 changes that include some hyperchromasia or nuclear
Squamous Abnormalities 37
Figure 1.43 Management guidelines for women with a Pap showing a high-grade squamous intraepithelial lesion (HSIL).
Immediate loop excision or colposcopy is acceptable for women with an HSIL Pap. The management options may vary if the woman
is pregnant, postmenopausal, or adolescent. (Reprinted with the permission of ASCCP © American Society for Colposcopy and
Cervical Pathology 2008.)
membrane irregularity are suggestive of LSIL and should but not conclusive for HSIL. Extensively keratinized
be categorized as ASC-US. SILs without definite HSIL are especially difficult to
grade159 (Fig. 1.44). In all such cases, a diagnosis of “SIL,
Distinguishing Low-Grade from High-Grade Squamous grade cannot be determined” (or “LSIL, cannot exclude
Intraepithelial Lesions. The distinction between cyto- HSIL”) is appropriate.160 This diagnosis accounts for 3%
logic LSIL and HSIL is an important one, with signifi- to 12% of all cytologic SILs.155,161–163 Patients with this
cantly different implications for clinical management. diagnosis have an intermediate risk (between that of
Proficiency in this distinction is an important skill of the cytologic LSIL and HSIL) of harboring histologic HSIL
cytology practitioner. As mentioned previously, HSIL is (CIN 2,3).160–162
usually a lesion of immature squamous cells, and nuclear
atypia (hyperchromasia, irregular chromatin distribu- Distinguishing High-Grade Squamous Intraepithelial
tion, and membrane contour irregularity) is more severe Lesion from Invasive Carcinoma. The criteria used to
than in LSIL. If a specimen is composed of both LSIL and distinguish HSIL from invasive carcinoma are by no
HSIL, it should be reported as an HSIL even if the HSIL means perfect. Not infrequently, a classic case of HSIL
cells are less numerous than the LSIL cells. In a small on cytology will turn out to be invasive squamous can-
percentage of cases, morphologic features intermediate cer on biopsy. Conversely, the possibility of invasive
between typical LSIL and HSIL make grading difficult.158 cancer is often raised in cases of HSIL in which the cells
Although there are generally fewer abnormal cells in an have marked nuclear abnormalities associated with
LSIL than in an HSIL, the quantity of cells is an unreli- abundant, heavily keratinized cytoplasm and unusual
able discriminator. cell shapes, but the lesion turns out to be only a kera-
tinizing HSIL on biopsy.78 Physicians understand that
Cytomorphologic patterns of “SIL, no diagnosis of HSIL on cytologic material excludes
grade cannot be determined”: the possibility of invasive cancer, and that colpos-
copy and biopsy are necessary for confirmation. Some
• few dysplastic cells
HSILs with features worrisome for invasive cancer
• extensive cytolysis
can be reported as “HSIL, with features suggestive of
• LSIL, with a small number of equivocal HSIL cells
invasive cancer.”
• extensively keratinized SILs, without definite HSIL
A B
Figure 1.44 Squamous intraepithelial lesion (SIL), cannot determine grade. When a lesion is extensively keratinized and there
is no definite high-grade squamous intraepithelial lesion (HSIL), it is difficult to grade. Colposcopically directed biopsies showed
A, CIN 1 and B, CIN 2,3.
menstrual bleeding. When associated with hyperchromatic Most SQCs are associated with an adjacent or over-
crowded groups of atypical cells or abundant atypical kera- lying HSIL, and therefore cytologic preparations from
tinized cells with unusual shapes (“tadpoles,” “fiber cells”), SQCs often contain a population of HSIL cells as well.
the pattern is diagnostic.
The cells of a nonkeratinizing SQC look like modified Differential diagnosis of squamous
HSIL cells (Figs. 1.46, 1.47). Like HSIL, they are hyper- cell carcinoma:
chromatic and have scant cytoplasm, but they have a
• HSIL
prominent nucleolus and a highly irregular pattern of
• atypia of atrophy
chromatin distribution. The cells of a keratinizing SQC
• atypia of repair
are often bizarrely elongated (Fig. 1.48). Some are long
• benign endometrial cells
and spindle shaped, with small condensed nuclei (“fiber
• Behçet disease
cells”). Others have a larger cytoplasmic body with a long
• pemphigus vulgaris
tail (“tadpole cells”). Such cells are uncommon in kera-
tinizing HSILs.
The differential diagnosis of SQC includes HSIL.
Prominent nucleoli and tumor diathesis are the principal
cytologic features that help distinguish SQC from HSIL,
but these features are not present in all smears from
patients with SQC. A significant number of women with
SQC are diagnosed as having HSIL because prominent
nucleoli and tumor diathesis are absent.167 Conversely, a
granular, tumor diathesis-like background is not specific
for invasive cancers and is seen in women with atrophic
vaginitis77 (see Fig. 1.7), severe cervicitis, and rare cases
of HSIL.78
In postmenopausal women, marked atrophy atypia is
one of the most common benign mimics of a keratiniz-
ing SQC (see Fig. 1.50B-D). The benign atypia of atrophy
contains scattered cells with large, dark nuclei and eosin-
ophilic or orangeophilic cytoplasm. Their large, dark
nuclei are alarming, but chromatin is usually smudgy.
Figure 1.46 Squamous cell carcinoma (SQC), nonkeratinizing. Such cells, if seen in a deeply atrophic squamous back-
The malignant cells have irregularly distributed chromatin and a ground, should be interpreted as ASC-US and not HSIL
prominent nucleolus, characteristic features of invasive SQCs. or invasive cancer.
A B
Figure 1.48 Squamous cell carcinoma, keratinizing. A, In keratinizing carcinomas, the cells have markedly aberrant shapes, as
seen here. “Fiber cells” are numerous. B, A tadpole cell and some tumor diatheses are seen in this tumor.
Marked repair atypia is another good mimic of nonke- ( surgical removal of the cervix) with lymph node dissec-
ratinizing SQC (see Fig. 1.52). Both repair and SQC con- tion. If intermediate- or high-risk histologic features are
tain large cells with prominent nucleoli, and mitoses are found after hysterectomy, postoperative radiotherapy
seen in both. Repair cells are recognized by their finely (with or without chemoradiation) improves local con-
textured chromatin pattern, the flatness and cohesion of trol and survival. Patients with higher stage disease (IB2
the sheets. If the nuclei have coarsely textured chroma- to IVA) are likely to be treated with external beam and
tin, show marked crowding, or demonstrate significant intracavitary radiation combined with cisplatin-based
dyshesion, SQC should be considered. chemotherapy.170 Women with metastatic cancer (stage
A minority of nonkeratinizing SQCs are composed IVB) are best treated with systemic chemotherapy, with
of small cells that are indistinguishable from endome- radiation therapy reserved for palliation of symptomatic
trial cells (see Fig. 1.13B). The blood that accompanies pelvic disease.
menstrual endometrial cells resembles the granular
necrosis that is tumor diathesis, adding to the similar-
ity. Mitoses, if identified, should raise the suspicion of
Atypical Squamous Cells
SQC. In some cases, knowledge that the patient has a Since the days of Papanicolaou, cytology laboratories
suspicious cervical mass or suspicious clinical symp- have used a borderline category to report findings of
toms (e.g., dyspareunia) may be the only clue to the cor- uncertain significance. Terminology was inconsistent
rect interpretation. and often confusing, however, because benign changes
Behçet disease, a chronic disease of uncertain cause were sometimes reported as “benign atypia.” In the
that is characterized by oral and genital ulcers, can mimic Bethesda System, recognizably benign cases, previously
SQC. Smears may show numerous isolated, keratinized called “benign atypia,” “inflammatory atypia,” or “reac-
cells with dark, pleomorphic nuclei and large nucleoli. tive atypia,” are excluded from this category. The 1988
A history of this disorder may be critical for correct diag- and 1991 Bethesda Systems used the term atypical squa-
nosis.168 Smears from patients with pemphigus vulgaris, mous cells of undetermined significance to designate
a blistering disorder that involves mucosal surfaces, may “cellular abnormalities that were more marked than
mimic a poorly differentiated SQC. A complete history those attributable to reactive changes but that quantita-
may be important to avoid making an overcall, although tively or qualitatively fell short of a definitive diagnosis of
cases of coexisting SQC and pemphigus vulgaris have SIL.” In the 2001 Bethesda System, ASC-US was replaced
been reported.169 by ASC and redefined in a subtle way. Instead of being
Treatment choices for women with cervical can- a diagnosis of exclusion, ASC is a diagnosis conveying a
cer include surgery (hysterectomy plus lymphadenec- suspicion of SIL.
