Current concepts in cervical pathology.
Abstract: Context. -- The correct diagnosis and reporting of cervical in situ and invasive carcinoma are essential for the appropriate clinical management of patients with human papillomavirus-associated disease.

Objectives. -- To review common mistakes made in the diagnosis of cervical dysplasia and invasive carcinoma, describe variants and benign mimics of high-grade squamous intraepithelial lesion and adenocarcinoma in situ, and discuss available ancillary studies that can be useful in making the distinctions as well as to review important factors related to prognosis that should be included in the pathology report.

Data Sources. -- Review of current literature.

Conclusions. -- There are many mimics and variants of cervical squamous and glandular lesions that can be resolved with ancillary studies and careful histologic examination. Prognostically important features, such as tumor size, presence of vascular invasion, and margin status, should always be included in the pathology report.
Article Type: Report
Subject: Cervical cancer (Diagnosis)
Cervical cancer (Causes of)
Pathology (Practice)
Papillomavirus infections (Health aspects)
Pathologists (Practice)
Tumor staging (Evaluation)
Authors: Park, Kay J.
Soslow, Robert A.
Pub Date: 05/01/2009
Publication: Name: Archives of Pathology & Laboratory Medicine Publisher: College of American Pathologists Audience: Academic; Professional Format: Magazine/Journal Subject: Health Copyright: COPYRIGHT 2009 College of American Pathologists ISSN: 1543-2165
Issue: Date: May, 2009 Source Volume: 133 Source Issue: 5
Topic: Event Code: 200 Management dynamics
Geographic: Geographic Scope: United States Geographic Code: 1USA United States
Accession Number: 230151994
Full Text: Cervical cancer is one of the most common cancers in women worldwide with about 493000 new cases each year resulting in 274 000 deaths globally. (1) In the United States it is the 14th most common cancer in women and affects nearly 12 000 women each year causing approximately 4000 deaths. (2) Most cervical carcinomas are etiologically related to the human papillomavirus (HPV) and about 70% of all cervical carcinomas are caused by 2 types of high-risk HPV, 16 and 18. (3) Once infected by HPV, cervical neoplasia undergoes a stepwise progression starting from preinvasive lesions that can be detected by screening and cured with complete excision. Therefore, the detection of such lesions is critical in the appropriate management of patients. Although it is important not to miss lesions, the overdiagnosis of benign atypias also has important clinical implications and can result in undue physical and psychologic distress, as well as waste valuable resources and funds. Pathologists are faced with many diagnostic difficulties when assessing cervical tissue, particularly in distinguishing benign from in situ disease and in situ from invasive carcinoma. This review focuses on variants and mimics of in situ squamous and glandular lesions, important prognostic factors in invasive disease that must be reported by the pathologist, and ancillary tests that are available to assist in the diagnosis of HPV-related lesions.

HUMAN PAPILLOMAVIRUS

Human papillomavirus preferentially infects cells in the cervical transformation zone, an area of active cell turnover. Basal cells, which feature the HPV receptor, are a natural target for infection. The virus remains in the cell in a latent state until activation causes viral replication and squamous dysplasia. Low-grade squamous intraepithelial lesion/cervical intraepithelial neoplasia 1 (LSIL/CIN 1) is characterized by koilocytotic atypia, nuclear enlargement, hyperchromasia, and perinuclear cytoplasmic clearing. These features are the result of viral proteins that affect DNA synthesis and the structure of intermediate filaments in the host cell cytoplasm. Human papillomavirus gene expression is tightly controlled in LSILs, which result from productive infections of cells that have begun the process of differentiation. High-grade squamous intraepithelial lesions/cervical intraepithelial neoplasia 2 to 3 (HSILs/CIN 2-3) do not support HPV genomic replication and, therefore, do not usually have the nuclear or cytoplasmic characteristics of an active viral infection.

