Benign Disorders of the Uterine Cervix and Vagina

Benign Disorders of the Uterine Cervix and Vagina


In this chapter, a number of benign epithelial cytologic abnormalities is described. The knowledge of these abnormalities is essential in the practice of cytopathology because some of these changes must be differentiated from malignant processes. The abnormalities will be presented according to the epithelium of origin but in practice may be found side by side in the same cytologic preparation.


Basal Cell Hyperplasia

This is a frequent finding in the biopsies of the cervix, as it is in other organs, such as the bronchus. On histologic examination, there is an increase in the number of layers of small basal cells. Consequently, the basal cells form onefourth to one-third of the thickness of the squamous epithelium (see Fig. 6-11). However, there are no nuclear abnormalities and the maturation of the epithelium proceeds
normally above the enlarged basal zone; hence, basal hyperplasia of squamous epithelium cannot be detected in smears unless there is a total loss of superficial cell layers or if the sample is obtained by energetic brushing. The etiology of the process is unknown but there is no evidence that it is related to cancer.


Leukoplakia (from Greek, leukos = white) is a clinical term describing a white patchy discoloration of the squamous epithelium. It is more common on the surface of the cervix than in the vagina. In most cases, it is a benign lesion caused by abnormal keratinization of the mucosal surface. It must be emphasized, however, that keratinizing precancerous lesions of the uterine cervix and occasionally, keratinizing invasive carcinoma, may appear as clinical leukoplakia (see Chap. 11).


In benign leukoplakia, the surface of the epithelium is covered by an eosinophilic layer of keratin of variable thickness. The keratinized layer is composed of anucleated cells, akin to those observed on the surface of the epidermis (Fig. 10-1A). The lesion may result from chronic trauma, prolapse of the uterus, pressure of a pessary, or previous cauterization of the cervix.


Anucleated squamous cells (squames), of characteristic pale yellow color in Papanicolaou stain, are shed from the surface of the keratinized squamous epithelium (Fig. 10-1B). All stages of transition between mature nucleated squamous cells and the anucleated variety may be observed: There is a gradual change in the color of the cytoplasm from pink to yellow and the gradual disappearance of the nucleus. In the fully keratinized squames, shadows of pre-existing nuclei can be observed. Occasionally, brown cytoplasmic granules, akin to those seen in the normal superficial squamous cells, may be noted. The actual frequency of anucleated squamous cells in cervicovaginal smears is low, estimated by Kern (1991) to be about 0.5%. In my opinion, the presence of anucleated squamous cells should be noted in the report because of the remote possibility that these cells may have originated from the surface of a squamous cancer, masquerading as leukoplakia. Such patients deserve a closer clinical look, although nearly all the lesions are benign. In squamous carcinomas with features of leukoplakia, anucleated squamous cells are usually accompanied by cells with abnormal, hyperchromatic nuclei that allow an accurate diagnosis, as discussed in Chapter 11.

Figure 10-1 Leukoplakia of uterine cervix. A. Biopsy showing a layer of keratin on the surface. B. Cervical smear from the same case. The yellow or yellow-orange superficial squamous cells are either anucleated (anucleated squames) or show residual traces of nuclei.

Parakeratosis (Pseudoparakeratosis) of Squamous Epithelium


Occasionally, the surface of an otherwise normal squamous epithelium of the uterine cervix is lined by a few layers of very small nucleated squamous cells, instead of the usual large, mature squamous cells. This disorder is somewhat similar to the abnormalities occurring on the surface of the epidermis of the skin (for example, in psoriasis), which are referred to as parakeratosis. Therefore, this abnormality of the cervix epithelium was named pseudoparakeratosis.


In cervical smears, large sheets of irregularly-shaped, small squamous cells, about 10 μm in diameter, with basophilic or eosinophilic cytoplasm, may be observed (Fig. 10-2A). The nuclei are spherical, of even sizes, often somewhat eccentric and usually pyknotic (Fig. 10-2B). The cause or the exact frequency of this benign abnormality is unknown. Voytek et al (1996) observed similar small cells in benign lesions but also in the presence of low- and high-grade squamous neoplastic lesions of the cervix. It is not certain whether this association is incidental. Still, the recognition of the pseudoparakeratotic cells as benign is sometimes very difficult because of their nuclear features. In case of doubt, colposcopy and biopsies of the cervix should be recommended. Frable (1994) included such cells among the litigation
cells that are difficult to classify and may be the subject of a dispute as to their exact significance.

Figure 10-2 Pseudoparakeratosis or parakeratosis. A. The cervical smear at survey magnification shows large sheets of very small squamous cells. B. At higher magnification, the spherical or slightly irregular cells have uniform, dark nuclei.

Somewhat similar cells were observed by Meisels and Fortin (1976) in human papillomavirus (HPV) infection and were named dyskeratocytes. Dyskeratocytes, as described and illustrated, are small elongated, spindly cells forming sheets with similar nuclear characteristics to pseudoparakeratotic cells described above. Meisels et al (1976) observed HPV virions in electron microscopy of dyskeratocytes. However, in pseudoparakeratosis, the underlying epithelium does not show any changes of papillomavirus infection and, therefore, the two entities may be similar but are possibly of different etiologies. (See Chap. 11 for further discussion of cytologic manifestations of human papillomavirus infection.)


Basal Cell Hyperplasia


Proliferation of the small, basal endocervical cells, akin to the basal hyperplasia of squamous epithelium, may result from chronic inflammation or it may occur spontaneously. In tissue sections, several layers of small, round cells are found beneath the columnar endocervical cells (see Fig. 6-12).


