Cytology, diagnostic principles and clinical correlates 3rd ed e cibas, b ducatman (saunders, 2009)

1600 John F. Kennedy Blvd.
Ste 1800
Philadelphia, PA 19103-2899

CYTOLOGY: DIAGNOSTIC PRINCIPLES AND
CLINICAL CORRELATES

ISBN: 978-1-4160-5329-3

Copyright © 2009 by Saunders, an imprint of Elsevier Inc.
Copyright © 2003 by Saunders, an imprint of Elsevier Ltd.
Copyright © 1996 by Saunders, an imprint of Elsevier Inc.
All rights reserved. No part of this publication may be reproduced or transmitted in any form
or by any means, electronic or mechanical, including photocopying, recording, or any information
­storage and retrieval system, without permission in writing from the publisher.
Permissions may be sought directly from Elsevier’s Rights Department: phone: (+1) 215 239 3804
(US) or (+44) 1865 843830 (UK); fax: (+44) 1865 853333; e-mail:
You may also complete your request on-line via the Elsevier website at evier.
com/permissions.

Notice
Knowledge and best practice in this field are constantly changing. As new research and ­experience
broaden our knowledge, changes in practice, treatment and drug therapy may become necessary or
appropriate. Readers are advised to check the most current information provided (i) on procedures
featured or (ii) by the manufacturer of each product to be administered, to verify the ­recommended
dose or formula, the method and duration of administration, and contraindications. It is the
­responsibility of the practitioner, relying on their own experience and knowledge of the patient,
to make diagnoses, to determine dosages and the best treatment for each individual patient, and
to take all appropriate safety precautions. To the fullest extent of the law, neither the Publisher
nor the Editors assume any liability for any injury and/or damage to persons or property arising

out of or related to any use of the material contained in this book.
The Publisher
Library of Congress Cataloging-in-Publication Data
Cibas, Edmund S.
  Cytology : diagnostic principles and clinical correlates / Edmund S.
Cibas, Barbara S. Ducatman. — 3rd ed.
   p. ; cm.
  Includes bibliographical references and index.
  ISBN 978-1-4160-5329-3
  1. Cytodiagnosis. I. Ducatman, Barbara S. II. Title.
  [DNLM: 1. Cytodiagnosis—methods. 2. Cytological Techniques. QY 95 C567c 2009]
  RB43.C47 2009
  616.07’582—dc22

Publishing Director: Linda Belfus
Acquisitions Editor: William Schmitt
Developmental Editor: Katie DeFrancesco
Project Manager: Bryan Hayward
Design Direction: Ellen Zanolle

Printed in China
Last digit is the print number:  9  8  7  6  5  4  3  2  1

2008027011


Dedicated to
Todd Bryant Stewart
and
Alan M. Ducatman

B978-1-4160-5329-3.00039-6

Preface to Third Edition

We hope this book will serve as a useful guide both for
the pathologist in practice and for the trainee—resident or fellow—who is looking to obtain expertise in this
subspecialty.
It has been 5 years since the publication of the second
edition of Cytology: Diagnostic Principles and Clinical
Correlates. Since then, cytology has continued to grow
and evolve as a subspecialty devoted to the diagnosis of
cellular tissue obtained by minimally invasive methods
(scraping, brushing, aspiration, etc.), and thus the need
for this updated edition. But we have retained many of
the qualities of the prior editions. As did the first two, this
edition aims to be concise yet comprehensive. We have
emphasized brevity and clarity. The text is grounded
firmly in an understanding of surgical pathology and
current diagnostic terminology. Where relevant, we have
illustrated the value of established ancillary studies (e.g.,
flow cytometry and immunohistochemistry) as well as
evolving techniques such as cytogenetics, which can be
helpful in the diagnosis of certain lymphomas, soft tissue tumors, renal neoplasms, and mesothelioma.
Although the book is multi-authored, the chapters follow a similar format: indications, sample collection and
preparation methods, recommended terminology for
reporting results, accuracy (including common ­pitfalls

that lead to false-negative and false-positive diagnoses),

a description of normal elements, and finally, a how-to
guide for the diagnosis of benign and malignant lesions,
with an emphasis on differential diagnosis. We have
retained the bulleted “capsule summaries,” particularly
for summarizing cytomorphologic features and differential diagnoses. We have continued to emphasize clinical correlation (hence the title). For example, Chapter 1
includes the recently revised algorithms of the American
Society for Colposcopy and Cervical Pathology for managing women with abnormal cervical cytologic diagnoses. Good cytologists are those who understand the
clinical implications of their interpretations.
Once again, we hope we have succeeded in conveying the beauty, strength, and challenge of cytology. With
this book we have tried to take some of the mystery out
of cytology. But mysteries remain; their solutions still
obscure. If this text inspires the reader to explore and
even solve some of them, we will consider ourselves
doubly rewarded.
Edmund S. Cibas
Barbara S. Ducatman
2008

vii


Contributors
Edmund S. Cibas, MD

Jeffrey F. Krane, MD, PhD

Associate Professor
Department of Pathology
Harvard Medical School
Director, Division of Cytopathology

Brigham and Women’s Hospital
Boston, Massachussetts

Assistant Professor
Department of Pathology
Harvard Medical School
Chief, Head and Neck Pathology Service
Brigham and Women’s Hospital
Boston, Massachusetts

Barbara S. Ducatman, MD

Xiaohua Qian, MD, PhD

Professor and Chair
Department of Pathology
Associate Dean for Faculty Services
Director, West Virginia University
National Center of Excellence
in Women’s Health
West Virginia University School of Medicine
Morgantown, West Virginia

Instructor
Department of Pathology
Harvard Medical School
Pathologist
Brigham and Women’s Hospital
Boston, Massachusetts

William C. Faquin, MD, PhD
Associate Professor
Department of Pathology
Harvard Medical School
Pathologist
Divisions of ENT Pathology
and Cytopathology
Massachusetts General Hospital
Boston, Massachusetts

Christopher A. French, MD
Assistant Professor
Department of Pathology
Harvard Medical School
Pathologist
Brigham and Women’s Hospital
Boston, Massachusetts

David W. Kindelberger, MD
Instructor
Department of Pathology
Harvard Medical School
Pathologist
Brigham and Women’s Hospital
Boston, Massachusetts

Andrew A. Renshaw, MD
Pathologist
Baptist Hospital
Miami, Florida

Jian Shen, MD, PhD
Instructor
Department of Pathology
Harvard Medical School
Pathologist
Massachusetts General Hospital
Boston, Massachusetts

Paul E. Wakely, Jr., MD
Professor
Department of Pathology
Ohio State University College of Medicine
Columbus, Ohio

Helen H. Wang, MD, DrPH
Associate Professor
Department of Pathology
Harvard Medical School
Medical Director of Cytology
Beth Israel Deaconess Medical Center
Boston, Massachusetts

ix


Acknowledgments

We owe a great debt to many individuals for their help
with this book.