tomy), radiation therapy, and chemoradiation, depending Most cytologists agree that this category is essen-
on tumor stage.170 Hysterectomy (simple or radical, tial. Eliminating ASC would result in increased report-
depending on histologic findings) is the treatment of ing of LSIL (which probably contributes little to cancer
choice for early-stage (IA1, IA2, and IB1), nonbulky prevention) and decreased reporting of HSIL.171 It is
disease. Women with early-stage disease who wish risky to eliminate an equivocal category because of the
to preserve fertility have the option of trachelectomy large number of women with underlying HSIL who are
Squamous Abnormalities 41
A B
Figure 1.49 Atypical squamous cells of undetermined significance (ASC-US). A, The nucleus of this mature squamous cell is
significantly enlarged and there is moderate hyperchromasia. Cells like this, particularly if few in number, are suggestive but not
diagnostic of a squamous intraepithelial lesion (SIL). B, Some cells have large cytoplasmic cavities but minimal nuclear atypia. It is
preferable to diagnose such cases as ASC-US when abnormal cells are few and the changes minimal.
42 Cervical and Vaginal Cytology
A B
C D
Figure 1.50 Atypical squamous cells of undetermined significance (ASC-US), associated with atrophy. A, Histologic section of
benign atrophy-associated atypia. B, Cytologic smear shows scattered large atypical cells in a granular background. C, Some cells
have a markedly enlarged, hyperchromatic nucleus. D, Often cells are poorly preserved, with smudgy nuclei and hypereosinophilic
cytoplasm. Follow-up in all cases was benign.
times a week for several months), followed by a repeat are difficult to distinguish from invasive carcinoma.
Pap test a week after completing the regimen.179 A signifi- Carcinomas often have a tumor diathesis and many iso-
cant squamous lesion will be more easily detected among lated atypical cells, features that are usually absent in
the mature cells, whereas a benign “atypia” resulting from repair reactions.
atrophy will be transformed into normal epithelium.
Squamous atypia in a postmenopausal woman is less
often associated with a biopsy-proven SIL (17%) than in
a premenopausal woman (46%).180 Further, the rate of
HPV detection in women with atypia is lower (10% ver-
sus 50%). In another study, squamous atypia in women
over the age of 50 was associated with histologic SIL in
less than 5% of cases.181
Parakeratosis with mild nuclear enlargement and mild
to moderate nuclear membrane irregularity (atypical
parakeratosis) suggests an SIL (Fig. 1.51). In some cases
such cells are accompanied by other changes diagnostic
of an SIL, but when the changes are mild, such cases are
best classified as ASC-US.
Highly exuberant atypical repair reactions can dem-
onstrate cellular crowding and overlap (in contrast with Figure 1.51 Atypical squamous cells of undetermined sig-
nificance (ASC-US), with features of atypical parakeratosis.
typical repair, which is in flat sheets), marked variation Small, keratinized squamous cells with mild variation in nuclear
in nuclear size, prominent and irregular nucleoli, and size and contour may represent either a reactive process or a
irregular chromatin distribution (Fig. 1.52). Such cases significant squamous lesion.
Squamous Abnormalities 43
Figure 1.53 Management guidelines for women with atypical squamous cells of undetermined significance (ASC-US). Human
papillomavirus (HPV) testing is preferred if liquid-based cytology or co-collection is available (“reflex HPV testing”). (Reprinted with
the permission of ASCCP © American Society for Colposcopy and Cervical Pathology 2008.)
44 Cervical and Vaginal Cytology
A B
Figure 1.55 Atypical squamous cells, cannot exclude high-grade squamous intraepithelial lesion (ASC-H). A, Immature squa-
mous metaplastic cells sometimes show some nuclear atypia that raises the possibility of high-grade squamous intraepithelial lesion
(HSIL), but the degree of nuclear enlargement, hyperchromasia, and membrane irregularity is insufficient for a definite diagnosis.
B, Subsequent colposcopy revealed benign immature squamous metaplasia, and a human papillomavirus (HPV) test on the residual
ThinPrep vial was negative for high-risk HPV.
Glandular Abnormalities 45
Figure 1.56 Management guidelines for women with atypical squamous cells, cannot exclude high-grade squamous intraepi-
thelial lesion (ASC-H). ASC-H warrants immediate colposcopy. (Reprinted with the permission of ASCCP © American Society for
Colposcopy and Cervical Pathology 2008.)
In the 2001 Bethesda System, AIS is a separate diag- cells in sheets reveal their glandular nature by “feath-
nostic category because there is a consensus that the ering,” a splaying out around the edges (Fig. 1.58B).
cytologic criteria are accurate and reproducible.187–189 Nuclei are hyperchromatic and crowded and there
Examination of the slide under low magnification is scant cytoplasm. Apoptotic bodies are seen in
reveals hyperchromatic crowded groups similar to most cases and are a useful clue to the diagnosis.190
those of HSIL (Fig. 1.57). Closer inspection reveals Mitoses are seen in some cases and are helpful, but
evidence of glandular differentiation: columnar cells only if accompanied by the typical nuclear changes
arranged in strips or rosettes (Fig. 1.58A). Columnar previously described.
A B
Figure 1.58 Adenocarcinoma in situ (AIS). A, Rosettes are highly characteristic of AIS and virtually never seen with high-grade
squamous intraepithelial lesion (HSIL), benign endocervical cells, or lower uterine segment (LUS) or endometrial epithelium. B, The
glandular nature of these neoplastic cells is betrayed by “feathering.”
A B
C D
Figure 1.59 Adenocarcinoma in situ (AIS) compared to tubal metaplasia. A, Endocervical AIS. Cells are columnar in shape,
dark, crowded, and arranged in a curved strip. B, A cone biopsy revealed AIS. C, Tubal metaplasia. Atypical glandular cells bear a
resemblance to those in A, except that cilia are identified. D, Subsequent biopsies showed tubal metaplasia of surface endocervical
epithelium.
B
Figure 1.60 Endocervical adenocarcinoma. A, The cells are round rather than elongated as in adenocarcinoma in situ (AIS). They are
crowded and hyperchromatic, and a tumor diathesis is present. Tumor diathesis on liquid-based preparations appears as clumps and as
a granular ring around groups of malignant cells (“clinging diathesis”). B, High magnification reveals the crowding and large nucleoli.
The rare, extremely well-differentiated tumor known they resemble AIS because, like AIS, the cells appear uni-
as minimal deviation adenocarcinoma (or adenoma form, and crowded, with mild to moderate aytpia. Like
malignum) is composed of mucinous glands that show AIS, strips and rosettes are seen,135 and there is no tumor
little if any atypia, and yet, if untreated, invade deeply diathesis.198 Few are diagnosed prospectively as an ade-
and metastasize. Patients sometimes present with nocarcinoma. Most are reported as benign or as “atypi-
vaginal discharge. In most cases, the neoplastic cells cal glandular cells.”198
on the Pap test look entirely like normal endocervi- The cytologic features of the rare adenoid cystic carci-
cal cells197 (Fig. 1.63A). Frequently, even cervical biop- noma and mucoepidermoid carcinoma of the cervix are
sies and endocervical curettings are misinterpreted as similar to their counterparts in the salivary gland and
benign. A correct diagnosis often requires at least a cone elsewhere (see Figs. 10.17, 10.18, 10.20 to 10.22).
biopsy to appreciate the invasive nature of the lesion
Endometrial Adenocarcinoma
(Fig. 1.63B).