Basal cell hyperplasia is rarely seen in cervical scrape smears. However, with the use of cervical brush instruments, particularly after a vigorous brushing, basal cell hyperplasia may be seen from time to time in the form of flat clusters of small, spherical or polygonal cells, measuring about 10 μm in diameter. The cells have scanty cytoplasm and relatively large, dense but regular nuclei, about 8 μm in diameter. The nuclei vary somewhat in size and shape and may show mitotic activity. The recognition of the basal endocervical cells is easier if well-differentiated columnar cells are attached to the periphery of such clusters (Fig. 10-3). The role of basal cell hyperplasia as a step in squamous metaplasia of the endocervical epithelium is discussed below.

The significance of basal cell hyperplasia of endocervical epithelium is in the similarity of these small benign cells to small cancer cells. The benign cells occur in flat clusters that are rarely dispersed. Also, the basal cells lack the nuclear features characteristic of cancer cells, described and discussed in Chapter 11. Errors of interpretation of such small cells may be twofold: benign basal cells may be mistaken for cancer cells but, more often, the small cancer cells are mistaken for benign cells, sometimes with disastrous consequences for the patient.

Squamous Metaplasia

Squamous metaplasia is a replacement of normal endocervical epithelium by squamous epithelium of varying degrees of maturity (see Fig. 6-12). This is a very frequent event that may be confined to a small area of the endocervix
or may be very extensive and involve the surface epithelium and the glands. Squamous metaplasia is a normal, physiological event during maturation of the female genital tract, notably in the epithelium of the transformation zone, as discussed in Chapter 8.

Figure 10-3 Basal cell hyperplasia of endocervix. A cluster of relatively small polygonal cells attached to well-differentiated endocervical cells.


Electron microscopic studies, notably by Bonilla-Musoles and Barbera (1970) and by Philipp (1975), documented that the small basal or reserve cells of the endocervical epithelium have the dual potential of differentiating either into mucus-producing normal endocervical cells or into squamous cells, a view confirmed by Tsutsumi et al (1993). It is likely that hormones, notably estrogens, play a role in the origin of squamous metaplasia. However, because squamous metaplasia is often observed in biopsy material from cervices showing chronic inflammation, it has also been linked to inflammatory processes. Squamous metaplasia is also observed in the presence of intrauterine contraceptive devices and, therefore, may be a reaction to mechanical pressure. The precise mechanisms leading to squamous metaplasia must still be elucidated. There may be some analogy between squamous metaplasia and the events occurring in the formation of the epidermis, as proposed by Sun et al (see Chap. 2). Sun documented that the transformation of a simple cuboidal epithelium into mature epidermis of the skin is accompanied by a sequential activation of keratin genes of ever higher molecular weights (see Fig. 2-25). Still, keratin polypeptides in metaplastic epithelium appear to be unique and differ from cytokeratins normally present in the squamous epithelium of the cervix and those found in the endocervical epithelium (Gigi-Leitner et al, 1986).

Figure 10-4 Squamous metaplasia of endocervix. A. A focus of squamous metaplasia undermining the endocervical glandular epithelium. B. Various stages of squamous metaplasia of the endocervix ranging from mature to immature manifestations of this process. C. A cluster of cells from a cervicovaginal smear corresponding to B and showing cohesive sheets of metaplastic cells. D. Sheets of metaplastic squamous cells showing the characteristic irregular configuration of the cytoplasm.

The metaplastic epithelium occurring within the area of the transformation zone (squamocolumnar junction) is commonly involved in initial neoplastic events affecting the uterine cervix, as discussed in Chapter 11.


The earliest stages of squamous metaplasia can be identified as a focus of multiplication of the basal cells of the endocervical epithelium. As these small cells grow towards the surface and become larger, their cytoplasm becomes eosinophilic and homogeneous (Fig. 10-4A). The glandular epithelium may remain on the surface while the underlying squamous epithelium forms increasingly mature cells. In most cases, however, the glandular surface epithelium is cast off and the endocervical epithelium is replaced by squamous
epithelium that may be either mature, resembling the squamous epithelium of the vagina, or immature, composed of smaller squamous cells of intermediate or parabasal type (Fig. 10-4B). The term immature, as used here, should not be confused with a malignant process. Various stages of transition between normal endocervical epithelium and mature metaplastic squamous epithelium may be observed. By special stains, mucus may nearly always be demonstrated in the cytoplasm of the metaplastic cells, indicating close relationship to the endocervical epithelium. Squamous metaplasia may replace the endocervical mucosa lining the endocervical canal or endocervical glands. Squamous metaplasia of endocervical glands may be discrete and focal, or diffuse. In extreme cases, one or several glands may be filled with squamous epithelium. If the metaplastic squamous epithelium is immature, the finding may mimic a neoplastic process, as discussed below under atypical metaplasia.


Only immature squamous metaplasia can be identified in cytologic material because cells derived from mature metaplasia cannot be distinguished from normal squamous cells. Squamous metaplasia of the endocervical mucosa can be diagnosed with certainty in cervical smears if flat sheets of polygonal parabasal squamous cells with basophilic or eosinophilic cytoplasm are contiguous with columnar endocervical cells (Fig. 10-4C). Within the sheets, the metaplastic squamous cells usually form clearly visible cell borders. One surface of the sheets is often flattened, corresponding to the surface of the metaplastic epithelium. Occasionally, the clusters of metaplastic cells are loosely structured and are composed of angulated squamous cells (Fig. 10-4D). On close inspection, the distinct flattening of the surface of the cluster is evident and sometimes there are transitions toward well-formed, mucus-producing columnar endocervical cells.

The cytoplasm of the metaplastic cells is either basophilic or eosinophilic and may show fine vacuoles in which mucus can be demonstrated by special stain. Occasionally, the vacuoles can be large and may be infiltrated with polymorphonuclear leukocytes. The nuclei of the metaplastic cells are spherical, measuring on the average 8 μm in diameter, but may be larger. Within the nuclei, small chromocenters and occasionally tiny nucleoli may be observed. Rarely, small spindly keratinized cells with slender pyknotic nuclei may originate from the surface of squamous metaplastic epithelium.