To Bill Schmitt, Kathryn DeFrancesco, Michael Troy,
and Kristin Saunders at Elsevier, who shepherded this
book gently to completion: a thousand thank yous. You
exemplified the spirit of teamwork, and we enjoyed
working with all of you.
Paula Delgrosso’s administrative skills and hard work
contributed immeasurably to this edition. Edmund
Carlevale heroically converted the previously unformatted references of the prior edition into EndNote format,
saving us hours of tedious work.
We express our deep appreciation to Mr. Dennis
Padget of Padget & Associates for his help with the complexities of billing in Chapter 17. He lent us his watchful eye through several versions of that section. We relied
extensively on his Pathology Service Coding Handbook
for the information set forth in that chapter. Readers who
want more information on pathology coding questions
can contact Dennis Padget at EZPathCoding@bellsouth.
net (502/722-8873) for information about subscribing to
that comprehensive electronic text.
We thank Drs. Robert Hasserjian and Tad Wieczorek
for their expertise and helpful comments on early drafts
of the Lymph Nodes chapter.
We are grateful to Kathleen Poole and the American
Society for Colposcopy and Cervical Pathology for

­ llowing us to reproduce their clinical management
a
algorithms in Chapter 1.
Thanks also to Sandy George and Deanna Reynolds at
West Virginia University, who were invaluable in providing their assistance.
We are indebted to many members of the staff of
the Brigham and Women’s Hospital and West Virginia

University School of Medicine and Hospital—the cytotechnologists, cytopathologists, and trainees—who
inspire us with their devotion to cytopathology and
who continue to challenge us. In particular, we wish
to acknowledge Dorothy Nappi, CT (ASCP), and Grace
Goffi, CT (ASCP) (IAC), who have helped us train so
many pathology residents and fellows over the years.
Without their help we would not have our extraordinary
collections of cytology teaching cases from which so
many of the images in this book are derived.
Finally, to our friends, families, and loved ones, especially Todd Stewart and Alan Ducatman, who tolerated
the long evening and weekend hours that deprived them
(temporarily!) of a large share of our time. This book
would not exist without their love and strength.
Edmund S. Cibas
Barbara S. Ducatman

xi


1
Cervical and Vaginal Cytology
Edmund S. Cibas

THE HISTORY OF THE PAP TEST
SAMPLING AND PREPARATION METHODS
Conventional Smears
Liquid-Based Cytology
ThinPrep Pap Test
SurePath Pap Test
MonoPrep Pap Test

AUTOMATED SCREENING
Historical Overview
FocalPoint Slide Profiler
ThinPrep Imaging System
ACCURACY AND REPRODUCIBILITY
DIAGNOSTIC TERMINOLOGY AND
REPORTING SYSTEMS
THE BETHESDA SYSTEM
Specimen Adequacy
General Categorization
Interpretation and Results
THE NORMAL PAP
Squamous Cells
Endocervical Cells
Exfoliated Endometrial Cells
Abraded Endometrial Cells and Lower Uterine
Segment
Trophoblastic Cells and Decidual Cells
Inflammatory Cells
Lactobacilli
Artifacts and Contaminants
ORGANISMS AND INFECTIONS
Shift in Flora Suggestive of Bacterial
Vaginosis
Trichomonas Vaginalis
Candida
Actinomyces
Herpes Simplex
Cytomegalovirus
Chlamydia Trachomatis

Rare Infections

The 20th century witnessed a remarkable decline in the
mortality from cervical cancer in many developed countries. This achievement is directly attributable to the
implementation of the Papanicolaou (Pap) test. In the
1930s, before Pap screening was introduced, cervical

BENIGN AND REACTIVE CHANGES
Benign Squamous Changes
Benign Endocervical Changes
Repair
Radiation Changes
Cellular Changes Associated with Intrauterine
Devices
Glandular Cells Status Post Hysterectomy
Other Benign Changes
VAGINAL SPECIMENS IN “DES DAUGHTERS”
SQUAMOUS ABNORMALITIES
Squamous Intraepithelial Lesions
Grading Squamous Intraepithelial Lesions
Low-Grade Squamous Intraepithelial Lesion
High-Grade Squamous Intraepithelial Lesion
Problems in the Diagnosis of Squamous
Intraepithelial Lesions
Squamous Cell Carcinoma
Atypical Squamous Cells
Atypical Squamous Cells of Undetermined
Significance
Atypical Squamous Cells, Cannot Exclude HSIL
GLANDULAR ABNORMALITIES

Endocervical Adenocarcinoma in Situ
Adenocarcinoma
Endocervical Adenocarcinoma
Endometrial Adenocarcinoma
Differential Diagnosis of Adenocarcinoma
Atypical Glandular Cells
Atypical Endocervical Cells
Atypical Endometrial Cells
OTHER MALIGNANT NEOPLASMS
Small Cell Carcinoma
Malignant Melanoma
Malignant Lymphoma
Malignant Mixed Mesodermal Tumors
Metastatic Tumors
ENDOMETRIAL CELLS IN WOMEN OLDER
THAN 40 YEARS OF AGE

cancer was the most common cause of cancer deaths in
women in the United States.1 Today, it is not even one of
the top ten.2
The incidence of cervical cancer in the United
States is approximately 11,000 cases, with 3670 deaths.2

1


2

Cervical and Vaginal Cytology

Worldwide, however, the cervical cancer incidence (over
500,000 cases annually) and mortality rates (288,000
deaths per year) are second only to those for breast
cancer.3 Screening programs, unfortunately, are rudimentary or nonexistent in many parts of the world. Fewer than
5% of women in developing countries have ever had a Pap
test.4 In contrast, 89% of women in the United States report
having had a Pap test in the preceding 3 years.

The History of the PAP Test
The Pap test is considered by many to be the most costeffective cancer reduction program ever devised.1 Credit
for its conception and development goes to George N.
Papanicolaou, an anatomist and Greek immigrant to the
United States. In 1928 he reported that malignant cells
from the cervix can be identified in vaginal smears.5
Later, in collaboration with the gynecologist Herbert
Traut, who provided him with a large number of clinical
samples, Papanicolaou published detailed descriptions
of preinvasive cervical lesions.6,7 Pathologists and physicians initially greeted this technique with skepticism,
but by the late 1940s Papanicolaou’s observations had
been confirmed by others. The Canadian gynecologist
J. Ernest Ayre suggested taking samples directly from the
cervix with a wooden spatula rather than from the vagina
with a pipette as originally described by Papanicolaou.8
Eventually, cytologic smears were embraced as an ideal
screening test for preinvasive lesions, which, if treated,
would be prevented from developing into invasive
cancer.
The first cervical cancer screening clinics were established in the 1940s.9 The Pap test was never evaluated
in a controlled, prospective study, but several pieces of
evidence link it to the prevention of cervical cancer.

First, the mortality rate from cervical cancer fell dramatically after screening was introduced, by 72% in British
Columbia10 and 70% in Kentucky.11 Second, there was a
direct correlation between the intensity of screening and
the decrease in mortality. Among Scandinavian countries,
the death rate fell by 80% in Iceland, where screening was
greatest; in Norway, where screening was lowest, the
death rate fell by only 10%.12 A similar correlation was
observed in high and low screening regions of Scotland13
and Canada.14 In the United States, the decrease in
deaths from cervical cancer was proportional to the
screening rates in various states.15 Finally, women who
do not develop invasive cancer are more likely to have
had a Pap test than women with cancer. In a Canadian
study, the relative risk for women who had not had a
Pap test for 5 years was 2.7,16 and screening history was
a highly significant risk factor independent of other
factors such as age, income, education, sexual history,
and smoking. In Denmark, a woman’s risk of ­developing
cervical cancer decreased in proportion to the number

of negative smears she had had, by 48% with just one
negative smear, 69% with two to four negative smears,
and 100% with five or more smears.17
Screening guidelines differ around the world. Even in
the United States, the recommendations of different organizations vary in some of their details.18-20 The American
Cancer Society (ACS) recommends the following:
• Cervical cancer screening should begin approximately 3 years after a woman begins having vaginal
intercourse, but no later than 21 years of age.
• Until age 30, cervical screening should be carried
out every year with conventional Pap tests or every