Villoglandular adenocarcinomas are rare low-grade Endometrial adenocarcinoma is predominantly a tumor
neoplasms that rarely if ever metastasize. Cytologically, of postmenopausal women, with a peak incidence in
Glandular Abnormalities 49
A B
C D
Figure 1.61 Endocervical adenocarcinoma compared to reactive endocervical cells. A, Endocervical adenocarcinoma, well dif-
ferentiated. The cells are enlarged and crowded, but the features are not conclusive for malignancy (note the absence of tumor dia-
thesis). A diagnosis of atypical glandular cells was made. B, Histologic sections showed adenocarcinoma. C, Reactive endocervical
cells. These cells appear similar to those in A. D, Biopsies in this patient confirmed reactive changes resulting from inflammation.
Figure 1.62 Endocervical adenocarcinoma. These malignant cells show variation in nuclear size, with prominent nucleoli and
coarsely granular chromatin.
50 Cervical and Vaginal Cytology
A B
Figure 1.63 Minimal deviation adenocarcinoma. A, The cells are sometimes impossible to distinguish from normal endocervical
cells, as in this case. B, A cone biopsy revealed deeply invasive, misshapen neoplastic glands.
A B
Figure 1.64 Endometrial adenocarcinoma compared to intrauterine device (IUD) effect. A, Endometrial adenocarcinoma, endo-
metrioid type. These malignant cells are large, vacuolated, and associated with neutrophils. B, IUD effect. Benign cells in women with
an IUD are indistinguishable morphologically from those of endometrial adenocarcinomas.
Glandular Abnormalities 51
background are characteristic. Compared with smears and tend to exfoliate as single cells and smaller clusters,
from the endometrioid type, smears from papillary often arranged as spheres, whereas the cells of endocer-
serous adenocarcinomas contain more malignant cells.201 vical adenocarcinomas are more columnar and more
Psammoma bodies are present in only 25% of cases.201 commonly shed as sheets of cells. Histiocytes commonly
Pap slides are more likely to contain malignant cells in accompany endometrial carcinomas and not endocervi-
patients with a serous rather than an endometrioid type cal carcinomas. Ultimately, the cytologist can usually only
of endometrial adenocarcinoma.202,203 suggest the possibilities, favoring one site over another;
the final classification rests on histologic examination.
Differential Diagnosis of Adenocarcinoma
Adenocarcinoma of the vagina is rare and often
Because there is significant morphologic overlap associated with a maternal history of DES use during
between adenocarcinomas of the cervix, endometrium, pregnancy.
and other sites, they are considered together. Adenocarcinomas from the ovaries and fallopian
tubes are more commonly associated with psammoma
Differential diagnosis of bodies,204 but this is not entirely reliable because endo-
adenocarcinoma: cervical and endometrial cancers sometimes contain
them as well.
• endocervical adenocarcinoma
Nonkeratinizing SQCs resemble endocervical ade-
• endometrial adenocarcinoma
nocarcinomas. Unless focal keratinization is identified,
• adenocarcinoma of other sites:
a definite distinction is not possible. The cells of IUD
• vaginal
effect are indistinguishable from those of endometrial
• ovarian
adenocarcinoma (Fig. 1.64B). If the woman has an IUD,
• tubal
it is likely that the cells represent IUD effect rather than
• metastatic
an adenocarcinoma.
• SQC
Enlarged, vacuolated cells with engulfed neutrophils
• IUD effect
(“bags of polyps”) are seen with inflamed endocervical
• endocervical polyp atypia
polyps and represent reactive endocervical cells (Fig.
• reactive endocervical cells
1.65), yet they mimic a similar cell that is characteris-
• AIS
tic of endometrial carcinoma. Morphologic distinction
• pemphigus vulgaris
can be impossible, and knowledge that the patient has
an endocervical polyp may be the only clue to correct
interpretation.
When adenocarcinoma cells are identified on a Pap slide, Reactive endocervical cells, including atypical
the two principal suspects are endocervical and endome- repair, can mimic adenocarcinomas and vice versa.71
trial adenocarcinoma. The age of the patient is helpful: The Reactive cells, paradoxically, often show more marked
older the patient, the more likely it is that the tumor has variation in nuclear size and nucleolar size and shape
arisen in the endometrium. Morphologic features are also than adenocarcinomas, which are often deceptively
helpful. Endometrial adenocarcinoma cells are rounder uniform.71 Reactive cells have thin nuclear membranes
A B
Figure 1.65 Inflamed endocervical polyp mimicking endometrial adenocarcinoma. A, The large vacuolated cells are associated
with neutrophils, just like the cells of endometrial adenocarcinoma. B, Histologic sections reveal an acutely inflamed polyp line by reac-
tive endocervical cells infiltrated by polyps.
52 Cervical and Vaginal Cytology
Figure 1.66 Management guidelines for women with atypical glandular cells (AGC). The guidelines are different for atypical
endometrial cells versus all other subcategories of atypical glandular cells. (Reprinted with the permission of ASCCP © American
Society for Colposcopy and Cervical Pathology 2008.)
Figure 1.68 Malignant melanoma of the vagina. The malignant spindled and epithelioid cells are dyshesive. There is focal finely
granular melanin pigment (arrow).
Endometrial Cells in Women Older than 40 Years of Age 55
Psammoma bodies are small, concentrically lami- a high nuclear-to-cytoplasmic ratio. Isolated cells may
nated calcifications that stain dark blue on Papanicolaou have a signet ring cell appearance. Tumor necrosis may
stains. They are commonly seen in some tumors of be present.
the ovary, fallopian tube, endometrium, and perito- Carcinomas of the bladder and urethra can also
neum, but are extremely rare in routine cervical vaginal spread to the vagina. Tumor cells are large with hyper-
smears.204 Their presence should prompt a search for a chromatic nuclei and without distinguishing features.
neoplasm, especially if they are associated with atypical Clinical correlation is needed for determining the site
cells204 (Fig. 1.69). of origin.
Carcinomas of the colon and rectum can spread Tumors from distant sites like the breast, kidney, pan-
directly to the vagina. Tumor cells frequently have a creas, and lung can metastasize to the female genital
columnar shape with large, hyperchromatic nuclei and tract. In general, precisely identifying the primary site
is impossible without the clinical history and previous
biopsy material for comparison.
PM, postmenopausal.
a
Cherkis RC, Patten SF, Andrews TJ, et al.: Significance of normal endometrial cells detected by cervical cytology. Obstet Gynecol 1988;71:242-244.
b
Gomez-Fernandez CR, Ganjei-Azar P, Capote-Dishaw J, Nadji M: Reporting normal endometrial cells in Pap smears: An outcome appraisal.
Gynecol Oncol 1999;74:381-384.
c
Gondos B, King EB: Significance of endometrial cells in cervicovaginal smears. Ann Clin Lab Sci 1977;7:486-490.
d
Ng ABP, Regan JW, Hawliczek S, Wentz B: Significance of endometrial cells in the detection of endometrial carcinoma and its precursors. Acta
Cytol 1974;18:356-361
e
Sarode VR, Rader AE, Rose PG, et al.: Significance of cytologically normal endometrial cells in cervical smears from postmenopausal women. Acta
Cytol 2001;45:153-156.
f
Yancey M, Magelssen D, Demaurez A, Lee RB: Classification of endometrial cells on cervical cytology. Obstet Gynecol 1990;76:1000-1005.
g
Zucker PK, Kasdon EJ, Feldstein ML: The validity of Pap smear parameters as predictors of endometrial pathology in menopausal women. Cancer
1985;56:2256-2263.
56 Cervical and Vaginal Cytology
Table 1.5 History of Bleeding in Postmenopausal Women with Endometrial Cells and
Biopsy-Proven Endometrial Cancer
Authors, year Cancers with endometrial History of bleeding, n (%)
cells on Pap, n
Yes No
Cherkis et al., 1988
a
20 15 5 (25%)
Zucker et al., 1985b 18 16 2 (11%)
Gomez-Fernandez et al., 1999c 6 6 0 (0%)
TOTAL 44 37 7 (16%)
a
Cherkis RC, Patten SF, Andrews TJ, et al.: Significance of normal endometrial cells detected by cervical cytology. Obstet Gynecol 1988;71:242-244.
b
Zucker PK, Kasdon EJ, Feldstein ML: The validity of Pap smear parameters as predictors of endometrial pathology in menopausal women. Cancer
1985;56:2256-2263.
c
Gomez-Fernandez CR, Ganjei-Azar P, Capote-Dishaw J, Nadji M: Reporting normal endometrial cells in Pap smears: An outcome appraisal.