Traditionally, cervical squamous lesions were thought to progress in a stepwise fashion from low grade to high grade, the theory being that a focus of HSIL arises in, gradually expands, and finally replaces LSIL as a monoclonal event. However, studies have shown that HSIL may also develop independently without progression or transformation from LSIL; instead, they likely arise de novo from epithelium adjacent to LSIL. (4)

Accurate histologic and cytologic classification of HPV-mediated lesions has important implications for guiding patient management, but studies have shown that interobserver and intraobserver reproducibility of diagnoses of cervical lesions is suboptimal. The atypical cells of undetermined significance (ASCUS)/LSIL Triage Study (ALTS)5 showed that variability of histologic diagnoses on biopsy is as great as with cytologic smears, even though histology is still considered the gold standard. Between the clinical center and central quality control groups, there was only moderate concordance, with the quality control group more frequently rendering low-grade diagnoses. CIN 1 had only 42% concordance, whereas the diagnostic extremes had much better concordance, 91% for normal and 77% for high grade. Most LSILs and even HSILs undergo spontaneous regression within 6 to 12 months due to host cell immune response. However, there is no way to prospectively differentiate among lesions that are destined to regress, persist, or progress to carcinoma. Overdiagnosis of benign lesions as LSIL adds unnecessary cost and wastes resources, as clinicians will often proceed directly to excisional biopsy for LSIL, especially in older patients. The patient's physical and psychologic distress should also not be underestimated.

SQUAMOUS LESIONS

The most common diagnostic challenges occur in distinguishing LSIL from benign reactive atypia, CIN 2 from CIN 3, and CIN 3 from atrophy and immature metaplasia.

Mild nuclear cytologic changes, especially those that occur in a background of inflammation, can be misinterpreted as LSIL. Normal squamous epithelium, with prominent cytoplasmic glycogen or vacuolization, is also commonly mistaken for koilocytosis. But true koilocytosis is often focal, whereas in normal epithelium there is no clear demarcation between abnormal and normal-appearing epithelium. Benign epithelium lacks nuclear enlargement, multinucleation (although binucleate cells are sometimes encountered), cellular disorganization, and maturation disturbances. Probably the most important feature is the lack of true nuclear atypia (Figure 1, A and B).

The distinction between CIN 2 and CIN 3 is not as crucial, because both lesions are in the high-grade category. Studies have shown that the subjective criteria used for separating the different grades and the fact that the thickness of the immature basaloid cells can vary within the same lesion lead to failures in reproducible diagnosis of CIN 2. (6)

The importance of the difference between CIN 3 and atrophy or metaplasia is obvious. Atrophy and immature metaplasia do not have the mitotic activity, nuclear pleomorphism, loss of cell polarity, nuclear hyperchromasia, or clumped chromatin of HSIL. Reparative change can also sometimes mimic HSIL with atypical basal cells occupying the lower half of the epithelium. These cells have regular nuclear contours, prominent nucleoli, and distinct cell borders, often accompanied by dense acute or chronic inflammation.

SQUAMOUS VARIANT LESIONS

Variants of squamous dysplasia that appear reactive or metaplastic may also cause diagnostic dilemmas. These include atypical immature metaplasia (AIM) and eosinophilic dysplasia (ED).

Atypical immature metaplasia is a loosely defined term for immature metaplastic cells with mild cytologic atypia, first described by Crum et al in 1983. (7) By their description, AIM was similar to condyloma, that is, low-grade lesion. But since then, there has been a shift in the diagnosis to include both high-grade and low-grade lesions. (8-10) This attests to the fact that this entity spans a morphologic spectrum and likely includes many diagnostically challenging cases with an immature appearance. There is lack of a uniform definition, which has resulted in inconsistencies in morphologic criteria. Several studies have shown that AIM can be HPV positive or negative, be variably positive for p16, and display a large range of proliferative rates with Mib-1 (8-10) (see section on ancillary studies on page 735). Follow-up studies have reported that HPV-associated AIM showing p16 expression and high proliferative rates subsequently recur as a bone fide dysplastic lesion in many instances. It has been hypothesized that AIM lesions that are HPV and p16 positive but hypoproliferative with Mib-1 may be early or regressing lesions. Atypical immature metaplasia is not commonly used as a diagnostic term in surgical pathology reports. Our group currently diagnoses HSIL when we encounter AIM lesions that diffusely express p16 and are hyperproliferative with Mib-1. Immunohistochemical stains that do not support a proliferative, HPV-associated lesion usually result in a report of AIM with a note describing the significance of the finding.