2 years using liquid-based Pap tests.
• At or after age 30, a woman who has had three normal test results in a row may be screened every
2 to 3 years with a Pap test (smear or liquid-based)
or every 3 years with a Pap plus human papillomavirus (HPV) test.
• A woman 70 years of age and older who has had
three or more normal Pap test results and no abnormal results in the previous 10 years may choose to
stop cervical cancer screening.
• A woman who has had a total hysterectomy may
choose to stop cervical cancer screening. (Exceptions
are women with a history of CIN 2,3, cervical cancer,
or in utero diethylstilbestrol [DES] exposure.)
Women with a history of cervical cancer, in utero DES
exposure, and who are immunocompromised (organ
transplantation, chemotherapy, chronic corticosteroid
treatment, or positive for human immunodeficiency
virus [HIV]) may benefit from more frequent screening.19
Adherence to these guidelines is critical for cervical cancer prevention. In the United States, more than 50% of
women who develop cervical cancer have not had a Pap
test in the 3 years before their cancer diagnosis.21
The recent development of two prophylactic HPV
vaccines provides a new opportunity for cervical cancer prevention.3 Both vaccines consist of empty protein
shells called virus-like particles that are made up of the
major HPV capsid protein L1. They contain no DNA and
are not infectious. One of the vaccines, Gardasil (Merck
& Co., Inc.), is a quadrivalent vaccine against HPV
types 6, 11, 16, and 18. The other is the bivalent vaccine
Cervarix (GlaxoSmithKline) that protects against HPV 16
and 18. They have shown extraordinary efficacy in preventing type-specific histologic CIN 2,3 lesions, with no
difference in serious adverse effects compared to placebo.22 The vaccines are administered in three doses to
females ages 9 to 26 years before the initiation of sexual

­activity. Continued Pap screening will remain important
for many decades, however, because these vaccines do
not protect against 30% of cervical cancers (i.e., those
not related to HPV 16 or 18); the duration of protection
is unknown; they are not effective in treating prevalent
HPV infections; and the cost of the vaccines might limit
their use in some populations.3




Sampling and Preparation Methods

As seen in the aforementioned ACS recommendations, the combination of a Pap test plus HPV test is
included as an option for screening women 30 years of
age or older. The rationale is to combine the superior sensitivity of HPV testing with the superior specificity of the
Pap test. This recommendation is controversial because
it increases screening costs. Moreover, questions remain
regarding the ideal management of women with discrepant results (e.g., HPV test positive and Pap negative). The
search for the best screening algorithm will undoubtedly
continue, particularly as molecular diagnostic methods
become more readily available.

Sampling and Preparation
Methods
To obtain an ideal Pap specimen, the following guidelines
have been established by the Clinical and Laboratory
Standards Institute.23
Patient instructions:
• Schedule the examination 2 weeks after the first

day of the last menstrual period. (It is preferable to
avoid examination during menses because blood
may obscure significant findings.)
• Do not use vaginal medication, vaginal contraceptives, or douches for 48 hours before the
appointment.
• Intercourse is not recommended the night before
the appointment.
Specimen collection:
• Specimens should be obtained after a nonlubricated speculum (moistened only with warm water
if needed) is inserted.
• Excess mucus or other discharge should be
removed gently with ring forceps holding a folded
gauze pad.
• The sample should be obtained before the application
of acetic acid or Lugol iodine.
• An optimal sample includes cells from the ectocervix
and endocervix.

Recent studies have challenged the prohibition
against a lubricated speculum and suggest that waterbased lubricants may be acceptable.24

Conventional Smears
Conventional smears are often obtained using the combination of a spatula and brush. The spatula is used first.
Although a wooden or plastic spatula is acceptable, the

3

plastic spatula is recommended because wooden fibers
trap diagnostic material. The spatula is rotated at least
360 degrees. The sample can be smeared on one half of

a slide and spray fixed (the other half should be covered
to avoid coating it with fixative before the endocervical sample is applied). Alternatively, one may set aside
the spatula sample momentarily while the endocervical
brush sample is obtained.
After the brush is inserted in the endocervical canal,
some bristles should still be visible. If inserted too
far, there may be inadvertent sampling of the lower
uterine segment (LUS), which causes diagnostic difficulties because its epithelium resembles a high-grade
squamous intraepithelial lesion (HSIL) and adenocarcinoma in situ (AIS). The brush should be rotated gently
only one-quarter turn. A larger rotation is unnecessary
because the circumferential bristles are in contact with
the entire surface the moment the brush is inserted.
The spatula sample, if not already applied and fixed,
should be applied to the slide, then the brush sample
rolled over the slide, followed by immediate fixation. The
two samples can be placed in quick succession on two
separate halves of the slide, or the endocervical sample can be rolled directly over the spatula sample, both
covering the entire slide. Immediate fixation (within
seconds) is critical to prevent air-drying artifact, which
distorts the cells and hinders interpretation.
The broomlike brush (“broom”) has a flat array of
plastic strips contoured to conform to the cervix, with
longer strips in the middle. This design allows simultaneous sampling of the endocervix and ectocervix. The long
middle strips are inserted into the os until the shorter
outer strips bend against the ectocervix. The broom is
rotated three to five times. To transfer the material, each
side of the broom is stroked once across the slide in a
painting motion.
The cotton swab moistened with saline is no longer
recommended because its fibers trap cells, reducing the

efficiency of cell transfer onto slides.
There are two options for smear fixation. Coating
fixatives contain alcohol and polyethylene glycol and
are applied by pump sprays, by droppers from dropper
bottles, or by pouring from an individual envelope
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
as described except that, instead of smearing the cells on
a slide, the collection device is rinsed in a vial containing
a liquid fixative. In the United States, the LBC Pap test is
more common than the smear.

Liquid-Based Cytology
An important landmark in the history of the Pap
test occurred in 1996 when the U.S. Food and Drug


4

Cervical and Vaginal Cytology

Administration (FDA) approved the ThinPrep™ (Hologic,
Marlborough, Mass.) as an alternative to the conventional cervicovaginal smear. This was followed 3 years
later by approval of the AutoCyte Prep™ (now known
as SurePath™; BD TriPath, Burlington, NC). The newest
LBC is the MonoPrep™ (MonoGen, Inc., Lincolnshire,
Ill.), which was approved in 2006. LBCs were an important step in the development of automated Pap screening
devices—an improved preparation was needed to minimize cell overlap so that automated screeners would

perform better in identifying abnormal cells. But LBC
performed so well in clinical trials against conventional
smears that it found a market independent of automated screening. Although there are exceptions,25 the
great majority of peer-reviewed studies, some of them
detailed in this chapter, show an increased detection
of low-grade squamous intraepithelial lesions (LSILs)
or HSILs with LBC.26 The debate over increased disease
detection with LBC continues, however, and the studies comparing LBC to smears have come under criticism
for allegedly sacrificing methodologic purity in their
design.26 Nevertheless, LBC offers several clear advantages over conventional smears: the opportunity to prepare duplicate slides and even cell block preparations
from the residual sample;27,28 the option of “out-of-vial”
aliquoting for HPV, chlamydia, and gonorrhea testing;
an improved substrate for automated screening devices;
and a thinner cell preparation that most pathologists and
cytotechnologists find less tiring to review than smears.

ThinPrep Pap Test
The practitioner obtains the ThinPrep Pap sample with
either a broom-type device or a plastic spatula/endocervical brush combination. The sampling device is swirled
or rinsed in a methanol-based preservative solution
(PreservCyt) for transport to the cytology laboratory and
then discarded. Red blood cells are lysed by the transport medium. The vials are placed one at a time on the
ThinPrep 2000 instrument. The entire procedure (Fig.
1.1A) takes about 70 seconds per slide and results in a
thin deposit of cells in a circle 20 mm in diameter (contrast with cytospin: diameter = 6 mm). A batch-processing version (the ThinPrep 3000) is also available. It uses
the same consumables (filters and solutions) but allows
automated processing of 80 samples at one time. In most
cases, only a fraction of the sample is used to prepare the
slide used for diagnosis. If needed, the residual sample is
available for additional ThinPrep slide preparation, cell

block preparation, or molecular diagnostic testing (e.g.,
high risk HPV, chlamydia, gonorrhea).
A multicenter, split-sample study found that the
ThinPrep detected 18% more cases of LSILs and more
serious lesions as compared to conventional smears,
with no significant difference in the detection of organisms.29 A number of studies have shown significant

increases in HSIL detection after the implementation of
the ThinPrep.30–35
The ThinPrep is equivalent to the conventional
smear in the detection of endocervical AIS.36 Data also
show comparable results between ThinPrep slides and
conventional smears for the detection of endometrial
pathology.37
The ThinPrep collection vial has been approved by
the FDA for direct testing for HPV, which is particularly
useful for managing women whose Pap tests show atypical squamous cells (ASC).38,39