Gynecol Oncol 1999;74:381-384.
Table 1.6 Meta-Analysis of Benign-Appearing Endometrial Cells in Women Over 40: Predictive
Value for Endometrial Hyperplasia and Carcinoma (Post-Bethesda 2001)
Authors, year Cases with biopsy, n Hyperplasia, n (%) Cancers, n (%) Hyperplasia or cancer, n (%)
Browne et al., 2005a 211 1 (0.5) 6 (2.8)* 7 (3.3)
Thrall et al., 2005b 159 9 (5.7) 0 9 (5.7)
Bean et al., 2006c 140 2 (1.4) 0 2 (1.4)
Kapali et al., 2007d 499 4 (0.8) 4 (0.8) 8 (1.6)
TOTAL 1099 16 (1.4) 10 (0.9) 26 (2.4)
*
Two women with cancer were premenopausal and asymptomatic.
a
Browne TJ, Genest DR, Cibas ES: The clinical significance of benign-appearing endometrial cells on a Papanicolaou test in women 40 years or
older. Am J Clin Pathol 2005;124(6):834-837.
b
Thrall MJ, Kjeldahl KS, Savik K, et al.: Significance of benign endometrial cells in Papanicolaou tests from women aged > or = 40 years. Cancer
2005;105(4):207-216.
c
Bean SM, Connolly K, Roberson J, et al.: Incidence and clinical significance of morphologically benign-appearing endometrial cells in patients age
40 years or older: The impact of the 2001 Bethesda System. Cancer 2006;108(1):39-44.
d
Kapali M, Agaram NP, Dabbs D, et al.: Routine endometrial sampling of asymptomatic premenopausal women shedding normal endometrial cells
in Papanicolaou tests is not cost effective. Cancer 2007;111(1):26-33.
A B
Figure 1.71 Mimics of endometrial cells. A, Follicular cervicitis. Lymphocytes are the same size as endometrial cells, but less tightly
clustered. B, Bare squamous cell nuclei. They are about the size of endometrial cells and sometimes aggregate. Cells that lack cyto-
plasm should not be interpreted as endometrial cells.
In the 2001 Bethesda System, this interpretation is A common mimic of endometrial cells in older
no longer categorized as an epithelial cell abnormality, women is the cluster of crushed, atrophic endocervical
and because of the small but definite risk of a significant cells. They are recognized on the basis of some resid-
endometrial lesion, neither is it categorized as NILM. ual columnar shape. Another is follicular cervicitis (Fig.
This orphan diagnosis, therefore, falls into the general 1.71A), manifested by lymphoid cell clusters. Lymphoid
categorization “Other,” a heading some laboratories sim- cells are smaller than exfoliated endometrial cells and
ply omit from the report, as in the preceding example. less tightly cohesive. Admixed larger, paler dendritic cell
Because the primary goal of the Pap test is the identifica- nuclei and tingible-body macrophages are typical of fol-
tion of squamous precursors, the explicit statement “neg- licular cervicitis. Clusters of naked squamous cell nuclei
ative for squamous intraepithelial lesion” is included. (Fig. 1.71B) are easily mistaken for endometrial cells,
This Pap diagnosis represents 0.5% to 1% of all Pap but can be identified because they have no cytoplasm.
reports.221–224 The detection of endometrial hyperpla- Naked squamous cell nuclei (often called “small blue
sia and cancer since the implementation of the 2001 cells”) are common in postmenopausal women and thus
Bethesda System is shown in Table 1.6. a frequent mimic of endometrial cells. They are seen in
21% of Pap samples from women over the age of 50, and
Differential diagnosis of their prevalence is proportional to the woman’s age.225 At
endometrial cells in women over 40: one time their presence was associated with tamoxifen,
• crushed endocervical cells a nonsteroidal estrogen used in the treatment and pre-
• follicular cervicitis vention of breast cancer, but the frequency of small blue
• small blue nuclei nuclei is no higher in these patients than in women who
are not taking tamoxifen.225
58 Cervical and Vaginal Cytology
44. Zahniser DJ: The Development of a Fully Automated System for 64. Nanda K, McCrory DC, Myers ER, et al.: Accuracy of the
the Prescreening of Cervical Smears. Nijmegen, Netherlands, Papanicolaou test in screening for and follow-up of cervical
1979. cytologic abnormalities: A systemic review. Ann Intern Med
45. Tucker JH, Shippey G: Basic performance tests on the 2000;132:810-819.
CERVIFIP linear array prescreener. Anal Quant Cytol 65. Attwood ME, Woodman CB, Luesley D, Jordan JA: Previous
1983;5:129-137. cytology in patients with invasive carcinoma of the cervix.
46. Tanaka N, Ikeda H, Ueno T, et al.: Automated cytologic Acta Cytol 1985;29:108-110.
screening system (Cybest model 4): An integrated image 66. Benoit AG, Krepart GV, Lotocki RJ: Results of prior cytologic
cytometry system. Appl Optics 1987;26:3301-3307. screening in patients with a diagnosis of Stage I carcinoma of
47. Nordin B: The Development of an Automated Prescreener the cervix. Am J Obstet Gynecol 1984;148:690-694.
for the Early Detection of Cervical Cancer: Algorithms and 67. Berkowitz RS, Ehrmann RL, Lavizzo-Mourey R, Knapp RC:
Implementation. Uppsala, 1989. Invasive cervical carcinoma in young women. Gynecol Oncol
48. Stenkvist B, Bergstrom R, Brinne U, et al.: Automatic analysis 1979;8:311-316.
of Papanicolaou smears by digital image processing. Gynecol 68. Dunn JE, Jr, Schweitzwer V: The relationship of cervical
Oncol 1987;27:1-14. cytology to the incidence of invasive cervical cancer and
49. Reinhardt ER, Blanz WE, Erhardt R, et al.: Automated mortality in Alameda Country, California. 1960-1974. Am J
classification of cytological specimens based on multistage Obstet Gynecol 1981;139:868-876.
pattern recognition. Proceedings of the 6th International 69. Fetherston WC: False-negative cytology in invasive cancer of
Conference on Pattern Recognition. Munich, 1982, the cervix. Clin Obstet Gynecol 1983;26:929-937.
pp. 153-159. 70. Gay JD, Donaldson LD, Goellner JR: False-negative results in
50. Ploem JS, van Driel-Kulker AMJ, Verwoerd NP: Leytas: cervical cytologic studies. Acta Cytol 1985;29:1043-1046.
A cytology screening system using the new modular image 71. Krane JF, Granter SR, Trask CE, et al.: Papanicolaou smear
analysis computer (MIAC) from Leitz. In Berger G, Ploem JS, sensitivity for the detection of adenocarcinoma of the cervix:
Goerttler K (eds): Clinical Cytometry and Histometry. London, a study of 49 cases. Cancer 2001;93:8-15.
Academic Press, 1987, pp. 24-35. 72. Kristensen GB, Skyggebjerg KD, Holund B, et al.: Analysis of
51. Tucker J, Stenkvist B: Whatever happened to cervical cytology cervical smears obtained within three years of the diagnosis
automation? Anal Cell Pathol 1990;2:259-266. of invasive cervical cancer. Acta Cytol 1991;35:47-50.
52. Lee JSJ, Kuan L, Seho O, et al.: A feasibility study of the AutoPap 73. Morell ND, Taylor JR, Snyder RN, et al.: False-negative
System location-guided screening. Acta Cytol 1998;42:221- cytology rates in patients in whom invasive cervical cancer
226. subsequently developed. Obstet Gynecol 1982;60:41-45.