Eosinophilic dysplasia (ED), another descriptive term, has been used for intraepithelial lesions showing both squamous metaplasia and dysplasia. (11) Eosinophilic dysplasia is commonly associated with HPV infection and adjacent foci of classic HSIL. Most ED lesions are diffusely and strongly positive for p16, hyperproliferative with Mib-1, and positive for HPV DNA by polymerase chain reaction, all high or intermediate risk. Therefore, ED likely represents a variant of HSIL with misleading morphologic features that suggest a benign diagnosis. Eosinophilic dysplasia is another term that is not commonly used in surgical pathology reports. Our group currently diagnoses HSIL when we encounter ED lesions that diffusely express p16 and are hyperproliferative with Mib-1. Immunohistochemical stains that do not support a proliferative, HPVassociated lesion result in a report of atypical metaplasia with a note describing the significance of the finding.

GLANDULAR LESIONS

Adenocarcinoma in situ (AIS) is much less common than squamous intraepithelial lesion (SIL). Unlike squamous neoplasms in which precursor lesions predominate over invasive carcinomas, the opposite is true for glandular lesions. It is more difficult to screen effectively for AIS than for SIL, so AIS may not be detected before the development of invasive adenocarcinoma. Histologic features of AIS include preservation of normal glandular architecture and involvement of part or all of the epithelium by enlarged, hyperchromatic, stratified nuclei with coarse chromatin, small nucleoli, mitoses, and apoptosis. The cytoplasm can be depleted or abundant, vacuolated, granular, and basophilic or eosinophilic (Figure 2). At least 90% of AIS cases are associated with HPV (most commonly HPV 18 and 16). It coexists with SIL or squamous cell carcinoma in 50% of cases. (12-23) Part of the difficulty in identifying AIS is that the lesion is not well visualized colposcopically as it can arise high up in the endocervical canal,17 and the cytologic criteria for identifying neoplastic glandular lesions are not as well defined as for SIL. Adenocarcinoma in situ has been described as multifocal, though the definitions of multifocality are often unclear. Various studies report that AIS can be multicentric (13%) (19) and multiquadrant (53%). (21) This is important because multifocality may affect the meaning of a negative margin on cone excisions. Studies have shown that cone biopsies for AIS with negative margins result in residual disease in a hysterectomy 6% to 44% of the time, whereas cone biopsies for AIS with positive margins are associated with residual disease 44% to 75% of the time. (24) Therefore, the standard treatment for AIS is simple hysterectomy or a cold knife cone excision (not loop electrosurgical excision procedure [LEEP]) with widely clear margins (10 mm ideal according to Goldstein and Mani (25)). Therefore, it is important to always mention the status of the margins and give the closest distance from the lesion if possible. Some AISs exhibit intestinal differentiation with goblet cells and sometimes Paneth or argentaffin cells. (26) Other related lesions include adenosquamous carcinoma in situ (27) and stratified mucin-producing intraepithelial lesion (SMILE), described subsequently.

[FIGURES 1-5 OMITTED]

AIS MIMICS

Glandular lesions that can mimic AIS include reactive glandular atypia, tubal/tuboendometrioid metaplasia, and so-called glandular dysplasia (GD).

Reactive glandular atypias can be secondary to inflammation, radiation, or viral infections. Glandular lesions showing cytologic atypia can also be found in the context of Arias-Stella reaction, microglandular hyperplasia, endometriosis, and mesonephric remnants. (28) In reactive changes, the nuclei are typically enlarged, have prominent nucleoli, but lack nuclear hyperchromasia. Mitotic activity is absent or minimal, as is pseudostratification. Papillary endocervicitis, featuring exaggerated papillary projections and a dense infiltrate of acute and/or chronic inflammation, can also occur. (29) Radiation change causes the nuclei and cytoplasm to enlarge, often with vacuolization, but with only a slight increase in the nuclear to cytoplasmic ratio. The nuclei can look round, look vesicular with prominent nucleoli, and rarely have a smudged hyperchromatic appearance. (30)