Unfortunately, metaplastic cells in their classical configuration in sheets are not always present in cell preparations, particularly those collected in a liquid medium and subsequently dispersed. In such preparations, the parabasal metaplastic squamous cells occur singly and are usually characterized by irregular, polygonal configuration with cytoplasmic processes, or spikes, as commonly observed in parabasal cells removed from their epithelial setting. The cytoplasmic processes are an artifact occurring during smear preparation by extensions of the cytoplasm at points of desmosomal junctions with adjacent cells. As the cells are being separated during smear preparation, the solid desmosomes resist rupture better than the elastic cytoplasm, which, as a consequence of mechanical stretching, becomes elongated at points of junction. Occasionally, one surface of these cells is flat, corresponding to the lining of the endocervical canal (see Fig. 10-8A).

As has been discussed in Chapter 8, the mere presence of parabasal squamous cells in a cervicovaginal smear is not diagnostic of metaplasia. Such cells may originate from the squamous epithelium of the vagina or the cervix under a variety of circumstances unrelated to metaplasia. The term metaplastic cells that has been suggested for parabasal cells, while picturesque, is not always scientifically sound.

Several attempts have been made to distinguish metaplastic cells from parabasal cells derived from native squamous epithelium by identification of keratins of various molecular weights. Thus, keratins 15, 16 and, occasionally, keratin 6 were observed in endocervical reserve cells and in squamous metaplasia (Smedts et al, 1993) whereas positive stain for keratin 17 was found useful in the differentization of metaplastic cells from normal parabasal cells (Martens et al, 1999).

Still, if angulated parabasal cells are trapped in the endocervical mucus or if the sample has been removed directly from the transformation zone or the endocervical canal by an instrument, such cells may be considered as adequate evidence that the smear is representative of the endocervical epithelium undergoing squamous metaplasia. The information on the value of such findings in liquid preparations as evidence of smear adequacy is not available. This issue is important in ensuring the adequacy of cervical smears, discussed in Chapter 8.

Atypical Squamous Metaplasia

Occasionally, the component cells of squamous metaplasia in tissue and smears show slight to severe abnormalities. The slight changes are cell and nuclear enlargement or binucleation confined to a few cells within the cluster (Fig. 10-5A). Although, in my experience, slight abnormalities of metaplastic cells are usually of no consequence to the patient and the biopsies in such cases disclose minimally atypical metaplasia, more severe changes, described below, often require further investigation of the patient.

More severe changes in metaplastic cells include significant cellular and nuclear enlargement, variability in nuclear sizes, coarse granulation of chromatin and the presence of prominent nucleoli (Fig. 10-5B-D). Some of these changes may resemble “repair,” discussed below. Because the precise identification of such changes is difficult, their classification as “metaplastic” or “endocervical” may depend on the preference of the observer. No doubt some of these abnormalities could be classified as atypical squamous or endocervical cells of unknown significance (ASCUS or AGUS), which will be discussed in detail in Chapter 11. On further investigation, many such patients
are subsequently shown to harbor precancerous lesions or even cancer of either squamous or endocervical type. Therefore, patients with marked nuclear changes (Fig. 10-5B-D) should have the benefit of a close, careful follow-up, including colposcopy and biopsies of cervix, particularly if, in addition to cell clusters, single abnormal cells are present in the smear. The role of testing for human papillomavirus in such cases is discussed in Chapter 11.

Figure 10-5 Atypical squamous metaplasia. A. A cohesive sheet of metaplastic endocervical cells showing nuclear enlargement and small nucleoli. B. A sheet of metaplastic cells showing significant nuclear enlargement. There was no evidence of a neoplastic lesion in this patient. C. A sheet of metaplastic endocervical cells showing variability in nuclear sizes. D. A sheet of metaplastic cells showing the “honeycomb” arrangement with clearly visible cell borders. The sizes of the nuclei are variable.

Tubal and Tubo-Endometrioid Metaplasia

Besides common squamous metaplasia, the endocervical epithelium occasionally shows features of tubal or endometrial epithelium, sometimes side by side. The condition has been labeled metaplasia, although in all probability, it represents a common variant of normal endocervical epithelium. As noted in Chapter 8 and described in detail by Babkowski et al (1996), ciliated endocervical cells are invariably present in the upper, proximal reaches of the normal endocervical canal.


In 1990, Suh and Silverberg described tubal metaplasia as a replacement of the normal epithelium, lining the endocervical surface and glands, by epithelium of tubal type, composed of columnar ciliated cells, clear secretory cells, and intercalated cells (Fig. 10-6A). The secretory cells may show apocrine snouts, or small cytoplasmic projections, on their surfaces. The variant is common, as it was found in 31% of surgical specimens by Jonasson et al (1992). Some of the endocervical glands may also show the features of endometrial epithelium, occasionally accompanied by endometrial stroma, and thus suggestive of endometriosis (Oliva et al, 1995). The tubal or endometrial lining may form papillary projections and show cystic dilatation. All these features confer on the endocervical glands an unusual aspect that may vaguely resemble precursor lesions of endocervical adenocarcinoma, discussed in Chapter 12. However, the cells lining the endocervical glands usually do not show any significant nuclear abnormalities and are, therefore, consistent with benign lesions. However, personal experience shows that in some of these lesions, the ciliated epithelium and adjacent endocervical glands may show marked abnormalities in the form of accumulation of small, highly abnormal cells with irregular, hyperchromatic nuclei, strongly suggestive of a small cell carcinoma. This observation received strong support from work by Schlesinger and Silverberg (1999) who described several
cases of endocervical carcinoma in situ of tubal type and stressed that the mere presence of ciliated epithelium does not guarantee that the lesion is benign.

Figure 10-6 Tubal metaplasia. A. Biopsy of endocervix showing glands with ciliated surface and minor variability in nuclear sizes. B. Ciliated endocervical cells with somewhat hyperchromatic nuclei of even sizes. C. A sheet of endocervical cells from the upper reaches of the endocervical canal showing dark, somewhat enlarged nuclei of even sizes.