53. Wilbur DC, Prey MU, Miller WM et al.: The AutoPap System for 74. Paterson ME, Peel KR, Joslin CA: Cervical smear histories of
primary screening in cervical cytology. Comparing the results 500 women with invasive cervical cancer in Yorkshire. BMJ
of the prospective intended-use study with routing manual 1984;289:896-898.
practice. Acta Cytol 1998;42:214-222. 75. Rylander E: Negative smears in women developing invasive
54. Cengel KA, Day SJ, Davis-Devine S, et al.: Effectiveness of the cervical cancer. Acta Obstet Gynecol Scand 1977;56:115-118.
SurePath liquid-based Pap test in automated screening and in 76. van der Graaf Y, Vooijs GP: False negative rate in cervical
detection of HSIL. Diagn Cytopathol 2003;29(5):250-255. cytology. J Clin Pathol 1987;40:438-442.
55. Troni GM, Cariaggi MP, Bulgaresi P, et al.: Reliability of sparing 77. Mount S, Harmon M, Eltabbakh G, et al.: False positive
Papanicolaou test conventional reading in cases reported as diagnosis in conventional and liquid-based cervical
No Further Review at AutoPap-assisted cytological screening: specimens. Acta Cytol 2004;48(3):363-371.
Survey of 30,658 cases with follow-up cytological screening. 78. Levine PH, Elgert PA, Mittal K: False-positive squamous cell
Cancer 2007;111(2):93-98. carcinoma in cervical smears: Cytologic-histologic correlation
56. Biscotti CV, Dawson AE, Dziura B, et al.: Assisted primary in 19 cases. Diagn Cytopathol 2003;28(1):23-27.
screening using the automated ThinPrep Imaging System. Am 79. Stoler MH, Schiffman M: Interobserver reproducibility of
J Clin Pathol 2005;123(2):281-287. cervical cytologic and histologic interpretations: realistic
57. Davey E, d’Assuncao J, Irwig L, et al.: Accuracy of reading estimates from the ASCUS-LSIL triage study. JAMA
liquid based cytology slides using the ThinPrep Imager 2001;285:1500-1505.
compared with conventional cytology: Prospective study. BMJ 80. Bethesda System. Web Atlas. Available at http://nih.techriver.
2007;335(7609):31. net/listing.php?histo=1.
58. Dawson AE: Can we change the way we screen? The ThinPrep 81. Maguire NC: Current use of the Papanicolaou class system in
Imaging System. Cancer 2004;102(6):340-344. gynecologic cytology. Diagn Cytopathol 1988;4:169-176.
59. Schledermann D, Hyldebrandt T, Ejersbo D, Hoelund B: 82. Richart RM, Barron BA: A follow-up study of patients with
Automated screening versus manual screening: A comparison cervical dysplasia. Am J Obstet Gynecol 1969;105:386-393.
of the ThinPrep imaging system and manual screening in a 83. Melnikow J, Nuovo J, Willan AR, et al.: Natural history of
time study. Diagn Cytopathol 2007;35(6):348-352. cervical squamous intraepithelial lesions: A meta-analysis.
60. Bolger N, Heffron C, Regan I, et al.: Implementation and Obstet Gynecol 1998;92(4 Pt 2):727-735.
evaluation of a new automated interactive image analysis 84. National Cancer Institute Workshop: The 1988 Bethesda
system. Acta Cytol 2006;50(5):483-491. System for reporting cervical/vaginal cytologic diagnoses.
61. Dziura B, Quinn S, Richard K: Performance of an imaging JAMA 1989;262:931-934.
system vs. manual screening in the detection of squamous 85. Solomon D, Davey D, Kurman R, et al.: The 2001 Bethesda
intraepithelial lesions of the uterine cervix. Acta Cytol System: Terminology for reporting results of cervical cytology.
2006;50(3):309-311. JAMA 2002;287:2114-2119.
62. Lozano R: Comparison of computer-assisted and manual scree 86. Solomon D, Nayar R (eds): The Bethesda System for Reporting
ning of cervical cytology. Gynecol Oncol 2007;104(1):134-138. Cervical Cytology: Definitions, Criteria, and Explanatory
63. Miller FS, Nagel LE, Kenny-Moynihan MB: Implementation of Notes, 2nd ed. New York, Springer, 2004.
the ThinPrep imaging system in a high-volume metropolitan 87. Davey DD, Neal MH, Wilbur DC, et al.: Bethesda 2001
laboratory. Diagn Cytopathol 2007;35(4):213-217. implementation and reporting rates: 2003 practices of
60 Cervical and Vaginal Cytology
participants in the College of American Pathologists 107. Gondos B, King EB: Significance of endometrial cells in
Interlaboratory Comparison Program in Cervicovaginal cervicovaginal smears. Ann Clin Lab Sci 1977;7:486-490.
Cytology. Arch Pathol Lab Med 2004;128(11):1224-1229. 108. Yancey M, Magelssen D, Demaurez A, Lee RB: Classification
88. Kurman RJ, Solomon D: The Bethesda System for Reporting of endometrial cells on cervical cytology. Obstet Gynecol
Cervical/Vaginal Cytologic Diagnoses: Definitions, Criteria, 1990;76:1000-1005.
and Explanatory Notes for Terminology and Specimen 109. de Peralta-Venturino MN, Purslow J, Kini SR: Endometrial
Adequacy. New York, Springer-Verlag, 1994. cells of the “lower uterine segment” (LUS) in cervical smears
89. Renshaw AA, Friedman MM, Rahemtulla A, et al.: Accuracy obtained by endocervical brushings: A source of potential
and reproducibility of estimating the adequacy of the diagnostic pitfall. Diagn Cytopathol 1995;12:263-271.
squamous component of cervicovaginal smears. Am J Clin 110. Lee KR: Atypical glandular cells in cervical smears from
Pathol 1999;111:38-42. women who have undergone cone biopsy: A potential
90. Geyer JW, Carrico C: Cellular constitution of Autocyte diagnostic pitfall. Acta Cytol 1993;37:705-709.
PREP cervicovaginal samples with biopsy-confirmed HSIL 111. Fiorella RM, Cheng J, Kragel PJ: Papanicolaou smears in
[Abstract]. Acta Cytol 2000;44:505. pregnancy: Positivity of exfoliated cells for human chorionic
91. Hutchinson ML, Isenstein LM, Goodman A, et al.: gonadotropin and human placental lactogen. Acta Cytol
Homogeneous sampling accounts for the increased diagnostic 1993;37:451-456.
accuracy using the ThinPrep processor. Am J Clin Pathol 112. Schnadig VJ, Davie KD, Shafer SK, et al.: The cytologist and
1994;101:215-219. bacterioses of the vaginal-endocervical area: Clues, commas
92. Bentz J, Rowe LR, Gopez EV, Marshall CJ: The unsatisfactory and confusion. Acta Cytol 1989;33:287-297.
ThinPrep Pap test: Missed opportunity for disease detection? 113. Amsel R, Totten PA, Spiegel CA, et al.: Nonspecific vaginitis.
Am J Clin Pathol 2002;117:457-463. Diagnostic criteria and microbial and epidemiologic
93. Islam S, West AM, Saboorian MH, Ashfaq R: Reprocessing associations. Am J Med 1983;74(1):14-22.
unsatisfactory ThinPrep Papanicolaou test specimens 114. Discacciati MG, Simoes JA, Amaral RG, et al.: Presence of 20%
increases sample adequacy and detection of significant or more clue cells: An accurate criterion for the diagnosis of
cervicovaginal lesions. Cancer 2004;102(2):67-73. bacterial vaginosis in Papanicolaou cervical smears. Diagn
94. Birdsong GG: Pap smear adequacy: Is our understanding Cytopathol 2006;34(4):272-276.
satisfactory … or limited? Diagn Cytopathol 2001;24:79-81. 115. Saslow D, Runowicz CD, Solomon D, et al.: American Cancer
95. Mauney M, Eide D, Sotham J: Rates of condyloma and dysplasia Society guideline for the early detection of cervical neoplasia
in Papanicolaou smears with and without endocervical cells. and cancer. CA Cancer J Clin 2002;52:342-262.