SurePath Pap Test
TriPath Imaging (acquired by Becton Dickinson in 2006)
developed the SurePath Pap test (formerly AutoCyte Prep
and CytoRich) for samples collected in an ethanol-based
transport medium. The process is shown in Figure 1.1B.
In contrast to the ThinPrep and MonoPrep methods, the
practitioner snips off the tip of the collection device and
includes it in the sample vial. The equipment to prepare slides includes a Hettich centrifuge and a PrepStain
robotic sample processer with computer and monitor.
The PrepMate™ is an optional accessory that automates
mixing the sample and dispensing it onto the density
reagent. Red blood cells and some leukocytes are eliminated by density centrifugation. In addition to preparing

an evenly distributed deposit of cells in a circle 13 mm in
diameter, the method incorporates a final staining step
that discretely stains each individual slide.
A multicenter, split-sample clinical trial showed a
7.2% increase in the detection of LSILs and more serious
lesions and a significant decrease in the percentage of
unsatisfactory specimens.40

MonoPrep Pap Test
The practitioner obtains the MonoPrep sample with
standard collection devices that are swirled or rinsed
in a preservative-filled collection vial, after which the
sampling device is discarded. As with the ThinPrep, red
blood cells are lysed by the transport medium. The vials
are delivered to the laboratory where slides are prepared
using the MonoPrep Processor, a fully automated, batchprocessing instrument capable of processing 40 samples
per hour, with a throughput capacity of 324 samples per
8-hour run. The process is shown in Figure 1.1C. In a
split-sample clinical trial similar in design to the ThinPrep
and SurePath trials, slides prepared by the MonoPrep
method showed a 26% increase in the detection of LSILs
and more serious lesions, with no significant difference
in relative specificity.41 MonoPrep provided a significant reduction in unsatisfactory slides, and there was no
difference in the presentation of endocervical or transformation zone component or the detection of benign
conditions.




Sampling and Preparation Methods

1. Dispersion

2. Cell collection

3. Cell transfer

2
Disaggregation

3
Transfer
to sedimentation
tube

5

A

1
Vortexing

B

1
High-speed
mixing

2
Turbidity

check

4
Sedimentation
2

3
Aspiration

5
Cell deposition
and staining

4
Cell deposition

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 technology 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

Automated Screening
Historical Overview
Automated cytology screening devices have been
under development since the 1950s. The first computerized screening system was developed in the United
States by Airborne Instruments Inc., and was called the
Cytoanalyzer.42 In preclinical trials it did not perform as
well as expected and the project was discontinued. The
difficulty of the task was soon appreciated, especially
the inherent problems with analyzing smears prepared
in the conventional manner. Despite setbacks, research
into cervical cytology screening continued, especially
in Europe and Japan, throughout the 1970s and 1980s,
with the development of the Quantimet,43 BIOPEPR,44
CERVIFIP,45 CYBEST,46 DIASCANNER,47,48 FAZYTAN,49
and LEYTAS.50 Some of these instruments are now in
museums, but others have served as prototypes for
systems that are commercially available or still under
development.
Although European investigators largely lost interest in cytology automation in the 1990s,51 researchers in
the United States and Canada, having established private enterprises supported by venture capital, retained
their enthusiasm. Foremost in the field have been AutoCyte
(formerly Roche Image Analysis Systems), Cytyc, Neopath,
and Neuromedical Systems. An important three-way
merger took place in 1999, when AutoCyte, after purchasing the intellectual property of Neuromedical Systems,
merged with Neopath to form a new company called

TriPath Imaging. In 2007, Cytyc Corporation, developer
of the ThinPrep Pap Test and ThinPrep Imaging System,
merged with Hologic Inc., and became a wholly-owned
subsidiary of Hologic.
In 1998, the FDA approved the AutoPap System
(now called the FocalPoint Slide Profiler™; BD TriPath
Imaging, Burlington, NC) as a primary screener for cervicovaginal smears, followed by approval in 2002 for
use with SurePath slides. In 2003, the FDA approved
the ThinPrep Imaging System™ as a primary screener
for ThinPrep Pap slides. Thus, these two automated
screening devices are designed for different preparation
methods. Although both rely on image analysis technology, there are also fundamental differences in the
way they integrate into the workflow of the laboratory.
Neither is approved for use for nongynecologic cytology
specimens.

FocalPoint Slide Profiler
The FocalPoint Slide Profiler (FPSP) is a self-contained
instrument that classifies Pap slides without human
intervention (Fig. 1.2A). It uses algorithms to measure
cellular features like nuclear size, integrated optical density, nuclear to cytoplasmic ratio, and nuclear contour—

morphologic features that pathologist Stanley Patten
established using planimetry and ocular micrometry for
the diagnosis of squamous and glandular lesions.52
AutoPap, the predecessor of FPSP, was originally
intended as a primary screening device that would eliminate the need to manually screen as many as one half of
all smears. It was temporarily redesigned as a quality control rescreening device called the AutoPap 300 QC System
and obtained FDA approval for this function in 1995. The
AutoPap 300 QC System did not find a wide audience,

however, and became obsolete in the year 2000. A redesign resulted in a new instrument (the AutoPap SystemPrimary Screener, later renamed FPSP) which obtained
FDA approval as a primary screening device in 1998.
In this mode, the device is used in the initial screening
of smears. It identifies approximately 25% of slides as
requiring “no further review.” Of the remaining slides that
require manual review, it also identifies at least 15% for a
second manual review, which may be used as a substitute
for the 10% review of negative Pap samples required of all
U.S laboratories (see Chapter 17).
A barcode is applied to each slide and slides are loaded
into slide trays. Up to 288 slides can be loaded at a time
(8 slides per tray, 36 trays). Each slide is analyzed using
preset algorithms at ×4 magnification for a visual map
of the entire slide, then 1000 fields are captured at ×20
magnification. After analysis, the device assigns a score
(from 0 to 1.0) to each slide according to the likelihood
of an abnormality. Slides below a cutoff are considered
no further review, and those above the cutoff are triaged
for full manual review. Any slide deemed unsuitable for
analysis because of preparation or cover slipping problems requires manual review.
The accuracy of the instrument was evaluated in a clinical trial at five laboratories.53 Each slide was first evaluated in the conventional manner. The same slides were
then processed by the AutoPap System, which detected
significantly more abnormal slides—atypical squamous
cells of undetermined significance (ASC-US) or greater—
than conventional practice (86% versus 79%).
Importantly, FPSP is not approved for women at high
risk for cervical cancer. Thus, a laboratory that uses FPSP
for primary screening must set aside all Paps from highrisk women for manual screening. It is up to the laboratory to define what constitutes a Pap from a high-risk
patient.
False-negative results are occasionally encountered

with the FPSP. In the clinical trial, there were 10 falsenegatives (5 ASC-US, 4 LSIL, and 1 HSIL) in the 1182
cases considered no further review by FPSP, and Cengel
and colleagues found 9 false-negatives (5 ASC-US and 4
LSIL) in the 296 cases considered no further review by
FPSP.54
The productivity gain with FPSP is modest, because
in practice the FPSP archives only about 16% to 17% of
Paps without full manual review.53,55




Automated Screening

7

A

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 manual 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.