Tubal metaplasia tends to occur in the upper endocervix. It is usually not difficult to distinguish from AIS because ciliated, intercalated (peg), and secretory cells, more characteristically found in the fallopian tube, are present at least focally in tubal metaplasia. Atypical tubal metaplasia is a form of tubal metaplasia in which the glands are lined by similar tubal type cells, but these are crowded and have larger, more hyperchromatic nuclei and are pseudostratified. Mitosis and apoptosis are not common like they are in AIS, but the degree of atypia can be mistaken for AIS (Figure 3). Tuboendometrioid metaplasia is essentially tubal metaplasia without ciliated cells, although it commonly displays more pseudostratification and higher nuclear to cytoplasmic ratios than other tubal metaplasia lesions; it lacks the architectural complexity, mitoses, apoptosis, and nuclear irregularity of AIS (Figure 4). Tubal and tuboendometrioid metaplasia can involve the deep cervical wall with dilated irregular gland formation and stromal reaction around the glands; this latter feature, in particular, may erroneously suggest a diagnosis of invasive adenocarcinoma (31) (Figure 5). Careful examination of the nuclear features should allow one to properly differentiate between benign and malignant.

Glandular dysplasia has been defined in a variety of ways, but in general, GD reportedly displays a lesser degree of abnormality than AIS. (32-35) The diagnostic criteria remain largely subjective. The most objective criteria for diagnosing noninvasive glandular lesions of the cervix were presented in a report by Ioffe et al (36) in which a numerical scoring system was proposed. In this scoring scheme, separate scores are assigned for the presence and degree of (1) nuclear atypia, (2) stratification, and (3) mitoses/apoptosis (counted in the 2 most active glands and the average number used). These 3 scores are then added to result in the total score (0-3, benign; 4-5, endocervical GD; 6-9, adenocarcinoma in situ). Although this method improves interobserver discordance, the biologic and clinical significance of GD remains to be determined. The association between GD and HPV infection is variable, and, as it only infrequently coexists with AIS, it has been suggested that GD might not represent an AIS precursor at all. It is therefore imprudent to make an unqualified diagnosis of GD. (24) Application of ancillary tests, discussed subsequently, can be contributory.

GLANDULAR VARIANT LESIONS

SMILE, first described by Park et al in 2000, (27) is an HPV-associated intraepithelial columnar cell neoplasm that is thought to arise from the reserve cells of the transformation zone. Morphologically, SMILEs are composed of immature stratified cells that, although appearing similar to SILs, display intracytoplasmic mucin or cytoplasmic vacuoles that separate the cells in the mid to lower layers of the epithelium (Figure 6). Overt gland formation, seen in typical AIS, is not found in SMILE. The most consistent feature of SMILE is the spacing of nuclei by mucin, which can be highlighted with a mucicarmine stain. SMILEs are commonly associated with HSIL, AIS, adenocarcinoma, squamous carcinoma, and adenosquamous carcinoma. The natural history of these lesions has not been fully elucidated, but it has been reported that lesions may recur in the form of SIL or AIS. As is the case with AIM and ED, SMILE should not be used as a stand-alone term in surgical pathology reports. Either AIS or adenosquamous carcinoma in situ can be used, depending on the context and the pathologist's preference.

INVASIVE CARCINOMA

The diagnosis of invasive squamous or adenocarcinoma is usually not challenging. Stage is the most important prognostic factor in cervical cancer, but unlike the practice in other organ systems, staging is primarily assessed by physical examination. In developed countries, imaging modalities may also be used, but physical examination is still the gold standard in countries where these tests are not widely available. Grossly inapparent carcinoma (carcinoma detected by Papanicolaou test and biopsy not by physical examination) is assessed on LEEP or cone biopsy, where stage is assigned by microscopic measurements of the extent of invasive carcinoma. Occult spread of disease may be found at the time of surgery, but this does not change the initial stage (though it may change the treatment plan). Although various staging systems exist, the International Federation of Gynecology and Obstetrics (FIGO) system is almost universally used (Table). There is a clear difference in survival depending on stage with near 99% 5-year survival for stage IA1 to 65% for IIB and 43% for IIIB. (37)

In pathology reports for LEEPs and cone biopsies, the size of the invasive tumor, including depth of stromal invasion and horizontal tumor spread, should always be included, as well as the presence of lymphovascular invasion (LVI) and margin status. LEEP and cone specimens should be inked with differential coloring of the endocervical margin, if possible.