Diagn Cytopathol 1990;6:18-21. 116. Petrin D, Delgaty K, Bhatt R, Garber G: Clinical and
96. Mintzer MP, Curtis P, Resnick JC, Morrell D: The effect of the microbiological aspects of Trichomonas vaginalis. Clin
quality of Papanicolaou smears on the detection of cytologic Microbiol Rev 1998;11(2):300-317.
abnormalities. Cancer 1999;87:113-117. 117. Fiorino AS: Intrauterine contraceptive device-associated
97. Vooijs PG, Elias A, van der Graaf Y, Veling S: Relationship actinomycotic abscess and Actinomyces detection on cervical
between the diagnosis of epithelial abnormalities and the smear. Obstet Gynecol 1996;87:142-149.
composition of cervical smears. Acta Cytol 1985;29:323-328. 118. Westhoff C: IUDs and colonization or infection with
98. Mitchell H, Medley G: Differences between Papanicolaou Actinomyces. Contraception 2007;75(6 Suppl):S48-S50.
smears with correct and incorrect diagnoses. Cytopathology 119. Gideon K, Zaharopoulos P: Cytomegalovirus endocervicitis
1995;6:368-375. diagnosed by cervical smear. Diagn Cytopathol 1991;7:625-
99. O’Sullivan JP, A’Hern RP, Chapman PA, et al.: A case-control 627.
study of true-positive versus false-negative cervical smears 120. Hunt JL, Baloch Z, Judkins A, et al.: Unique cytomegolovirus
in women with cervical intraepithelial neoplasia (CIN) III. intracytoplasmic inclusions in ectocervical cells on a cervical/
Cytopathology 1998;9:155-161. endocervical smear. Diagn Cytopathol 1998;18:110-112.
100. Bos AB, van Bellegooijen M, ven den Akker-van Marle E: 121. Bernal JN, Martinez MA, Dabancens A: Evaluation of proposed
Endocervical status is not predictive of the incidence of cytomorphologic criteria for the diagnosis of Chlamydia
cervical cancer in the years after negative smears. Am J Clin trachomatis in Papanicolaou smear. Acta Cytol 1989;33:309-
Pathol 2001;115:851-855. 313.
101. Kivlahan C, Ingram E: Papanicolaou smears without endo 122. Arroyo G, Elgueta R: Squamous cell carcinoma associated with
cervical cells: Are they inadequate? Acta Cytol 1986;30:258-260. amoebic cervicitis: Report of a case. Acta Cytol 1989;33:301-
102. Mitchell H: Longitudinal analysis of histologic high- 304.
grade disease after negative cervical cytology according to 123. deMoraes-Ruehsen M, McNeill RE, Frost JK, et al.: Amebae
endocervical status. Cancer 2001;93:237-240. resembling Entamoeba gingivalis in the genital tracts of IUD
103. Weir MM, Bell DA: Transitional cell metaplasia of the cervix: users. Acta Cytol 1980;24:413-420.
A newly described entity in cervicovaginal smears. Diagn 124. Chandra M: Cervical smear with intracellular organisms from
Cytopathol 1998;18:222-226. a case of granuloma venereum (donovanosis) [Letter]. Acta
104. Jonasson JG, Wang HH, Antonioli DA, Ducatman BS: Tubal Cytol 1991;35:143-145.
metaplasia of the uterine cervix: A prevalence study in 125. Kapila K, Verma K: Intracellular bacilli in vaginal smears in a
patients with gynecologic pathologic findings. Int J Gynecol case of malacoplakia of the uterine cervix [Letter]. Acta Cytol
Pathol 1992;11:89-95. 1989;33:410-411.
105. Vooijs GP, van der Graaf Y, Vooijs MA: The presence of 126. Cibas ES, Browne TJ, Bassichis MH, Lee KR: Enlarged
endometrial cells in cervical smears in relation to the day of squamous cell nuclei in cervical cytologic specimens from
the menstrual cycle and the method of contraception. Acta perimenopausal women (“PM Cells”): A cause of ASC
Cytol 1987;31:427-433. overdiagnosis. Am J Clin Pathol 2005;124(1):58-61.
106. Chang BS, Pinkus GS, Cibas ES: Exfoliated endometrial 127. Benoit JL, Kini SR: “Arias-Stella reaction”-like changes in
cell clusters in cervical cytologic preparations are derived endocervical glandular epithelium in cervical smears during
from endometrial stroma and glands. Am J Clin Pathol pregnancy and postpartum states—A potential diagnostic
2006;125(1):77-81. pitfall. Diagn Cytopathol 1996;14:349-355.
references 61
128. Yahr LJ, Lee KR: Cytologic findings in microglandular 151. Carter JJ, Koutsky LA, Hughes JP, et al.: Comparison of human
hyperplasia of the cervix. Diagn Cytopathol 1991;7:248-251. papillomavirus types 16, 18, and 6 capsid antibody responses
129. Fornari ML: Cellular changes in the glandular epithelium of following incident infection. J Infect Dis 2000;181(6):1911-
patients using IUCD—A source of cytologic error. Acta Cytol 1919.
1974;18:341-343. 152. Lee KR, Minter LJ, Crum CP: Koilocytotic atypia in Papani
130. Gupta PK, Burroughs F, Luff RD, et al.: Epithelial atypias colaou smears. Cancer 1997;81:10-15.
associated with intrauterine contraceptive devices (IUD). 153. Sherman ME, Schiffman MH, Erozan YS, et al.: The Bethesda
Acta Cytol 1978;22:286-291. System: A proposal for reporting abnormal cervical smears
131. Tambouret R, Pitman MB, Bell DA: Benign glandular cells in based on the reproducibility of cytopathologic diagnoses.
posthysterectomy vaginal smears. Acta Cytol 1998;42:1403- Arch Pathol Lab Med 1992;116:1155-1158.
1408. 154. Atypical Squamous Cells of Undetermined Significance/Low-
132. Novotny DB, Maygarden SJ, Johnson DE, Frable WJ: Tubal Grade Squamous Intraepithelial. Lesions Triage Study (ALTS)
metaplasia: A frequent potential pitfall in the cytologic Group: Human papillomavirus testing for triage of women
diagnosis of endocervical glandular dysplasia on cervical with cytologic evidence of low-grade squamous intraepithelial
smears. Acta Cytol 1992;36:1-10. lesions: Baseline data from a randomized trial. J Natl Cancer
133. Young RH, Scully RE: Minimal-deviation endometrioid Inst 2000;92:397-402.
adenocarcinoma of the uterine cervix: A report of five cases of 155. Jones BA, Novis DA: Cervical biopsy-cytology correlation.
a distinct neoplasm that may be misinterpreted as benign. Am A College of American Pathologists Q-Probes study of
J Surg Pathol 1993;17:660-665. 22,439 correlations in 348 laboratories. Arch Pathol Lab Med
134. O’Connell F, Cibas ES: Cytologic features of ciliated 1996;120:523-531.
adenocarcinoma of the cervix: A case report. Acta Cytol 156. Wright TC, Jr, Massad LS, Dunton CJ, et al.: 2006 consensus
2005;49(2):187-910. guidelines for the management of women with cervical
135. Chang AR: Carcinoma in situ of the cervix and its malignant intraepithelial neoplasia or adenocarcinoma in situ. Am J
potential: A lesson from New Zealand. Cytopathology Obstet Gynecol 2007;197(4):340-345.
1990;1:321-328. 157. Patten SE: Diseases of the uterine cervix. In Keebler CM,
136. Koss LG, Stewart FW, Foote FW, et al.: Some histological Somrak TM (eds): The Manual of Cytotechnology, 7th ed.
aspects of behavior of epidermoid carcinoma in situ and Chicago, American Society of Clinical Pathologists 1993,
related lesions of the uterine cervix: A long-term prospective pp. 106-107.
study. Cancer 1963;16:1160-1211. 158. McGrath CM, Kurtis JD, Yu GH: Evaluation of mild-to-
137. Walboomers JMM, Jacobs MV, Manos MM, et al.: Human moderate dysplasia on cervical-endocervical (Pap) smear:
papillomavirus is a necessary cause of invasive cervical cancer A subgroup of patients who bridge LSIL and HSIL. Diagn
worldwide. J Pathol 1999;189:12-19. Cytopathol 2000;23:245-248.