Cytology of Tubal Metaplasia

With the widespread use of rigorous endocervical brushing techniques the presence of ciliated endocervical cells has become much more common in cervical smears than with the use of cervical scrapers. Thus “tubal metaplasia” became a recognized cytologic entity. In brush smears in some cases of tubal metaplasia, one can observe sheets of essentially normal ciliated endocervical cells with enlarged hyperchromatic nuclei (Fig. 10-6B,C). In some such cases, the nuclear abnormalities may be significant and the cells may be classified as atypical glandular cells of unknown significance (AGUS), requiring further investigation. Ducatman et al (1993) also claimed that one could recognize in smears the “peg cells,” characterized by “dark and granular cytoplasm and elongated nuclei.” We have not been able to identify such cells with certainty.

Cytologic Abnormalities in Tubal Metaplasia

Novotny et al (1992) claimed that tubal metaplasia may shed abnormal cells that can be mistaken for cells derived from precursor lesions of endocervical adenocarcinoma. These authors provided an elaborate table, listing the cytologic features of both types of lesions. This has not been my experience. As mentioned above, in several cases seen in consultation in which smears contained ciliated cells accompanied by small abnormal or suspicious cells, the corresponding biopsy material disclosed marked abnormalities of endocervical glands and adjacent endocervical epithelium, consistent with a malignant process. Within the neoplastic epithelium, ciliated cells were occasionally present. These observations confirm that cytologic abnormalities always have their counterpart in corresponding histologic material and are in keeping with the concept of carcinomas with ciliated cells, perhaps developing in tubal metaplasia, as described by Schlesinger and Silverberg (1999). For further discussion of precancerous lesions of the endocervical epithelium, see Chapter 12.


It has been mentioned above that Oliva et al (1995) observed the presence of endometrial stromal cells in areas of endometrioid metaplasia of the endocervix, consistent with the diagnosis of endometriosis. Similar observations were reported by Yeh et al (1993). No specific histologic or cytologic abnormalities were attributed by these authors to the histologic findings. However, Mulvany and Surtees (1999) reported that the cervical brush smears in 7 of 10 cases of endometriosis located in the uterine cervix or vagina contained cells interpreted as consistent with endocervical adenocarcinoma, either invasive or in situ. The cytologic
abnormalities consisted of nuclear hyperchromasia, large cohesive cell clusters, and the presence of mitotic activity and of necrotic cells, described as “apoptosis.” Similar observations were reported by Hanau et al (1997). The histologic findings in these cases disclosed various degrees of atypia in the glands which, in two cases, was “severe.” Although Mulvany and Surtees did not classify the abnormalities as malignant, it is well known that endometrioid carcinomas may originate in the endometriotic glands (Koss, 1963; Brooks and Wheeler, 1977; Mostoufizadeh and Scully, 1980). These abnormalities must be considered in the differential diagnosis of endocervical adenocarcinoma.

Lesions Erroneously Classified as Metaplasia

There are few, if any, fields of microscopy that are as difficult to interpret as cervicovaginal cytopathology. The difficulty is compounded by the facts that many of the precancerous lesions of the uterine cervix have an unpredictable behavior and may require many years of follow-up until their true nature becomes evident, as discussed in Chapter 11. The interpretation of the cytologic material and of histologic patterns in biopsy material may be equally difficult and the separation of the atypical metaplastic processes from true neoplasia may depend on the judgment, training, and experience of the observer. Some investigators have suggested that human papillomavirus typing may help in the final classification of such difficult lesions (Crum et al, 1997) but this expensive technique is not always available and not necessarily reliable. It is, therefore, not surprising that under the term of “metaplasia,” several entities have been described in recent years that in the judgment of this writer are an erroneous and misleading misclassification of lesions that belong to the spectrum of precancerous lesions of the uterine cervix.

Trivijitslip et al (1998) described a papillary immature metaplasia (PIM), a form of somewhat atypical metaplasia with formation of papillary folds. This abnormality was thought to represent an immature form of condyloma of the uterine cervix epithelium and was attributed by Mosher to infection with nononcogenic papillomaviruses types 6 and 11. Still, in a further study of PIM by Trivijitslip and Mosher (1998), three of nine patients with PIM harbored a high-grade squamous intraepithelial lesion. Thus, PIM should be included in the spectrum of precancerous events in the uterine cervix.

A particularly dangerous concept has been introduced into cervical cytology by Dressel and Wilbur (1992) under the term atypical immature squamous metaplastic cells. Atypical metaplastic cells are small cancer cells that have some superficial similarity to metaplastic cells but differ markedly by the configuration of nuclei. Such cells are described and discussed in Chapter 11. In an elaborate study, Geng et al (1999) confirmed the neoplastic nature of “atypical immature metaplasia.” In 12 of 15 patients in whom the presence of human papillomavirus DNA was documented by polymerase chain reaction, a concurrent or subsequent diagnosis of a high-grade squamous epithelial lesion was established on biopsies. Park et al (1999) concluded that observer agreement in the differential diagnosis of this lesion from a high-grade neoplastic lesion was poor and not assisted by HPV typing.

Another misleading concept has been introduced by Egan and Russel (1997) under the name of transitional (urothelial) cell metaplasia of the uterine cervix. The entity was also described by Weir et al (1997) but was strenuously rejected by Koss (1998). In my judgment, many of the lesions described represent intraepithelial neoplastic lesions of the uterine cervix undergoing atrophy and occurring mainly in postmenopausal women. Harnden et al (1999) admitted that the so-called transitional metaplasia lacks the cardinal features of the urothelium. In many such cases, the cytologic and biopsy findings suggested a high-grade squamous intraepithelial lesion (HGSIL).

The entities inappropriately classified as metaplasia are discussed again in Chapter 11 together with other precancerous lesions of the cervix.