Lymphovascular invasion has been shown to be an independent risk factor in relation to disease-free survival in a Gynecologic Oncology Group (GOG) prospective study of 645 patients (disease-free survival, 89% vs 77% without and with LVI, respectively). (38) However, reproducible and accurate diagnosis of LVI can be problematic. The most difficult issue concerns artifactual retraction or shrinkage of stroma around invasive nests that results in spaces that resemble vessels. Immunostains for vascular markers, such as CD31, CD34, and lymphatic marker D240 (podoplanin) can help if the focus is present in subsequent sections. (39) Despite problems with recognizing LVI, most studies have shown that there is a relationship between the depth of invasion and the presence of LVI and that LVI is a predictor of residual invasive disease in the hysterectomy specimen. (40) For some gynecologists, the presence of LVI in the context of microinvasive carcinoma (FIGO stage IA1) triggers radical resection instead of conservative management.

[FIGURES 6-8 OMITTED]

In a prospective study from the GOG by Sedlis et al, (41) patients with stage IB cervical carcinomas were randomized to pelvic radiotherapy following radical hysterectomy and lymphadenectomy if their tumors fulfilled 2 of the following 3 criteria: large clinical tumor size, deep (greater than one-third) stromal invasion, and LVI. They found that patients receiving adjuvant radiotherapy suffered fewer recurrences than controls. Surgical pathology reports of radical hysterectomy specimens should therefore include the same information as reported for LEEP and cone biopsies in addition to the percentage invasion into the cervical wall, the gross size of the tumor, and the status of the parametrium.

The distinction between squamous and adenocarcinoma might have clinical relevance, as it has been postulated that adenocarcinomas and adenosquamous carcinomas have a worse prognosis than pure squamous carcinomas, with higher rates of lymph node metastases. However, a GOG study looking at 626 patients with locally advanced disease treated with radiation showed no difference based on histologic cell type. (42) The impact of tumor grade on prognosis has also been debated. Another GOG study by Zaino et al (40) showed that in squamous carcinomas grade has little prognostic value. On the other hand, tumor differentiation may have a significant role in adenocarcinoma. (43) There are many different types and variants of primary cervical carcinomas, but this review is limited to commonly encountered squamous cell and adenocarcinomas. A partial list of other carcinomas, some of which have important clinical differences with the tumors discussed herein, include verrucous carcinoma, lymphoepithelial-like carcinoma, intestinal adenocarcinoma, gastriclike adenocarcinoma, endometrioid adenocarcinoma, clear cell and serous adenocarcinoma, minimal deviation adenocarcinoma, mesonephric adenocarcinoma, glassy cell carcinoma, mucoepidermoid carcinoma, adenoid cystic carcinoma, adenoid basal carcinoma, large cell and small cell neuroendocrine carcinoma, and mixed epithelial and mesenchymal tumors.

Correctly measuring the extent of invasion is crucial for staging cervical cancer, especially in early-stage (microscopically detected) disease because stage dictates treatment and prognosis. However, superficial invasion is frequently overcalled. Of 265 purported cases of microinvasion submitted to a group of reference pathologists of the GOG, about 50% were rejected. (44) Pitfalls of misdiagnosing SIL with microinvasion include SIL involving endocervical glands and areas of prior biopsy site with entrapped SIL. Areas of prior biopsy will have associated acute and chronic inflammatory infiltrate and may have entrapped neoplastic epithelium buried within the stroma as a result of disrupted SIL from a punch biopsy. Definitive examples of invasive squamous carcinoma should demonstrate stromal desmoplasia, irregular or scalloped margins, and/or abundant, paradoxically mature eosinophilic cytoplasm.

