Identification of glandular lesions in cervical cytology begins with an adequate sampling of the cervical transformation zone and endocervical canal. This is achieved by performance of a speculum examination of the cervix which entails opening of the vaginal canal with a speculum in order to visualize the ectocervix and cervical os. Cells are obtained from the exposed external cervix, the ectocervix, and from the endocervical canal, in order to assure evaluation of the transformation zone, the area of highest risk for preneoplastic and neoplastic squamous and glandular lesions. Several devices are available for cytology sampling. Cotton/dacron swabs had been widely used in the past but have been found to fail to release abnormal cells for examination so their use is no longer recommended [1]. Wooden spatulas should not be used with liquid-based fixatives, but may be used to make direct smears. Currently the most commonly used devices are the plastic spatula, the endocervical brush, and the cervical broom (Fig. 2.1). All devices are intended to circumferentially scrape the cervix. When using the endocervical brush, the bristles nearest the operator should be at the level of the external os of the cervix in order to obtain the best sample; the brush should be rotated 90–180° in one direction being careful not to over-rotate. The use of both the plastic spatula for the ectocervix and the endocervical brush for the endocervix has been shown to provide more endocervical cells than the spatula alone [2]. The cervix broom is a device designed to sample both the ecto- and endocervix simultaneously. The central longer bristles are inserted into the endocervical canal while the shorter, more peripheral bristles are splayed over the ectocervix. The bristles of the broom are angled in such a way that an adequate sample can only be obtained if the broom is turned clockwise for 360° five times.

For conventional smears (CS), the cell samples are smeared from the spatula (either wooden or plastic) and rolled from the brush onto a glass slide and immediately fixed in alcohol, either by direct immersion into a solution of ethanol or by the use of aerosol spray fixation (Fig. 2.2). The CS is now ready for staining with the Papanicolaou (Pap) stain, the stain mandated by the Clinical Laboratory Improvement Amendment of 1988. The Pap stain uses hematoxylin to color the nucleus and several other colors to stain the cytoplasm. The stain was originally developed to analyze the degree of keratinization of the squamous cell cytoplasm but also adequately stains endocervical cells.

Alternatively, the sampling device (only plastic) can be directly immersed into a proprietary vial of alcohol-based fixative in order to capture and release the cells; this method constitutes “liquid-based cytology” (LBC). In North America, two types of LBC are commercially available: ThinPrep^® Pap Test (Hologic, Inc., Marlborough, MA) and BD SurePath™ Pap Test (Becton-Dickinson) (Fig. 2.2). The tests have features that produce subtle but important differences in cervical cytology, especially relative to glandular epithelium.

ThinPrep (TP) employs a methanol-based fixative. The clinician must vigorously agitate the sampling device to extract the maximum amount of cells prior to removing and discarding the device. SurePath (SP) uses an ethanol-based fixative of relatively low concentration. The sampling devices used with SP have detachable heads which are simply dropped into the fixative.

Sample processing differs for the two products leading to a different presentation of the sample on the cytology slide. The TP sample is drawn up through a membrane without any preliminary extraction of unwanted material (mucus, red blood cells, or leukocytes) to which the cells adhere while fluid passes through to be discarded. When the cellularity is sufficient, the cells are pressed to a glass slide. The TP process gives the cell sample a relatively flattened presentation on the slide. SP processing begins with centrifugation with density gradient material to remove excess blood and leukocytes. An aliquot of sample is then allowed to settle by force of gravity onto the slide producing a more three-dimensional quality to the presentation especially prominent in tightly cohesive groups of glandular cells.

The decision to use CS vs. LBC depends on several factors including the rate of fully adequate samples, the reliability of the screening process, the ability to perform additional concurrent tests, and the cost. The cost of LBC is more than CS due to the need for disposables and additional laboratory equipment which is mitigated when additional tests are performed as a second separate sample is unnecessary for the other tests. Despite these differences between CS and LBC, the American College of Obstetricians and Gynecologists does not advocate for one method over the other citing a meta-analysis report of eight studies and a randomized trial that failed to show a significant difference [3–5]. Regarding the unsatisfactory rates between TP and SP, a recent systematic review and meta-analysis of 14 SP studies and 28 TP studies found significantly fewer unsatisfactory SP samples as compared to TP [6].

Performance of CS as compared to LBC has not found significant differences in the rate of detection of high grade squamous intraepithelial lesions [5, 7–9]. On the other hand, observational studies suggest that CS does not perform as well as LBC in the detection of glandular abnormalities [7, 10–14].