138. Ayre JE: The vaginal smear: “Precancer” cell studies using 159. Faquin WC, Brown FB, Krane JF, et al.: Extensively keratinized
a modified technique. Am J Obstet Gynecol 1949;58: squamous intraepithelial lesions of the cervix are difficult to
1205-1219. grade. Am J Clin Pathol 2001;115:80-84.
139. Papanicolaou GN: Atlas of Exfoliative Cytology. Cambridge, 160. Nasser SM, Cibas ES, Crum CP, Faquin WC: The significance
Mass, Harvard University Press, 1954. of the Papanicolaou smear diagnosis of low-grade squamous
140. Koss LG, Durfee GR: Unusual patterns of squamous epithe intraepithelial lesion cannot exclude high-grade squamous
lium of the uterine cervix: Cytologic and pathologic study of intraepithelial lesion. Cancer 2003;99(5):272-276.
koilocytotic atypia. Ann N Y Acad Sci 1956;63:1235-1261. 161. Elsheikh TM, Kirkpatrick JL, Wu HH: The significance of
141. Meisels A, Fortin R: Condylomatous lesions of the cervix and “low-grade squamous intraepithelial lesion, cannot exclude
vagina: I. Cytologic pattern. Acta Cytol 1976;20:505-509. high-grade squamous intraepithelial lesion” as a distinct
142. Purola E, Savia E: Cytology of gynecologic condyloma squamous abnormality category in Papanicolaou tests.
acuminatum. Acta Cytol 1977;21:26-31. Cancer 2006;108(5):277-281.
143. Hills E, Laverty CR: Electron microscopic detection of 162. Owens CL, Moats DR, Burroughs FH, Gustafson KS: “Low-
papilloma virus particles in selected koilocytotic cells in a grade squamous intraepithelial lesion, cannot exclude high-
routine cervical smear. Acta Cytol 1979;23:53-56. grade squamous intraepithelial lesion” is a distinct cytologic
144. Menezes G, Euscher E, Schwartz B, et al.: Utility of the in situ category: Histologic outcomes and HPV prevalence. Am J Clin
detection of HPV in Pap smears diagnosed as within normal Pathol 2007;128(3):398-403.
limits. Acta Cytol 2001;45:919-926. 163. Shidham VB, Kumar N, Narayan R, Brotzman GL: Should LSIL
145. Bonfiglio TA, Stoler MH: Human papillomavirus and cancer of with ASC-H (LSIL-H) in cervical smears be an independent
the uterine cervix. Hum Pathol 1988;19:621-622. category? A study on SurePath specimens with review of
146. Dyson N, Howley PM, Munger K, Harlow E: The human literature. Cytojournal 2007;4:7.
papilloma virus-16 E7 oncoprotein is able to bind to the 164. Wang SS, Sherman ME, Hildesheim A, et al.: Cervical
retinoblastoma gene product. Science 1989;243:934-936. adenocarcinoma and squamous cell carcinoma incidence
147. Matlashewski G, Schneider J, Banks L, et al.: Human trends among white women and black women in the United
papillomavirus type 16 DNA cooperates with activated ras in States for 1976-2000. Cancer 2004;100(5):1035-1044.
transforming primary cells. EMBO J 1987;6:1741-1746. 165. Clifford GM, Smith JS, Plummer M, et al.: Human papi
148. McCance DJ, Kopan R, Fuchs E, Laimans LA: Human llomavirus types in invasive cervical cancer worldwide:
papillomavirus type 16 alters human epithelial cell differen a meta-analysis. Br J Cancer 2003;88(1):63-73.
tiation in vitro. Proc Natl Acad Sci USA 1988;85:7169-7173. 166. Tavassoli FA, Devilee P (eds): World Health Organization
149. Schiffman M, Castle PE, Jeronimo J, et al.: Human papilloma Classification of Tumours. Pathology and Genetics of Tumours
virus and cervical cancer. Lancet 2007;370(9590):890-907. of the Breast and Female Genital Organs. Lyon, France, IARC
150. Thomison J, 3rd, Thomas LK, Shroyer KR: Human papi Press, 2003.
llomavirus: Molecular and cytologic/histologic aspects 167. Rushing L, Cibas ES: The frequency of tumor diathesis in
related to cervical intraepithelial neoplasia and carcinoma. smears from women with squamous cell carcinoma of the
Hum Pathol 2008;39(2):154-166. cervix. Acta Cytol 1997;41:781-785.
62 Cervical and Vaginal Cytology
168. Wilbur DC, Maurer S, Smith NJ: Behçet’s disease in a vaginal 188. Biscotti CV, Gero MA, Toddy SM, et al.: Endocervical
smear: report of a case with cytologic features and their adenocarcinoma in situ: an analysis of cellular features. Diagn
distinction from squamous cell carcinoma. Acta Cytol Cytopathol 1997;17:326-332.
1993;37:525-530. 189. Lee KR, Manna EA, Jones MA: Comparative cytologic features
169. Dvoretsky PM, Bonfiglio TA, Patten SF, Helmkamp BF: of adenocarcinoma in situ of the uterine cervix. Acta Cytol
Pemphigus vulgaris and microinvasive squamous cell 1991;35:117-126.
carcinoma of the uterine cervix. Acta Cytol 1985;29:403-410. 190. Biscotti CV, Hart WR: Apoptotic bodies: a consistent mor
170. Long HJ, 3rd, Laack NN, Gostout BS: Prevention, diagnosis, and phologic feature of endocervical adenocarcinoma in situ. Am
treatment of cervical cancer. Mayo Clin Proc 2007;82(12):1566- J Surg Pathol 1998;22:434-439.
1574. 191. Lee KR, Minter LJ, Granter SR: Papanicolaou smear sensitivity
171. Pitman MB, Cibas ES, Powers CN, et al.: Reducing or for adenocarcinoma in situ of the cervix: A study of 34 cases.
eliminating the category of atypical squamous cells of Am J Clin Pathol 1997;107:30-35.
undetermined significance decreases the diagnostic accuracy 192. Pacey F, Ayer B, Greenberg M: The cytologic diagnosis of
of the Papanicolaou smear. Cancer 2002;96:128-134. adenocarcinoma in situ of the cervix uteri and related lesions:
172. Kinney WK, Manos MM, Hurley LB, Ransley JE: Where’s the III. Pitfalls in diagnosis. Acta Cytol 1988;32:325-330.
high-grade cervical neoplasia? The importance of minimally 193. Lee KR, Genest DR, Minter LJ, et al.: Adenocarcinoma in situ
abnormal Papanicolaou diagnoses. Obstet Gynecol in cervical smears with a small cell (endometrioid) pattern:
1998;91:973-976. Distinction from cells directly sampled from the upper
173. Nascimento AF, Cibas ES: The ASC/SIL ratio for cytopatho endocervical canal or lower segment of the endometrium. Am
logists as a quality control measure: A follow-up study. Am J J Clin Pathol 1998;109:738-742.
Clin Pathol 2007;128(4):653-656. 194. Castellsague X, Diaz M, de Sanjose S, et al.: Worldwide human
174. Juskevicius R, Zou KH, Cibas ES: An analysis of factors that papillomavirus etiology of cervical adenocarcinoma and its
influence the ASCUS/SIL ratio of pathologists. Am J Clin cofactors: implications for screening and prevention. J Natl
Pathol 2001;116:331-335. Cancer Inst 2006;98(5):303-315.
175. Collins LC, Wang HH, Abu-Jawdeh GM: Qualifiers of atypical 195. Sasagawa M, Nishino K, Honma S, et al.: Origin of
squamous cells of undetermined significance help in patient adenocarcinoma cells observed on cervical cytology. Acta
management. Mod Pathol 1996;9:677-681. Cytol 2003;47(3):410-414.
176. Genest DR, Dean B, Lee KR, et al.: Qualifying the cytologic 196. DiTomasso JP, Ramzy I, Mody DR: Glandular lesions of the
diagnosis of “atypical squamous cells of undetermined cervix: validity of cytologic criteria used to differentiate
significance” affects the predictive value of a squamous reactive changes, intraepithelial lesions, and adenocarcinoma.
intraepithelial lesion on subsequent biopsy. Arch Path Lab Acta Cytol 1996;40:1127-1135.