Once the presence of invasion is confirmed, measuring the depth of invasive squamous carcinoma is the next important step. For consistency, the depth of stromal invasion should be measured following these guidelines: the depth of neoplastic projections should be measured from the initial site of invasion, either from the basal lamina of the surface epithelium or from the endocervical glands replaced by intraepithelial lesions (Figure 7); in cases in which a direct histologic continuum between invasive foci and SIL cannot be demonstrated, it is assumed that the invasion originated from the basal cells of the overlying SIL. Horizontal extent of the invasive carcinoma only needs to be measured in cases in which the tumor depth is 5 mm or less because once the tumor invades greater than 5 mm, it is stage IB regardless of the extent of horizontal spread. FIGO includes horizontal measurement in staging but does not delineate how to measure it. When there is only unifocal invasion, measuring horizontal extent is not difficult. However, invasive carcinoma can often be multifocal. There is no absolute method for how to best measure the horizontal extent of invasion in these circumstances, but because tumor volume is thought to be an important factor in prognosis (45, 46) with higher rates of lymph node metastases and recurrence in women with greater tumor volume, it may be best to measure in a way that most accurately approximates volume. In an article by Reich and Pickel, (47) they describe 3 patterns of microinvasion and how each of those should be measured:

Type I -- multiple discrete foci of invasion are contiguous with surface epithelium and the origin of invasion can be detected. In this case, measure each invasive focus separately and then add together for a total horizontal length. The HSIL in between the invasive foci is not measured.

Type II -- the origin of invasion cannot be seen in all of the early invasive foci but the greatest lateral extent can be seen in a single section. The breadth of the lesion is considered to lie between the 2 most lateral foci counting everything, including normal and HSIL, in between.

Type III -- like type II, except that the lesion is present in multiple contiguous sequential sections; the greatest horizontal length cannot be measured in a single slide. In this case, the width must be calculated by measuring the distance between the sections; this is usually assumed to measure 2 to 3 mm per section.

Therefore, the growth pattern of the invasive foci is important and the entire cervix must be embedded and evaluated when the lesion is microscopic. Examining multiple levels may also be useful.

The nomenclature used to describe bona fide examples of superficially invasive carcinomas is, unfortunately, not standardized. Proposals for definitions of microinvasion or microinvasive carcinoma include examples showing invasion less than 3 mm deep or those with invasion less than 5 mm deep, with or without LVI. The horizontal extent of the invasive front is also inconsistently mentioned as a criterion. The term has also been used, historically, more frequently in reference to squamous cell carcinoma than adenocarcinoma, as discussed subsequently. In practice, therefore, we do not advocate the use of the term microinvasive carcinoma. Instead, we report the depth of invasion, measure the horizontal extent of the invasive component, and mention whether LVI is present. If it is a pathologist's practice to use the term microinvasive carcinoma it is recommended that the term only be used when definitive measurements can be made (and surgical margins are negative for invasive carcinoma) and the patient's gynecologist is informed about the criteria the pathologist uses for such tumors.

Although the FIGO staging system for cervical cancer applies to both squamous and adenocarcinoma, there is considerable controversy as to whether microinvasive adenocarcinoma exists as a histologically recognizable entity. One reason is that the irregular distribution and architecture of the normal endocervical crypts in the cervical stroma make it difficult to differentiate between early stromal invasion and AIS. There are 2 unequivocal features of invasion in adenocarcinoma: individual cells/fragmented glands/incomplete glands lined by malignant cells at a stromal interface and malignant glands with desmoplastic stromal response. Other less definitive "soft" features that are suggestive of early invasion include complex architectural growth with papillae, cribriforming and irregular glands growing in a confluent or labyrinthine pattern, malignant glands adjacent to medium- or thick-walled vessels, and the presence of malignant glands below the level of normal glands. Normal cervical glands extend to the inner one-third of the cervical wall and can measure up to 1 cm. Nabothian cysts, tunnel clusters, endocervical hyperplasia, deep endocervical glands, and mesonephric duct remnants can all extend to the outer third of the cervical wall.

Admittedly, in about 20% of the cases, it is impossible to distinguish between AIS and early invasive adenocarcinoma. (48) Because there is no overlying basement membrane from which to measure the starting point of invasion, as there is in squamous carcinoma, measuring the exact depth of invasion of cervical adenocarcinoma is problematic. Most sources, including the World Health Organization, (49) recommend measuring the thickness of the tumor from the mucosal surface, rather than depth of invasion as defined previously.