Obscuring blood may result in an unsatisfactory CS sample so the first few days of the menstrual cycle should be avoided [15]. Appropriately processed LBC and HPV testing results do not appear to be affected by the timing of the menstrual cycle [16]. Likewise gel lubricant on the speculum does not appear to interfere with the cytology [17, 18].

The cervical cytology sample must be accompanied by a requisition form which should be completed by the clinician in order to provide the basic patient information needed to interpret the sample. Required elements by US Federal Regulations include patient name, age, date of birth, ordering clinician, and whether the patient is undergoing routine screening or is at high risk for cervical disease. The most important elements for the specific cytologic evaluation of glandular cells are the patient’s age, the last menstrual period, gestational history, use of an intrauterine device (IUD) as a method of contraception, history of diethylstilbestrol exposure in utero, and any history of prior abnormal cytology samples or gynecologic pathology (Table 2.1). During the analysis of the slide, the cytologist should correlate the cytologic findings with the clinical information provided. If a discrepancy is identified between the information provided and the cytologic findings, the cytologist must make an attempt to explain the differences. If the electronic medical record and laboratory information system are available, the records can be searched for clarification. As an example, changes in glandular cells suggestive of the presence of an IUD may be identified, but the clinician has not indicated the placement of an IUD. Not infrequently in searching the medical record the history of IUD placement will be found.

Screening of cervical cytology slides is often assisted by image analysis (also known as automated review) prior to the review by the cytotechnologist. Two systems are FDA approved and available for image analysis: the ThinPrep Imager (TPI) and the SurePath FocalPoint family (Slide Profiler, Guided System [GS] and Location Guided Screening (LGS)) of which the GS is nearly exclusively used in North America. Both systems (TPI and FP GS) evaluate LBC slides and select fields of view (FOV) most likely to harbor abnormal cells for presentation to the cytotechnologist for initial review. The FOV make up approximately one third of the total slide area for both devices. If cytologic abnormalities are not identified by the cytotechnologist on initial review, the case is signed out as negative. If a potential abnormality is identified, the full slide is manually screened by the cytotechnologist. Both systems rely on robotic instrumentation that incorporates automated microscopy with high-speed image analysis. Both systems use complex computer algorithms to analyze morphologic findings. At the heart of the ThinPrep system is the use of a modified “Feulgen-like” hematoxylin stain which enables the image analysis software to calculate the relative stoichiometric DNA content of the epithelial cells. FOVs are presented in a geographic fashion for each slide. The FocalPoint family uses traditional Papanicolaou-stained slides and hierarchically “scores” each slide and the FOVs on each slide, presenting the FOVs to the reviewer in order of probability of abnormality (highest probability FOVs first). In addition, the slides within a given run are ranked according to their potential to harbor abnormal cells.

Both the TPI system and FP GS system have been found to accurately triage slides with glandular cells abnormalities to full manual review [19, 20]. However, as compared to the ranking of slides containing high grade squamous lesions (HSILs are routinely ranked at the highest scores), the FP GS system may not always assign a slide containing a glandular lesion to the highest scoring ranks (slides with glandular lesions can occur at any score level) and hence all slides need to be carefully reviewed for their presence [21].

In 1988 a new nomenclature, the Bethesda System (TBS), for reporting cervical cytology results was devised in order to standardize terminology, to produce clinically useful reports improving communication with clinicians and other cytologists, and to improve the reproducibility of diagnostic criteria for each entity. The Bethesda terminology has undergone one revision which took place at a second consensus conference in 2001. In the newest edition of the terminology, recommended components of the report include specimen type (conventional smear vs. LBC), adequacy (satisfactory or unsatisfactory), interpretation (negative for intraepithelial lesion or malignancy (NILM), other (endometrial cells in a woman over 40 years of age) or epithelial cell abnormalities (squamous or glandular) or other malignant neoplasms), ancillary testing, and automated review. An optional quality indicator component allows the report to signal features such as the absence of an endocervical/transformation zone component in an otherwise satisfactory sample, which may have clinical utility in selected circumstances. Under the interpretative category of NILM are included reactive changes; for example, reactive endocervical cells resulting from the presence of an intrauterine contraceptive device. Epithelial cell abnormalities: glandular are subcategorized according to cell-type (endocervical, endometrial, or extrauterine) and as to severity (atypical not otherwise specified, atypical favor neoplastic, adenocarcinoma in situ (AIS), and adenocarcinoma (ACA)) (Table 2.2). The first edition of TBS used the umbrella term, atypical glandular cells (AGC) of undetermined significance (AGUS) and required qualification as to cell of origin (endocervical or endometrial). In the second edition, AGUS was abandoned for at the interpretive designation AGC.