Med 1998;122:338-341. 197. Granter SR, Lee KL: Cytologic findings in minimal deviation
177. Lachman MF, Cavallo-Calvanese C: Quantification of adenocarcinoma (adenoma malignum) of the cervix: a report
atypical squamous cells of undetermined significance in an of seven cases. Am J Clin Pathol 1996;105:327-333.
independent laboratory: is it useful or significant? Am J Obstet 198. Ballo MS, Silverberg SG, Sidawy MK: Cytologic features of
Gynecol 1998;179:421-429. well-differentiated villoglandular adenocarcinoma of the
178. Schooland M, Sterrett GF, Knowles SAS, et al.: The “inconclusive- cervix. Acta Cytol 1996;40:536-540.
possible high-grade epithelial abnormality” category in 199. Zhou J, Tomashefski JF, Jr, Khiyami A: ThinPrep Pap tests
Papanicolaou smear reporting. Cancer 1998;84:208-217. in patients with endometrial cancer: A histo-cytological
179. Wright TC, Cox JT, Massad LS, et al.: 2001 consensus guidelines correlation. Diagn Cytopathol 2007;35(7):448-453.
for the management of women with cervical cytological 200. Thrall M, Kjeldahl K, Gulbahce HE, Pambuccian SE: Liquid-
abnormalities. JAMA 2002;287:2120-2129. based Papanicolaou test (SurePath) interpretations before
180. Symmans F, Mechanic L, MacConnell P, et al.: Correlation of histologic diagnosis of endometrial hyperplasias and
cervical cytology and human papillomavirus DNA detection in carcinomas: Study of 272 cases classified by the 2001 Bethesda
post-menopausal women. Int J Gynecol Pathol 1992;11:204-209. system. Cancer 2007;111(4):217-223.
181. Kaminski PF, Sorosky JI, Wheelock JB, Stevens CW: The 201. Wright CA, Leiman G, Burgess SM: The cytomorphology of
significance of atypical cervical cytology in an older papillary serous carcinoma of the endometrium in cervical
population. Obstet Gynecol 1989;73:13-15. smears. Cancer 1999;87:12-18.
182. Sherman ME, Castle PE, Solomon D: Cervical cytology of 202. Kuebler D, Nikrui N, Bell D: Cytologic features of endometrial
atypical squamous cells-cannot exclude high-grade squamous papillary serous carcinoma. Acta Cytol 1989;33:120-126.
intraepithelial lesion (ASC-H): characteristics and histologic 203. Todo Y, Minobe S, Okamoto K, et al.: Cytological features
outcomes. Cancer 2006;108(5):298-305. of cervical smears in serous adenocarcinoma of the
183. Boon ME, Baak JP, Kurver PJ, et al.: Adenocarcinoma in situ endometrium. Jpn J Clin Oncol 2003;33(12):636-641.
of the cervix: An underdiagnosed lesion. Cancer 1981;48:768- 204. Kern SB: Prevalence of psammoma bodies in Papanicolaou-
773. stained cervicovaginal smears. Acta Cytol 1991;35:81-88.
184. Friedell GH, McKay KD: Adenocarcinoma in situ of the 205. Wright C, Pipingas A, Grayson W, Leiman G: Pemphigus
endocervix. Cancer 1953;6(5):887-897. vulgaris of the uterine cervix revisited: Case report and review
185. Krumins I, Young Q, Pacey F, et al.: The cytologic diagnosis of the literature. Diagn Cytopathol 2000;22:304-307.
of adenocarcinoma in situ of the cervix uteri. Acta Cytol 206. Schnatz PF, Guile M, O’Sullivan DM, Sorosky JI: Clinical
1977;21(2):320-329. significance of atypical glandular cells on cervical cytology.
186. Bousfield L, Pacey F, Young Q, et al.: Expanded cytologic Obstet Gynecol 2006;107(3):701-708.
criteria for the diagnosis of adenocarcinoma in situ of the 207. Garcia MT, Acar BC, Jorda M, et al.: Use of p63 for distinction
cervix and related lesions. Acta Cytol 1980;24(4):283-296. of glandular versus squamous lesions in cervicovaginal
187. Ayer B, Pacey F, Greenberg M, Bousfield L: The cytologic specimens. Cancer 2007;111(1):54-57.
diagnosis of adenocarcinoma in situ of the cervix uteri 208. Cherkis RC, Patten SF, Dickinson JC, Dekanich AS: Significance
and related lesions: I. Adenocarcinoma in situ. Acta Cytol of atypical endometrial cells detected by cervical cytology.
1987;31:397-411. Obstet Gynecol 1987;69:786-789.
References 63
209. Stoler MH, Mills SE, Gersell DJ, Walker AN: Small cell 218. Koss LG, Durfee GR: Cytologic diagnosis of endometrial
neuroendocrine carcinoma of the cervix: A human carcinoma: Results of ten years of experience. Acta Cytol
papillomavirus type 18-associated cancer. Am J Surg Pathol 1962;6:519-531.
1991;15:28-32. 219. Nguyen TN, Bourdeau JL, Ferenczy A, Franco EL: Clinical
210. Harris NL, Scully RE: Malignant lymphoma and granulocytic significance of histiocytes in the detection of endometrial
sarcoma of the uterus and vagina: A clinicopathologic analysis adenocarcinoma and hyperplasia. Diagn Cytopathol
of 27 cases. Cancer 1984;53:2530-2545. 1998;19:89-93.
211. Gomez-Fernandez CR, Ganjei-Azar P, Capote-Dishaw J, Nadji M: 220. Tambouret R, Bell DA, Centeno BA: Significance of histiocytes
Reporting normal endometrial cells in Pap smears: An in cervical smears from peri/postmenopausal women. Diagn
outcome appraisal. Gynecol Oncol 1999;74:381-384. Cytopathol 2001;24:271-275.
212. Sarode VR, Rader AE, Rose PG, et al.: Significance of 221. Thrall MJ, Kjeldahl KS, Savik K, et al.: Significance of benign
cytologically normal endometrial cells in cervical smears endometrial cells in Papanicolaou tests from women aged
from postmenopausal women. Acta Cytol 2001;45:153-156. > or = 40 years. Cancer 2005;105(4):207-216.
213. Cherkis RC, Patten SF, Andrews TJ, et al.: Significance of 222. Kapali M, Agaram NP, Dabbs D, et al.: Routine endometrial
normal endometrial cells detected by cervical cytology. Obstet sampling of asymptomatic premenopausal women shedding
Gynecol 1988;71:242-244. normal endometrial cells in Papanicolaou tests is not cost
214. Zucker PK, Kasdon EJ, Feldstein ML: The validity of Pap effective. Cancer 2007;111(1):26-33.
smear parameters as predictors of endometrial pathology in 223. Bean SM, Connolly K, Roberson J, et al.: Incidence and
menopausal women. Cancer 1985;56:2256-2263. clinical significance of morphologically benign-appearing
215. Avis NE, McKinlay SM: The Massachusetts Women’s Health endometrial cells in patients age 40 years or older: The impact
Study: An epidemiologic investigation of the menopause. of the 2001 Bethesda System. Cancer 2006;108(1):39-44.
J Am Med Womens Assoc 1995;50(2):45-49. 224. Browne TJ, Genest DR, Cibas ES: The clinical significance of
216. Ng ABP, Regan JW, Hawliczek S, Wentz B: Significance of benign-appearing endometrial cells on a Papanicolaou test in
endometrial cells in the detection of endometrial carcinoma women 40 years or older. Am J Clin Pathol 2005;124(6):834-
and its precursors. Acta Cytol 1974;18:356-361. 837.
217. Montz FJ: Significance of “normal” endometrial cells in 225. Opjorden SL, Caudill JL, Humphrey SK, Salomao DR: Small
cervical cytology from asymptomatic postmenopausal cells in cervical-vaginal smears of patients treated with
women receiving hormone replacement therapy. Gynecol tamoxifen. Cancer 2001;93(1):23-28.
Oncol 2001;81:33-39.