Microglandular Hyperplasia of Endocervical Glands

In 1967, Taylor et al reported that, in biopsies of patients receiving contraceptive hormones, a marked proliferation of small endocervical glands may be observed. The lesion may be polypoid and imitate an endocervical polyp or, in rare cases, adenocarcinoma of the cervix or endometrium (Young and Scully, 1989). Numerous other reports were published subsequent to Taylor’s paper, generally attributing this lesion to contraceptive medication containing progesterone (recent summary in Candy and Abell, 2001). There is no doubt, however, that such lesions may also be observed in the absence of contraceptive medication, in pregnant and postmenopausal women, and, in about onefourth of all cases, without an obvious cause at all.

Histology and Cytology

In tissue sections, microglandular hyperplasia consists of a grouping of well-formed endocervical glands of markedly variable sizes. The glands are lined by normal endocervical cells. The proliferation of smaller glands may deceptively suggest an invasion of the stroma and thus adenocarcinoma (Fig. 10-7A). Unusual histologic presentation of these lesions may include areas of solid growth and mucus-containing signet ring cells (Leslie and Silverberg, 1984; Young and Scully, 1989). In some such cases, the differentiation from endocervical or endometrial adenocarcinoma may cause problems. Clinical history and absence of mitotic activity and nuclear abnormalities are usually sufficient to reach the correct diagnostic conclusion.

On the other hand, some well-differentiated carcinomas may sometimes mimic microglandular hyperplasia. Nuclear abnormalities and mitotic activity are present in such lesions (Young and Scully, 1992). It has been suggested by Daniele et al (1993) and Alvarez-Santin et al (1999) that microglandular hyperplasia can be recognized in cervical brush smears by the presence of 2- and 3-dimensional fenestrated clusters of endocervical cells, some forming small, gland-like structures
(Fig. 10-7B). In order to secure such clusters, a very energetic brushing of the endocervix is required. In our experience, there are no specific cytologic abnormalities that would allow a reproducible recognition of microglandular hyperplasia. Some of the findings described are most likely artifacts caused by endocervical brushings (see below). However, some observers reported marked cytologic abnormalities associated with microglandular hyperplasia (Valente et al, 1994; Selvaggi and Haefner, 1997; Selvaggi, 2000). It is our judgment that, in such cases, the microglandular hyperplasia is an incidental biopsy finding. There is excellent evidence that the presence of microglandular hyperplasia does not rule out the presence of precancerous intraepithelial lesions of the cervix, as first pointed out by Nichols and Fidler (1971), or, for that matter, one of the rare cervical or endometrial adenocarcinomas mimicking microglandular hyperplasia (Young and Scully, 1992). Thus, the presence of atypical cells in smears calls for a careful investigation of the uterine cervix by methods discussed in Chapters 11 and 12, regardless whether or not microglandular hyperplasia is present.

Figure 10-7 Microglandular hyperplasia. A. Section of endocervix showing numerous glands of uneven sizes and configuration, separated by fibrous stroma and lined by cuboidal cells of even sizes. B. A large complex fragment of endocervical epithelium in a brush specimen, conceivably reflecting microglandular hyperplasia.

Other Benign Abnormalities of Endocervical Glands

Tunnel clusters is an arrangement of endocervical glands forming small tubules or “tunnels” (Fluhmann, 1961A). A cystic form of this abnormality has been described (Segal and Hart, 1990).

Mesonephric hyperplasia is a proliferation of the remanents of Gartner (mesonephric) ducts usually located within the smooth muscle of the cervix (Ferry and Scully, 1990; Jones and Andrews, 1993).

Endocervicosis is a proliferation of histologically normal endocervical glands in an abnormal location, such as the outer rim of the uterine cervix or the urinary bladder (Young and Clement, 2000).

These lesions may be confused with endocervical adenocarcinoma in tissue sections. There is no systematic study of the cytologic presentation of these benign lesions. Anecdotal evidence suggests that there are no specific cytologic abnormalities that would allow their recognition. The cytologic presentation of Gartner (mesonephric) duct carcinoma is described in Chapter 12.


As discussed in Chapter 8, the squamous and endocervical types of epithelium usually meet within the squamocolumnar junction or the transformation zone of the uterine cervix. In some young women, the two epithelia meet on the surface of the vaginal portio of the cervix, outside of the external os. As a consequence, a sharply circumscribed red patch appears on the vaginal portio of the cervix, adjacent to the external os. The patch is lined by the delicate endocervical glandular epithelium, which is transparent. The visible vessels of the cervical stroma give the area the red appearance on visual inspection. Because to the naked eye the patch may mimic an ulcer, the faulty clinical term of erosion has been applied to the lesion. The correct term is eversion or ectropion (from Greek: ek = out, trope = a turning). The ectropion may occupy a small segment of the visible cervix or surround the external os (circumoral erosion). The importance of the ectropion is largely clinical, inasmuch as the red patch must be differentiated from true ulceration of the cervix due to inflammation or to cancer.

Histology and Cytology

The patch is lined by typical endocervical epithelium, sometimes forming papillary folds, containing a normal, occasionally hyperemic stroma (Fig. 10-8A). It is not uncommon to find a few normal endocervical glands in the area
of the ectropion. Direct cervical smears contain only fragments of benign endocervical tissue and single columnar endocervical cells (Fig. 10-8B). Because the delicate epithelium lining the area is readily damaged during the process of obtaining smears, fresh blood is often present in the cytologic specimen. If a portion or all of the everted endocervical mucosa has undergone squamous metaplasia, squamous cells of varying degrees of maturity will appear in smears. The cytologic picture is vastly different from inflammatory or neoplastic lesions that may also occur in this part of the cervix, as described below and in Chapter 11. The eversion requires no treatment because, with the passage of time, it will undergo squamous metaplasia.