Some cervical adenocarcinomas, particularly villoglandular adenocarcinomas, and many papillary squamous cell carcinomas can grow as exophytic, polypoid masses that are obvious at physical examination. Substantial architectural complexity and the presence of a gross, massforming lesion qualify for invasive carcinoma (as opposed to HSIL or AIS), but measurement of the depth of invasion can be problematic because these lesions frequently grow up into the endocervical canal, instead of down into the cervical stroma. The depth of stromal invasion here should not reflect thickness, rather it should be measured from a plane that underlies the exophytic tumor. Regardless of this measurement, these tumors are regarded as at least FIGO stage IB because they are easily detected on physical examination.

ANCILLARY STUDIES

Various immunohistochemical and molecular assays are available as ancillary studies in the workup of difficult squamous and glandular lesions.

p16INK4a is a tumor suppressor gene that encodes a protein involved in cell cycle regulation. It acts in a negative feedback loop with another tumor suppressor, retinoblastoma protein, so that cell proliferation is held in check. When high-risk HPV DNA integrates into the host cell genome, the viral oncoprotein E7 binds to retinoblastoma protein rendering it inactive. The feedback loop is lost and p16 is overexpressed as a result. This manifests as diffuse, strong, cytoplasmic and/or nuclear staining in squamous and glandular lesions associated with high-risk HPV infection (Figure 8, A and B). (50-53) Interpreting p16 immunostaining, however, can be complicated, as both the context in which p16 is applied and also the distribution and intensity of its staining pattern must be considered before using the stain to confirm the presence of an HPV-associated neoplasm. A variety of lesions unassociated with HPV can show focal or patchy p16 staining, including tubal metaplasia, squamous metaplasia, and endometrioid adenocarcinoma (Figure 9). Serous carcinomas can show diffuse, strong staining as well. (51, 54, 55) Atypical immature metaplasia lesions that are p16 positive have been shown to have a higher incidence of subsequent HSILs and are likely HSIL lesions that have an immature metaplastic appearance. (9) Eosinophilic dysplasias are also usually p16 positive, supporting the contention that these are variants of dysplastic lesions arising from cervical reserve cells. (11) Atypical, proliferative glandular lesions that are diffusely and strongly p16 positive are also likely HPV-associated adenocarcinomas as opposed to one of the reactive lesions discussed earlier. The diffuse staining pattern of HPV-associated glandular lesions can be especially helpful in curettage specimens where the differential is often with endometrial endometrioid adenocarcinomas. A useful panel in this situation includes p16, estrogen and progesterone receptors, monoclonal carcinoembryonic antigen (CEAM), and vimentin. (56, 57) Although overlapping patterns are seen, the general trend is for cervical adenocarcinomas to be positive for CEA(m) and p16 and negative for estrogen and progesterone receptors and vimentin, whereas the opposite is likelier in endometrial endometrioid adenocarcinomas.

[FIGURES 9-10 OMITTED]

Mib-1 is a monoclonal antibody that reacts with Ki-67, a nuclear cell proliferation-associated antigen that is expressed in all active parts of the cell cycle ([G.sub.1], S, [G.sub.2]/M). Positive staining in parabasal cells is found under normal conditions, so interpreting the results of this stain rests on examination of the middle and upper thirds of the squamous epithelium. High-grade squamous intraepithelial lesion (CIN 2 and CIN 3) usually shows diffuse nuclear positivity scattered throughout all layers of the epithelium, (58) whereas the distribution of Mib-1 staining in CIN 1 can be less diffuse with only small clusters of squamous cells staining in the upper two-thirds of the epithelium. One should be careful not to overinterpret positive staining in the presence of inflammation because intraepithelial lymphocytes will be positive for Mib-1 and inflammation can also cause reactive proliferation. It should also be noted that tangential sectioning sometimes results in the impression that Mib-1-positive parabasal cells are superficially placed, leading to confusion with a SIL (59, 60) (Figure 10, A and B). Adenocarcinoma and AIS should show strong diffuse staining with Mib-1, whereas benign mimics such as tubal metaplasia, radiation change, and reactive atypia should show only focal or patchy staining at most. (61, 62) SMILE should also show diffuse positivity of the entire thickness of the lesion, including the mucin-producing cells. (27) Atypical immature metaplasia, when diffusely positive for Mib-1, is likely an immature variant of HSIL. (9)