ThinPrep Imaging System
The ThinPrep Imaging System (TIS) uses location-guided
screening to aid the cytotechnologist in reviewing a
ThinPrep Pap slide. The TIS consists of two components,
the image processor (“imager”) and the review scope
(RS; Fig. 1.2B). Stained and cover slipped ThinPrep slides

are placed in a cartridge (each cartridge holds 25 slides),
and up to 10 cartridges are loaded onto the benchtop
imager. The imager has the capacity to screen over 300
slides per day. It scans the slides and identifies 22 fields
of view (FOV) on each slide based on optical density
measurements and other features. The x and y coordinates of the 22 FOV are stored in a database and retrieved
at a later time. The server is electronically linked to one
or more RSs in the laboratory. An RS resembles a standard microscope but is augmented with an automated
stage, a pod that controls the stage and objectives, and a
keypad. The scope also has a camera that reads the slide
identifier when the slide is loaded onto the stage. When
a valid slide identifier is recognized, the server sends
its coordinate information to the scope, permitting the

c­ ytotechnologist to navigate to the 22 FOV using the
pod. Navigation to each FOV is done geographically, that
is, using the shortest distance from one FOV to the next.
The cytotechnologist uses the pod to advance forward
or return back through the FOV, changing objectives as
needed. If no abnormal cells are found in any of the FOV,
the case has been completed and can be reported as negative. If any abnormal cells are found in any of the FOV, a
review of the entire slide must be performed. This can be
done using the autoscan function on the RS, with preset,
customized user screening preferences. The RS has both
electronic and physical slide dotting capabilities.
The accuracy of the TIS was evaluated in a clinical trial
at four laboratories. ThinPrep slides were first screened
manually and the results recorded. They were then
rescreened using the TIS. Truth adjudication was performed by expert review of all abnormal cases and a proportion of negative slides. The TIS detected significantly
more abnormal slides (ASC-US or greater) than manual

review (82% versus 76%).56 A later split-sample study
comparing conventional smear cytology versus the TIS


8

Cervical and Vaginal Cytology

for ThinPrep slides showed a significantly higher detection rate of histologic HSIL (CIN 2,3) with the TIS.57
Because 22 FOV represent approximately 25% of the
ThinPrep cell spot,58 implementation of the TIS comes
with a significant productivity enhancement, and in
some laboratories the productivity of cytotechnologists
has as much as doubled.56,59,60
Implementing the TIS requires adopting the proprietary ThinPrep Pap stain, to which some adjustment is
necessary because it yields darker nuclear staining of
metaplastic and endocervical cell clusters than most
traditional Pap stains. Like FPSP, TIS does not eliminate
false-negatives, which are still encountered, albeit less
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
of the TIS.61–63

Accuracy and
Reproducibility
The sensitivity of cytology for detecting preinvasive
squamous and glandular lesions is difficult to establish,
but it is clearly far from perfect. Most studies of preinvasive lesions suffer from verification bias (i.e., cases are
referred for biopsy on the basis of an abnormal smear,

and women with negative Pap tests are not biopsied).
The few relatively unbiased studies show that the mean
sensitivity of the Pap test is 47% (range 30% to 80%), and
the mean specificity is 95% (range 86% to 100%).64
The sensitivity of cytology is less than ideal for invasive cancers as well, and estimates range widely (16% to
82%). Many women with cervical cancer have a history
of one or more negative smears.65–76 The relative contributions of sampling and laboratory error vary from one
study to another and likely depend on how carefully
retrospective rescreening is performed.
False-positive diagnoses of cervical cancer occur in
10% to 15% of cases.77,78 The chief culprits are the atrophic smear with benign squamous atypia in a granular,
pseudonecrotic background; reparative changes; and
keratinizing HSILs.
The interobserver reproducibility of cytologic interpretations is less than perfect. In a large study of women,
most of whom had mild cytologic abnormalities, the
unweighted κ statistic for four categories of diagnosis—
negative, atypical, LSIL, and HSIL—was 0.46, indicating
moderate reproducibility.79 (Roughly, a κ of 0 or less represents poor agreement, 0 to 0.2 slight agreement, 0.2 to
0.4 fair agreement, 0.4 to 0.6 moderate agreement, 0.6 to
0.8 very good agreement, and 0.8 to 1.0 almost perfect
agreement.) In the same study, the reproducibility of histologic interpretations of cervical biopsies, also for four
categories of diagnosis, was identical (0.46). The ­greatest
disagreement with Pap tests involved those originally

interpreted as showing ASC-US; the second reviewer
agreed with only 43% of cases. The greatest disagreement with biopsies involved those originally interpreted
as LSIL; the second reviewer concurred in only 43% of
cases.79
A graphic demonstration of the relative reproducibility of various cytologic findings is available on the
Bethesda System Web Atlas, which contains the results

of the Bethesda Interobserver Reproducibility Project.
A large number of images were reviewed by hundreds
of observers, who were asked to place the images into
one of the Bethesda System categories. The results are
displayed for each image as a histogram.80

Diagnostic Terminology
and Reporting Systems
Papanicolaou devised a numerical system for reporting
cervical smears, which was originally intended to convey
his degree of suspicion that the patient had cancer: class I,
absence of atypical or abnormal cells; class II, atypical but no evidence of malignancy; class III, suggestive
of but not conclusive for malignancy; class IV, strongly
suggestive of malignancy; and class V, conclusive for
malignancy. Over time, however, the Papanicolaou class
system underwent many modifications and was not
used in a uniform fashion.81 It persisted in many laboratories well into the 1980s, however. In other laboratories it
was replaced (or supplemented) by descriptive terms
borrowed from histologic classifications of squamous
lesions. Squamous cancer precursors were originally
divided into carcinoma in situ, which was a high-risk
lesion of immature, undifferentiated atypical cells, and
dysplasia (subdivided into mild, moderate, and severe),
considered to be a low-risk lesion composed of more
mature squamous cells. In the 1960s, Richart challenged
the duality of dysplasia/carcinoma in situ and proposed
a new term, cervical intraepithelial neoplasia (CIN).
CIN was graded from 1 to 3, but Richart believed that
CIN 1 (mild dysplasia) had a strong propensity to progress to CIN 3 and cancer. The high rate of progression
found in his study most likely related to stringent entry

criteria; for inclusion, CIN 1 had to be confirmed on three
consecutive Paps.82 Richart’s data showed a higher progression rate for mild dysplasia than most other natural
history studies.83 The CIN concept was highly influential,
however, and for many years squamous precursors were
treated as much on the basis of their size and location as
on their grade. This situation remained for two decades.
In 1989 the Bethesda System was introduced to standardize the reporting of cervical cytology results and
incorporate new insights gained from the discovery of
HPV.84 The name for a squamous cancer precursor was
changed to squamous intraepithelial lesion (SIL), subdivided into only two grades (low and high) based on the




evolving understanding of the biology of HPV. In this system, LSIL encompasses CIN 1, and HSIL encompasses
CIN 2 and 3. This was a shift away from the CIN concept,
one based on a reevaluation of the existing evidence,
which demonstrated that most LSILs are, in fact, transient HPV infections that carry little risk for oncogenesis,
whereas most HSILs are associated with viral persistence
and a significant potential for progression to invasive
cancer.
The first Bethesda System workshop in 1988 was followed by two others in 1991 and 2001, which made modifications to the original framework and terminology. The
2001 workshop broadened participation by using a dedicated Web site on the Internet, and an electronic bulletin board received more than 1000 comments regarding
draft recommendations. The 2001 Bethesda System, like
its predecessors, recommends a specific format for the
cytology report, starting with an explicit statement on
the adequacy of the specimen, followed by a general categorization and an interpretation or result.85,86