Figure 10-8 Eversion of endocervical mucosa. A. Histologic section showing the surface of the transformation zone lined by endocervical epithelium. Foci of squamous metaplasia are present. B. Sheet of normal endocervical cells characteristic of eversion in cervical smears.


The term repair has been introduced into the field of gynecologic cytology by Bibbo et al (1971) and by Patten (1978). These authors described atypical cells of endocervical and squamous origin with a number of abnormal cytoplasmic and nuclear features in patients with recent past history of radiotherapy to the uterine cervix, recent hysterectomy, other clinical procedures, such as cautery or biopsy, past history of severe cervicitis (Bibbo et al, 1971), and “partial or complete destruction (of the epithelium) by infection and inflammation” (Patten, 1978). Thus, this is a very heterogeneous group of patients wherein many different factors may account for the cellular abnormalities. Most important, perhaps, histologic evidence of true repair of a damaged epithelium (i.e., epithelial regrowth over a defect) has not been provided by Bibbo and to a very limited extent by Patten.

The concept of repair is valid but only under well-defined circumstances, for example, after a conization of the uterine cervix or other documented form of epithelial injury. In the histologic material, tongues of poorly formed, young epithelial cells bridging the defect caused by prior surgery may be observed (Fig. 10-9A). The mechanisms of “repair” or healing of a wounded epithelium are extremely complex (Singer and Clark, 1999) and have not been studied in the uterine cervix but probably resemble those in the skin.


A smear obtained approximately 1 week after a procedure damaging the cervical epithelium may show some rather characteristic cytologic features. The smears show flat sheets, composed of tightly fitting cells, generally resembling metaplastic cells (Fig. 10-9B,C). The cells may vary in size and their cytoplasm may be vacuolated and infiltrated with polymorphonuclear leukocytes. Occasionally, the cells may show bizarre, sometimes elongated configuration. The nuclei of these cells also vary in size, may show some degree of hyperchromasia, and, most importantly, contain one or more clearly visible nucleoli of variable sizes. Mitotic figures can be observed (Fig. 10-9D). There are usually few, if any, single cells with similar characteristics. The background of the smear usually shows a great deal of fresh blood and inflammation.

The manifestations of repair may be particularly difficult to interpret in smears of postmenopausal women with epithelial atrophy. The nuclei of the epithelial cells are enlarged, of uneven sizes, and hyperchromatic, mimicking malignant lesions. Because the thin epithelial lining offers little protection, bundles of spindly cells representing cervical stroma, may be present. We observed such cells after cervical biopsies, conization, or energetic curettage. “Repair” reaction may occur after surgical procedures, as illustrated above, as a reaction to chronic inflammatory events or in the presence of a foreign body or object in the endocervical canal. For example, intrauterine contraceptive devices or endocervical polyps may be the cause of such abnormalities (see below). There remain, however, a number of patients in whom similar cell abnormalities may be observed in smears in the absence of any known events that could account for the “repair.” It is likely that, in such patients, the cell abnormalities represent an exuberant
or florid squamous metaplasia of the endocervical epithelium.

Figure 10-9 “Repair.” A. The appearance of the cervical epithelium 4 days after a cone biopsy. The surface is lined by small, atypical epithelial cells. B. A sheet of epithelial cells from the same patient showing marked variability in sizes and configuration, corresponding to the young epithelium lining the surface of the cervical defect. C. Irregular flat sheets of endocervical or metaplastic cells showing prominent nucleoli. D. A cluster of endocervical cells showing marked variability in nuclear sizes and prominent nucleoli. Mitotics figure are present (arrows).

The cytologic changes attributed to repair are similar to those occurring in inflammatory changes in the endocervical cells, described further on in this chapter, or in atypical squamous metaplasia, described above. The conspicuous atypia, such as shown in Figure 10-9D, may reflect a neoplastic lesion, such as a high-grade squamous epithelial lesion or an endocervical adenocarcinoma. In such cases, however, the cell clusters are usually accompanied by single cells with similar morphologic features (see Chaps. 11 and 12). Geirsson et al (1977) analyzed the features of the cells in repair and compared them with cells of adenocarcinoma. There was a significant overlap between the two lesions, particularly in reference to nucleolar abnormalities. It is virtually impossible to arrive at a conclusive cytologic diagnosis in such cases and patients should be evaluated further by colposcopy and cervical and endocervical biopsies. The Bethesda System of cervicovaginal smear classification, described in Chapter 11, calls for separation of “typical” from “atypical” repair. This proposition is not reasonable because the cytologic pattern of repair is atypical by definition. For example, Rimm et al (1996) observed that 25% of patients with “atypical repair” pattern in smears harbored squamous intraepithelial lesions of low or high grade. Colgan et al (2001) reported that in a major survey of laboratories in the United States, repair was the most common source of false-positive and false-negative smears.

In the absence of supporting clinical or histologic evidence, the concept of repair, although occasionally correct, is a dangerous one as it may mislead even an experienced observer. Cell abnormalities of considerable magnitude not based on secure clinical and histologic data must be investigated further, regardless of label.


This common benign tumor may originate in any area of the endocervical canal. Histologically, a polyp is composed of a central connective tissue stalk lined by gland-forming endocervical mucosa. Squamous metaplastic epithelium may replace portions of the outer lining and may extend to the glands. Inflammation of varying type and intensity is often observed in the stroma. Very uncommonly, cervix cancer may start in a polyp.

Figure 10-10 Repair reaction caused by an endocervical polyp. A,B. Marked abnormalities of metaplastic squamous cells with very large nuclei and bizarre cell shapes.

An endocervical polyp cannot be recognized in cervical smears, except for the rare events when a fragment of polyp may be observed (Ngadiman and Yang, 1995). Still, polyps may cause nonspecific atypias of squamous and endocervical cells. Small, keratinized squamous cells with fairly large pyknotic nuclei may be occasionally observed in cervical smears corresponding to areas of somewhat atypical squamous metaplasia on the surface of the polyp. In rare instances, the pressure of the polyp on adjacent endocervical epithelium may cause a florid repair reaction (Fig. 10-10A,B). Upon removal of the polyp, the cytologic abnormalities promptly disappear. Nevertheless, this uncommon occurrence represents a potentially important source of cytologic error.