The Bethesda System
Specimen Adequacy

One of the most important advances of the Bethesda
System is its recommendation that each Pap report begin
with a statement of adequacy. In 1988, the Bethesda
System proposed three categories for specimen adequacy: “satisfactory,” “less than optimal” (renamed
“satisfactory but limited by … ” in 1991), and “unsatisfactory.” The 2001 Bethesda System eliminated the middle category because it was confusing to physicians and
prompted unnecessary repeat Pap tests. Nevertheless, the
2001 Bethesda System advocates mentioning the presence or absence of a transformation zone ­component
and permits comments on obscuring elements. The 2001
Bethesda System criteria for adequacy are listed in Table 1.1.
They are somewhat arbitrary, because scientific data
on adequacy are limited, particularly regarding the
minimum number of cells needed for an adequate sample.
It is easy to determine whether a specimen is adequate
or unsatisfactory in most cases. Slides received without
patient identification or broken beyond repair should
be rejected as unsatisfactory. An appropriately labeled
smear with an adequate complement of well-preserved
squamous and endocervical cells is clearly satisfactory.
On average, about 0.5% of Pap samples are interpreted
as unsatisfactory.87 Unsatisfactory Pap samples can be
finalized by a cytotechnologist and need not be reviewed
by a cytopathologist (see Chapter 17).
One of the components of an adequate smear is an
adequate squamous component. In the 1988 and 1991
Bethesda Systems, the requirement for an adequate
squamous component was defined as “well-preserved

The Bethesda System

9

Table 1.1  The 2001 Bethesda System Categories
for Specimen Adequacy
Satisfactory for Evaluation
A satisfactory squamous component must be present (see
text).
Note the presence or absence of endocervical or
transformation zone component.
Obscuring elements (inflammation, blood, drying artifact,
other) may be mentioned if 50% to 75% of epithelial cells
are obscured.
Unsatisfactory for Evaluation
Specimen rejected or not processed because (specify
reason). Reasons may include:
• lack of patient identification.
• unacceptable specimen (e.g., slide broken beyond
repair).
or:
Specimen processed and examined, but unsatisfactory for
evaluation of an epithelial abnormality because (specify
reason). Reasons may include:
• insufficient squamous component (see text).
• obscuring elements cover more than 75% of epithelial
cells

and well-visualized squamous epithelial cells should
cover more than 10% of the slide surface.”88 This guideline, however, was interpreted differently by different
cytologists. Even in laboratories that interpreted it literally, observers consistently overestimated the percentage of slide coverage by squamous cells.89 With the 2001
Bethesda System modification, the requirement was
redefined as a minimum “estimated number of squamous cells,” the minimum being different for conventional versus liquid-based preparations.

The minimum number of squamous cells for
adequacy depends on the preparation method:
• liquid-based: 5000
• conventional: 8000 to 12,000
The minimum number of 5000 squamous cells for
an adequate LBC Pap was based on correlations made
between the false-negative rate and squamous cell cellularity.90 Because LBCs likely represent a more homogeneous representation of the material obtained by the
collection device,91 a more stringent squamous ­cellularity
requirement was imposed on conventional smears.
Whether or not a slide contains an adequate squamous
cell component is immediately apparent in most cases. In
borderline cases, techniques are available for estimating
adequacy: reference images for conventional smears and
a spot-counting procedure for liquid-based preparations.
Reference images of known cell counts are useful for estimating cellularity.89 Because of this, the 2001 Bethesda
System published images to assist in the estimation of
squamous cellularity on conventional smears.86


10

Cervical and Vaginal Cytology

A spot-counting method is used to evaluate LBCs
with borderline squamous cellularity. A minimum of
10 fields are counted along a diameter that includes the
center of the slide (Fig. 1.3A). If the cell circle has blank
spots, these should be represented in the fields counted
(Fig. 1.3B). The average number of squamous cells is
then compared against tables that take into account the

objective, the eyepiece field number, and the diameter
of the circle that contains cellular material.86 For example, with an FN20 eyepiece, and a ×40 objective, the sample is adequate if the average number of cells counted is
greater than 3.1 for a ThinPrep slide.
Additional slides can usually be generated from the
residual vial of an LBC sample. In some laboratories, an
additional slide is prepared when the initial slide has
insufficient cellularity. The addition of a washing step
with 10% glacial acetic acid increases the percentage of
satisfactory ThinPrep Pap samples, uncovering occasional cases of SIL and invasive cancer.92,93
The cellularity of the squamous cell component
is estimated; laboratories are not expected to count
individual cells. Squamous cellularity is sometimes
particularly difficult to estimate, for example, when
there is marked cell clustering or cytolysis. In certain
­clinical settings, particularly in women with atrophy,
a lower number may be adequate. In these situations,
cytologists are expected to use their judgment when
­evaluating adequacy.86
In the 2001 Bethesda System, the presence or absence
of an endocervical or transformation zone component is
noted on the report. An endocervical component is considered present if 10 or more endocervical or squamous

A

metaplastic cells, either isolated or in groups, are ­present.
The data on the endocervical component as a measure
of adequacy are contradictory.94 The importance of
endocervical cells was first suggested by cross-sectional
­studies, which showed that smears are more likely to
contain SIL when endocervical cells are ­present.95–97 Data

from retrospective case-control studies, however, do not
support this; investigators have found no association
between false-negative Pap samples and the absence of
endocervical cells.98,99 Retrospective cohort studies have
shown that women whose initial smears lack endocervical cells do not develop more lesions on follow-up than
women whose smears do have an endocervical component,100–102 implying that an ­endocervical component
is not essential. Currently, a smear without endocervical cells is not considered unsatisfactory, although the
absence of an endocervical or ­transformation zone
component is mentioned as a “quality ­indicator.” This
is not to imply that a repeat Pap is necessary. Physicians
are expected to use their judgment and to consider
repeating the Pap if the patient is at high risk for ­cervical
cancer.

General Categorization
The general categorization is an optional component of
the 2001 Bethesda System.
Three categories:
• negative for intraepithelial lesion or malignancy
• epithelial cell abnormality
• other
The 1991 Bethesda categories “within normal limits”
and “benign cellular changes” were combined into a single “negative” category in 2001. “Other” includes cases
that do not fit neatly into one of the other two categories:
non-epithelial malignancies, such as melanoma and
lymphoma, and benign-appearing endometrial cells in
women over 40 years of age.
Specimens are categorized according to the most significant abnormality identified.

Interpretation and Results

B
Figure 1.3  Method for estimating the adequacy of the squamous component of liquid-based preparations. A, At ×40, 10
fields are counted starting at the edge (horizontal or vertical)
and including the center of the preparation. B, An attempt is
made to include “holes” in proportion to their size, making sure
that the fields counted cover both cellular and sparsely cellular
areas in proportion to their size.

Recommended terminology for reporting findings is
listed in Table 1.2.
Non-neoplastic findings, other than organisms, are
optional, given that many physicians desire the Pap test
report to be as concise as possible. Findings of no clinical consequence, if mentioned, may result in confusion and even unnecessary repeat testing. Nevertheless,
many cytologists believe it is important to document
that certain findings were interpreted as benign, particularly those that can mimic a neoplasm.




The Normal PAP

Table 1.2  The 2001 Bethesda System
Specimen Adequacy (see Table 1.1)
General Categorization (Optional)
Negative for intraepithelial lesion or malignancy (NILM)
Epithelial cell abnormality
Other
Interpretation/results
NILM

  Organisms
  Trichomonas vaginalis
  Fungal organisms morphologically consistent with
  Candida species
  Shift in flora suggestive of bacterial vaginosis
  Bacteria morphologically consistent with Actinomyces
  species
  Cellular changes consistent with herpes
simplex virus
Other non-neoplastic findings (optional to report; list not
comprehensive)
  Reactive cellular changes associated with: inflammation
  (includes typical repair); radiation; intrauterine
  contraceptive device (IUD)
  Glandular cells status post hysterectomy
  Atrophy
Epithelial cell abnormalities
  Squamous cell
   Atypical squamous cells (ASC)
   - of undetermined significance (ASC-US)
   - cannot exclude HSIL (ASC-H)
   Low-grade squamous intraepithelial lesion (LSIL)
   High-grade squamous intraepithelial lesion (HSIL)
   Squamous cell carcinoma (SQC)
  Glandular cell
   Atypical glandular cells (AGC); specify endocervical,
   endometrial, or not otherwise specified
   AGC, favor neoplastic (specify endocervical or not
   otherwise specified)
  Endocervical adenocarcinoma in situ (AIS)

  Adenocarcinoma
Other
  Endometrial cells in a woman older than 40 years
  of age

11

In the United States, a pathologist is required to review
cases that show reactive or reparative changes and any
abnormality at the level of ASC-US or higher. This represents about 10% to 20% of the total Pap volume in
most laboratories.