Sagiroglu and Sagiroglu (1970) documented that intrauterine contraceptive devices produce a chronic inflammatory reaction in the endometrium, resulting in the presence of leukocytes and macrophages in uterine lavage and in smears obtained directly from IUDs after removal. These authors suggested that the principal contraceptive effect of these devices is related to the presence of macrophages, which are capable of phagocytosis of spermatozoa. The effect of IUDs on the endometrial cytology and smear pattern has been discussed in Chapter 8 and is discussed again in reference to endometrial pathology in Chapter 13.

Endocervical Epithelium

The mechanical effect of IUD on endocervical epithelium may result in the shedding of columnar endocervical cells with distended, vacuolated cytoplasm. Occasionally, the vacuoles may be infiltrated by polymorphonuclear leukocytes, thus mimicking cells of endometrial adenocarcinoma (see Chap. 13). The cells may be similar to those observed in some instances of squamous metaplasia, described above. The pressure of the IUD on the endocervical epithelium may also result in florid squamous metaplasia or the “repair reaction,” as shown in Figures 10-9 and 10-10. Although Sagiroglu and Sagiroglu (1970) observed innumerable macrophages in smears obtained directly from IUDs after removal, there is no evidence that the population of these cells is significantly increased in routine cervical or vaginal samples of women wearing IUDs.

Other Findings

An important finding in women wearing IUDs is the presence of bacteria, Actinomyces, discussed in detail in the second part of this chapter and observed mainly in women wearing devices made of plastic for 3 or more consecutive years without replacement. The presence of amoebae in smears of IUD users was described by Arroyo and Quinn (1989) and by DeMoraes-Ruehsen et al (1980).

An occasionally disturbing finding in cervical smears from such patients is amorphous debris that are sometimes calcified (Fig. 10-11A-C) and occasionally form small, concentrically calcified spherical bodies, akin to psammoma bodies. The latter may be surrounded by macrophages (Fig. 10-11D). Schmidt et al (1980, 1982, 1986) has shown, by electron microscopy, that the calcified debris in IUD wearers represents fragments of plastic. There are usually few problems with the identification of the calcified debris. However, because psammoma bodies are commonly associated with ovarian and sometimes endometrial cancer (see Chaps. 13 and 15), the presence of debris mimicking psammoma bodies calls for a precautionary examination of the patient after removal of the IUD. The most commonly applied procedure is ultrasound of the pelvic organs.

A number of studies, beginning with the study by Melamed et al (1969) and subsequently repeatedly confirmed (Boyce et al, 1972; Ory et al, 1975; Sandmire et al, 1976) noted that women using oral contraceptives or wearing IUD were at a higher risk of neoplastic cervical lesions than women using barrier contraceptives, such as a diaphragm. These differences may be caused by exposure to human papillomaviruses, as discussed in Chapter 11. Therefore,
the screening and evaluation of smears from patients wearing IUDs must be thorough and careful, and any abnormalities that cannot be clearly attributed to the device itself should be further evaluated and investigated.

Figure 10-11 Effects of intrauterine contraceptive devices (IUD). A,B. Calcified debris characteristic of this condition. C. Calcified debris with a foreign body giant cell. D. A very small calcified fragment with concentric calcification, surrounded by macrophages, shown under high magnification. This structure, although very small, was similar to a psammoma body. There was no evidence of neoplasm on careful examination.


The widespread use of endocervical brush instruments had for its purpose securing cells from the endocervical canal to insure adequacy of sampling. However, rigorous use of these instruments may also result in cytologic abnormalities that may cause difficulties of interpretation. The resulting smear may contain thick clusters of endocervical cells, sometimes of complex configuration (such as loose peripheral cells, a phenomenon known as “feathering”) that may mimic changes attributed to endocervical carcinoma (Fig. 10-12A-C). It is possible that some of the cellular changes attributed to various abnormalities of the uterine endocervix, such as tubal metaplasia or microglandular hyperplasia are actually brush-induced artifacts. These clusters may persist in liquid preparations and may also cause problems of interpretation. The fundamental principle of cytopathology requires that all cell abnormalities in smears must find their counterpart in histologic material. Lesions incidentally found in histologic material are not necessarily the source of cytologic abnormalities and vice versa, without appropriate documentation.

Wilbur (1995) also pointed out that endocervical brushings present the observer with numerous difficult-to-interpret cell images that may be mistaken for various neoplastic lesions. Babkowski et al (1996) pointed out that the most complex cell clusters are derived from the upper reaches of the endocervical canal and maybe associated with tubal metaplasia (see above). Such clusters are occasionally too dense or too complex to interpret as normal and may either lead to unnecessary biopsies or result in a request for additional sampling. If the additional sampling is performed before the brush-induced injury to the cervical epithelium has healed (6 to 12 weeks), the resulting atypia of repair may cause additional interpretative difficulties (Fig. 10-12D). In my experience, the interpretation of thick, 3-dimensional endocervical cell clusters should be conservative, unless the smear also contains abnormal cells singly or in small clusters that are easier to evaluate. It is rare for a neoplastic
lesion to occur in the form of tightly knit clusters of endocervical cells, without some ancillary evidence of disease, as discussed in the appropriate chapters.

Figure 10-12 Artifacts induced by brushes. A. A very thick, large cluster of endocervical cells difficult to analyze. B. A very large sheet of endocervical cells difficult to interpret. C. Thick clusters of endocervical cells with “feathering” mimicking the appearance of endocervical adenocarcinoma. D. Marked endocervical repair reaction in a smear obtained 2 weeks after previous brushing.