Squamous Cells
The ectocervix is lined by a stratified squamous epithelium that matures under the influence of estrogen.
The most mature squamous cells are called superficial
cells. They have a small, pyknotic nucleus that is 5 to 6
μm in diameter. Intermediate cells have a larger nucleus
measuring 8 μm in diameter, which is not pyknotic but
instead has a finely granular texture. Intermediate cells
are occasionally binucleated and even multinucleated.
Both superficial and intermediate cells are large poly­
gonal cells with transparent pink or green cytoplasm
(Fig. 1.4). Superficial and intermediate cells are the predominant cells in cytologic samples from women of
reproductive age.
Immature squamous cells are called parabasal cells
and basal cells. Because a Pap test does not usually
scrape off the entire thickness of the epithelium but only
the upper few layers, immature cells near the base of a
mature epithelium are not sampled. An immature epithelium, however, is composed throughout its thickness
by parabasal-type cells or basal-type cells. Immature

epithelium is common at the transformation zone,
where it is called squamous metaplasia, and whenever
there is squamous epithelial atrophy as a result of a low
estrogen state. Thus, parabasal and basal cells are typically obtained from squamous metaplasia or atrophic
epithelium.
Squamous atrophy is encountered in a variety of
clinical settings associated with a low estrogen state.

Automated Review and Ancillary Testing (Include as
Appropriate)
Educational Notes and Suggestions (Optional)

The Normal PAP
A normal Pap test result begins with a statement
of adequacy, followed by “negative for intraepithelial lesion or malignancy” (NILM). Additional findings (e.g., reactive changes, infectious organisms) are
listed subsequently. Approximately 91% of Pap tests
are interpreted as such.87 Normal Pap tests, with the
exception of those cases that show reactive or reparative changes, can be finalized by a cytotechnologist and
need not be reviewed by a ­pathologist (see Chapter 17).

Figure 1.4  Superficial and intermediate squamous cells.
The mature squamous epithelium of the ectocervix in women of
reproductive age is composed throughout most of its thickness
by superficial (arrowhead) and intermediate (arrow) cells.


12

Cervical and Vaginal Cytology

Low estrogen states include:
• premenarche
• postpartum
• postmenopause
• Turner syndrome
• status post bilateral oophorectomy

Immature, parabasal cells are round or oval rather
than polygonal and have a variably sized nucleus that is
usually larger than that of an intermediate cell. Basal cells
are even smaller and have scant cytoplasm (Fig. 1.5).
Basal and parabasal cells are the hallmark of atrophy.
With a deeply atrophic cervical epithelium, no superfi-

Figure 1.5  Parabasal and basal cells (postpartum smear).
Parabasal cells (large arrow) are oval and typically have dense
cytoplasm. Basal cells (small arrow) are similar but have less cytoplasm. Many cells have abundant pale-yellow staining glycogen,
a characteristic but nonspecific feature of squamous cells of
pregnancy and the postpartum period.

A

cial or intermediate cells are seen, only parabasal and
basal cells. In addition, atrophic epithelium, particularly in postmenopausal women, is prone to injury and
inflammation and often shows a spectrum of changes
that must be recognized as normal and not confused
with a significant lesion. The sheets of immature cells are
crowded and syncytium-like, mimicking the crowded
cells of an HSIL (Fig. 1.6). Nevertheless, the chromatin texture in atrophy is finely granular and evenly distributed, and nuclear contours remain mostly smooth
and thin. A curious variant, transitional cell metaplasia,

is notable for prominent longitudinal nuclear grooves
(“coffee-bean nuclei”), wrinkled nuclei, and small perinuclear halos (Fig. 1.6B).103 Cellular degeneration is
seen in some cases of atrophy (Fig. 1.7A). Air-drying, a
common artifact with smears, causes artificial nuclear
enlargement. Dark blue, rounded, amorphous masses
known as “blue blobs,” thought to represent either condensed mucus or degenerated bare nuclei, are sometimes seen (Fig. 1.7B), as is a granular background (see
Fig. 1.7A) that resembles the necrosis associated with
invasive cancers.
Parabasal cells are also the constituents of squamous
metaplasia of the endocervix. Squamous metaplasia
is a common morphologic alteration of the endocervical epithelium usually limited to the transformation
zone in women who otherwise have good squamous
­maturation. It is identified on smears as flat sheets
of immature squamous cells (parabasal cells), often
arranged in an interlocking fashion like paving stones
(Fig. 1.8). The parabasal cells may show mild variation
in nuclear size, with slightly irregular contours and slight
hyperchromasia.
Squamous metaplasia, as defined cytologically, is
always composed of parabasal cells (immature squamous cells). So-called mature squamous metaplasia, 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 resembles 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

A

13

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
overlying endocervical glands, is not recognized as such
on cytologic preparations.
Other normal changes of squamous cells are hyperkeratosis and parakeratosis. Hyperkeratosis is a benign
response of stratified squamous epithelium as a result
of chronic mucosal irritation, as in uterine prolapse.
Anucleate, mature, polygonal squamous cells appear
as isolated cells or plaques of tightly adherent cells (Fig.
1.9A). Such cells are benign and should not be considered abnormal. This cytologic picture is mimicked by
contamination of the slide by squamous cells of the
vulva or skin from the fingers of the persons handling
the slide.

Parakeratosis, a benign reactive change also caused
by chronic irritation, is characterized by small, heavily

keratinized squamous cells with dense orangeophilic
cytoplasm and small, pyknotic nuclei (Fig. 1.9B). When
such densely keratinized cells show nuclear atypia in the
form of enlargement and membrane irregularity with
hyperchromasia, they are called “dyskeratocytes” or
“atypical parakeratosis” and should be categorized as an
epithelial cell abnormality.

Endocervical Cells
The endocervix is lined by a mucin-producing ­columnar
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.

granular chromatin texture and abundant vacuolated
cytoplasm. Nucleoli are inconspicuous but become
quite prominent in reactive conditions, such as cervicitis
(see section on reactive changes). Endocervical cells are
often identified in strips or sheets rather than as isolated
cells (Fig. 1.10). When arranged as strips, the cells have

the appearance of a picket fence. When in sheets, they
resemble a honeycomb because of the well-defined cell
borders and uniform cell arrangement. Rarely, mitoses
are identified. They should not raise suspicion of a neoplasm if the cells are otherwise normal in appearance.
Tubal metaplasia is a benign alteration of the endocervical epithelium found in about 30% of cone biopsy and
hysterectomy specimens (Fig. 1.11).104

Exfoliated Endometrial Cells
Spontaneously exfoliated, menstrual endometrial cells
are seen if the Pap is taken during the first 12 days of the
menstrual cycle.105

A

Cytomorphology of exfoliated
endometrial cells:
• balls of small cells
• isolated small cells
• scant cytoplasm
• dark nucleus
• nuclear molding
• nuclear fragmentation

Exfoliated endometrial cells are most easily ­recognized
when they are arranged in spherical clusters (Fig. 1.12).
They are small cells with a dark nucleus and (usually)
scant cytoplasm. Occasional cells may have more abundant clear cytoplasm. Clusters have a scalloped contour
as a result of the slight protrusion of individual cells.
Apoptosis is common. Isolated endometrial cells are
also seen, but they are less conspicuous because of their

small size.

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.