Another consequence of rigorous brushings is the presence of normal endometrial cells in the sample. Such small cells, when seen in the endocervical sample may be mistaken for small cancer cells. This issue is discussed further in Chapter 8.

In general, the endocervical brush instruments, hailed as an important advance in cervicovaginal cytology because they insure sampling of the endocervical canal, are also a source of potential diagnostic errors.



Inflammatory processes within the female genital tract may be caused by infections with a variety of microorganisms and parasites or by physical and chemical agents. Sometimes the causes of the inflammation remain unknown as in Behçet’s disease (Sakane et al, 1999). In this chapter, the basic mechanism of inflammation and diseases caused by bacterial, fungal, viral, and parasitic agents will be described. The physical and chemical agents are discussed in Chapter 18. Most of the organisms responsible for inflammatory processes may be recognized in Papanicolaou-stained smears although, occasionally, special stains or procedures may be required for identification. The principal infectious agents are as follows:

Bacterial agents

Cocci and coccoid bacteria

Gram-positive cocci: species of Streptococcus and Staphylococcus

Gram-negative cocci: Gonococcus

Gardnerella vaginalis (Haemophilus vaginale or vaginalis)


Calymmatobacterium granulomatis Donovan (granuloma inguinale)

Mycoplasma and Ureaplasma

Chlamydia trachomatis

Acid-fast organisms: Mycobacterium tuberculosis, Mycobacterium avium


Spirochaeta pallida (syphilis)

Organisms that are normally saprophytic but may be associated with infections:

Lactobacillus (Döderlein bacillus) and Leptothrix

Other uncommon bacterial agents

Fungal agents

Candida species: C. albicans (monilia), C. glabrata (Torulopsis glabrata)

Aspergillus species



Cryptococcus species


Viral agents

Herpesvirus types I, II, VIII


Human polyomavirus



Molluscum contagiosum (vulva)

Human papillomavirus, various types (see Chap. 11)

Other rare viruses

Parasitic infections and infestations


Trichomonas vaginalis

Entamoeba histolytica

Entamoeba gingivalis

Balantidium coli

Helminths (worms)

Schistosoma haematobium, S. mansoni, S. japonicum


Intestinal worms

Enterobius vermicularis (pinworm)

Trichuris trichiura (whipworm)

Teniae (flat worms)

T. solium (intermediate host: swine)

T. saginata (intermediate host: cattle)

T. echinococcus (intermediate host: dog)

Other uncommon parasites


With the spread of the acquired immunodeficiency syndrome (AIDS), in which normal immune defenses of the human body are reduced or abolished, unusual organisms (not listed above) may be encountered.


The female genital tract is a common site of inflammatory processes that may involve the vulva, vagina, uterine cervix, endometrium and spread to organs located in the bony pelvis, such as the fallopian tubes, ovaries, and parametria. Many, although not all, of the agents causing inflammation may be sexually transmitted (Borchardt and Noble, 1996).

The predisposing factors are: not fully mature squamous epithelium of the vagina and cervix, such as that seen in prepubertal girls or in menopausal women; injury to the endocervical canal, especially during pregnancy or delivery; or trauma of any kind. The variations of the vaginal acidity (pH) are also significant. Alkaline pH, such as observed during the menstrual flow, favors the growth of the common parasite, Trichomonas vaginalis, which does not prosper in an acid pH of less than 5. Other microorganisms have other pH requirements.

Three basic pathways of infection of the genital tract are recognized:

  • Direct invasion of the genital tract by pathogens, often sexually transmitted

  • Spread of an infectious process from an adjacent organ

  • Blood-borne infections

Regardless of the pathway or the causative organism, all infectious processes may lead to an acute or chronic inflammatory reaction. Although any component of the female genital tract can be affected by an inflammatory process, certain agents may favor one type of tissue to another. For example, Trichomonas vaginalis infestation affects mainly the squamous epithelium of the vagina, cervix, and urethra. Many pus-producing bacteria, such as staphylococci and streptococci, find favorable conditions for survival in the endocervical canal, whence the infection may spread into the endocervical glands. From the cervix, the infectious agents may ascend the endocervical canal and reach the endometrium, the fallopian tubes, and thence the pelvic organs, causing pelvic inflammatory disease. Other organisms may have a different behavior pattern, as will be discussed below. Because the epithelial changes accompanying inflammation may result in considerable cytologic atypias, they are reported in some detail.

Sequence of Events in Inflammatory Processes

An inflammatory process is a complex reaction of the living tissue to various forms of injury that may be caused by a variety of agents, many of them listed above. An excellent summary of the current state of knowledge pertaining to mechanisms of inflammation can be found in a paper by Luster (1998). Only a simple summary of the key events is provided here.

Acute Inflammation

The first event in an acute inflammatory process is injury to the tissue. The common cause of injury is pathogenic organisms or physical and chemical agents. The injury causes cell death or necrosis which attracts polymorphonuclear leukocytes, principally neutrophiles, that invade the affected tissue exiting from the dilated regional capillary vessels. This function is governed by chemotactic agents known as chemokines (Luster, 1998). The role of the neutrophiles is to phagocytize, neutralize and destroy the agent(s) causing cell death. There are several consequences of this initial sequence of events. The capillary vessels may be injured and blood may seep into the affected area. The neutrophiles often die while performing the initial defensive role and release into the surrounding tissues a number of proteolytic enzymes that damage the tissues further, increasing the area of necrosis. The second echelon of defenses is vested in B and T lymphocytes and in macrophages,
which enter the area of injury. The macrophages are activated to phagocytize the debris and eliminate the damages. In some parasitic infestations, eosinophilic leukocytes may play a key role.

There are several possible outcomes of the initial, acute inflammatory process:

Jun 8, 2016 | Posted by in PATHOLOGY & LABORATORY MEDICINE | Comments Off on Benign Disorders of the Uterine Cervix and Vagina

Full access? Get Clinical Tree

Get Clinical Tree app for offline access