In a woman 40 years of age or older, benign-appearing
endometrial cells are reported because of the small associated risk of endometrial neoplasia.
Differential diagnosis of exfoliated
endometrial cells:
• HSIL
• squamous cell carcinoma
• AIS
• small cell carcinoma
Figure 1.12  Endometrial cells. Spontaneously exfoliated
endometrial cells, as in menses, are small cells arranged in balls.
Cytoplasm is scant. Nuclei around the perimeter appear to be
wrapping around adjacent cells (arrow), a characteristic but nonspecific feature.

Occasionally, endometrial cell clusters consist of an
obvious dual cell population with small, dark stromal
cells (in the center) and larger glandular cells (around the
edges). Most endometrial cell clusters, however, do not
have this dual population. “Monocontoured clusters” like
that in Figure 1.12 may consist of glandular endometrial
cells, stromal endometrial cells, or a mix of both.106
Shedding endometrial cells after day 12 (“out of phase”)
is associated with endometritis, endometrial polyps, and
intrauterine devices (IUDs). In a young woman, abnormal shedding is almost never a result of endometrial
adenocarcinoma.107,108 For this reason, endometrial cells
do not need to be mentioned in the report for women
under 40 years of age. Some laboratories do so anyway, to
document that the cells were identified and interpreted
as benign endometrial cells. Endometrial cells are notorious for their ability to cause diagnostic difficulty, because
a variety of neoplastic cells resemble endometrial cells.

The differential diagnosis includes a number of significant lesions that mimic endometrial cells and thus
are sometimes mistakenly interpreted as normal, particularly if the woman is in the first 12 days of her menstrual 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
cells. Like endometrial cells, their nuclei are dark, and
they have scant cytoplasm (Fig. 1.13A). HSIL cells, even
when small, are usually bigger than endometrial cells,
vary more in size, and have denser cytoplasm. HSIL
clusters are usually less well circumscribed and are not
as spherical as endometrial cell balls. Some poorly differentiated squamous cell carcinomas (SQCs) are composed of small dark cells that mimic endometrial cells
to perfection (Fig. 1.13B). In such cases, suspicious
clinical findings (e.g., postcoital bleeding) might be the
only clue to the correct interpretation. Most AIS have a

columnar cell morphology, but a minority are made
up of smaller and rounder cells (Fig. 1.13C), particularly on LBC preparations. Careful examination for
focal ­columnar differentiation and mitoses can be quite
helpful. The rare small cell carcinoma of the cervix may
display crush artifact (Fig. 1.13D), which is rarely seen
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 carcinoma (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 endometrioid 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.

Abraded Endometrial Cells and
Lower Uterine Segment
The endocervical sampling device occasionally inadvertently samples the LUS or endometrium.109 This is especially likely when the endocervical canal is abnormally
shortened, as it is after a cone biopsy.110
Cytomorphology of abraded

endometrium and lower uterine
segment:
• large and small tissue fragments
• glands and stroma
• stromal cells
• uniform
• oval or spindle shaped
• finely granular chromatin
• occasional mitoses
• capillaries traverse larger fragments

• glands
• tubular
• straight or branching
• mitoses (some cases)
• extreme nuclear crowding
• scant cytoplasm
The characteristic feature is the combination of glands
and stroma, often in large fragments (Fig. 1.14A-C),
either together or separated. Glandular cells of the LUS
resemble endocervical cells, but have a higher nuclear to
cytoplasmic ratio, are more hyperchromatic, and can be
mitotically active. Because of their high nuclear to cytoplasmic ratio, they can be confused with a significant
squamous or glandular lesion.109

Trophoblastic Cells and Decidual Cells
Syncytiotrophoblastic cells from placental tissue are seen
rarely, perhaps in about 0.1% of smears from pregnant

The Normal PAP

17

A

B

C

Figure 1.14  Endometrial cells,
directly sampled. A, An intact
endometrial tubule is surrounded
by well-preserved endometrial
stromal cells. B, Benign stromal
cells are elongated and mitotically active (arrow) and may
suggest a high-grade squamous
intraepithelial lesion (HSIL) or
a malignancy. The pale, finely
granular chromatin and the association with intact endometrial
glands are clues to a benign diagnosis. C, The glandular cells are
crowded and mitotically active
(arrow), but evenly spaced.


18

Cervical and Vaginal Cytology

women.111 The cells are large, with abundant blue or pink
cytoplasm. They have multiple nuclei that have a granular chromatin texture and slightly irregular contours.
Trophoblastic cells can be distinguished from multinucleated histiocytes because their nuclei are darker and
more irregular in contour (Fig. 1.15). They do not show
the prominent molding and ground-glass appearance
of nuclei of herpes simplex infection. Immunostains for
human chorionic gonadotropin and human placental
lactogen can be used to confirm their identity as trophoblastic cells. The presence of syncytiotrophoblastic cells
is not a reliable predictor of an impending abortion.111

Decidual cells are isolated cells with abundant granular cytoplasm, a large vesicular nucleus, and a prominent nucleolus. They often show degenerative changes.

Inflammatory Cells
Neutrophils are seen in all Pap samples and do not
necessarily indicate infection, but they are ­present
in increased numbers after injury or infection.
Lymphocytes and plasma cells are rare, but ­occasionally—
most often in older women—they are ­numerous
(Figs. 1.16, 1.71A). This pattern is called follicular
cervicitis because biopsies show lymphoid follicle formation. The lymphocytes of follicular cervicitis can be
confused with HSIL cells, endometrial cells, and lymphoma. Histiocytes are associated with a myriad of
conditions (e.g., menses, pregnancy, foreign bodies,
radiotherapy, and endometrial hyperplasia and carcinoma) (Fig. 1.17), but by themselves are a nonspecific
finding of no clinical significance.

Lactobacilli

Figure 1.15  Syncytiotrophoblast. The nuclei of these multinucleated cells are dark and coarsely granular, unlike those of
histiocytes.

The vagina is colonized by gram-positive rod-shaped
bacteria of the genus Lactobacillus. They are beneficial because they produce lactic acid, which reduces
the ambient acid-base balance (pH) and possibly
protects from infection by Candida and other pathogens. Lactobacilli metabolize the glycogen contained
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

Organisms and Infections
Shift in Flora Suggestive of
Bacterial Vaginosis

Figure 1.17  Histiocytes. Histiocytes have abundant multivacuolated cytoplasm and an oval, occasionally folded nucleus.

pattern, commonly seen during the second (luteal)
phase of the menstrual cycle, is known as cytolysis—
bare intermediate cell nuclei, fragments of squamous
cytoplasm, and abundant bacterial rods (Fig. 1.18).
Cytolysis can interfere with one’s ability to evaluate
nuclear to cytoplasmic ratio, an important criterion in

grading SILs.

Artifacts and Contaminants
The more commonly encountered artifacts and specimen contaminants are illustrated in Figure 1.19.

A steep reduction in the proportion of lactobacilli, with
a concomitant predominance of coccobacilli, is associated with bacterial vaginosis, a disorder characterized
by a thin, milky vaginal discharge and a foul, fishy odor.
At one time attributed solely to Gardnerella vaginalis,
it is now clear that bacterial vaginosis can be caused
by other bacteria as well.112 The diagnosis is made by
correlating morphologic findings on a Pap or wet prep
with other test results (vaginal pH and the amineodor “whiff” test after adding potassium hydroxide
[KOH]).113
Cytomorphology of a shift in flora:
• short bacilli (coccobacilli), curved bacilli, or mixed
bacteria
• no lactobacilli
• “filmy” appearance
• “clue cells”
The cytologic hallmark is the replacement of
the normal lactobacilli by shorter bacilli (coccobacilli), curved bacilli, and mixed bacteria (Fig. 1.20).
These small organisms are numerous and give a filmy
appearance to the preparation. They frequently adhere
to squamous cells, completely covering them like a
shag carpet (“clue cells”). Clue cells are not specific

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.