ProEx C is a recently developed immunohistochemical assay that targets the expression of topoisomerase II-[alpha] and minichromosome maintenance protein-2, two genes that have been shown to be overexpressed in cervical cancers. (63-70) The assay is a nuclear stain that is positive in cervical dysplasia and has been validated in cytologic specimens for the detection of HSIL. (66, 67) Studies have also shown that ProEx C is comparable to p16 and Mib-1 in the detection of high-grade lesions in formalin-fixed tissue sections and in distinguishing them from benign mimics. (68-70) The staining pattern of ProEx C in histologic sections is similar to that of Mib-1 in that only nuclear positivity above the basal one-third layer is considered positive. In a recent study by Pinto et al (70) it was shown that the combination of p16 and ProEx C predicted more nonSILs than p16 combined with Mib-1. This suggests that ProEx C may be more efficient than Mib-1 for distinguishing reactive epithelial changes from squamous lesions.

In situ hybridization (ISH) is a direct signal detection assay that allows visualization of HPV DNA within infected cells. Unlike other molecular detection methods (eg, polymerase chain reaction), the ISH assay can be automated and does not require the skills of highly trained laboratory personnel or the stringent laboratory setup required to prevent contamination. The interpretation of ISH is similar to that of immunohistochemistry staining in tissue sections (look for positive nuclear staining). The ISH signal patterns of HPV DNA have been associated with the physical status of HPV in infected cells, that is, episomal or integrated forms. (71, 72) Episomal forms result in blocklike nuclear labeling, whereas integrated forms result in punctate, nuclear signals. Punctate signals are most frequently found in HSILs and invasive carcinomas but not LSILs. Because the integration of oncogenic HPV into the human genome is a critical step for cervical cancer carcinogenesis, the signal pattern could be a useful marker for predicting precancerous lesion progression. In a recent study by Guo et al, (73) the authors showed that the Inform HPV III (Ventana Medical Systems, Tuscon, Arizona) in situ hybridization assay is comparable to polymerase chain reaction and may be a useful adjunct in detecting HPV-positive lesions and carcinomas. However, ISH has also been shown to have low sensitivity so its utility has been questioned. (74)

CONCLUSIONS

There are many common difficulties that arise in the daily assessment of cervical specimens. Practicing pathologists should not overcall reactive squamous atypia as squamous dysplasia and they should be aware of variants of squamous lesions and AIS that may appear metaplastic or reactive (AIM, ED, SMILE). Stage is the most important prognostic factor for cervical cancer, but other pathologic indices are also important. One should include all prognostically relevant information in pathology reports for best patient care, including the extent of invasion, margin status, and the presence of LVI. The depth of invasion should be measured from the nearest overlying basement membrane in squamous carcinoma, but the thickness of tumor (rather than depth of invasion) should be reported in adenocarcinoma. Radical hysterectomy reports should also contain information about the greatest tumor dimension and the status of the parametria. It may be difficult to distinguish AIS from early invasive adenocarcinoma, but using the criteria discussed previously should be helpful. Finally, ancillary studies (p16, MIB-1, ProEx C, HPV in situ) can be useful in difficult cases, but they must be interpreted in the proper context.

International Federation of Gynecologists and Obstetricians Staging

Stage I--confined to the cervix

IA--diagnosed only by microscopy; no visible lesions

IA1--stromal invasion less than 3 mm in depth and 7 mm or less in horizontal spread

IA2--stromal invasion between 3 and 5 mm with horizontal spread of 7 mm or less

IB--visible lesion or a microscopic lesion with more than 5 mm of depth or horizontal spread of more than 7 mm

IB1--visible lesion 4 cm or less in greatest dimension

IB2--visible lesion more than 4 cm

Stage II--invades beyond cervix

IIA--without parametrial invasion, but involves upper two thirds of vagina

IIB--with parametrial invasion

Stage III--extends to pelvic side wall or lower one-third of the vagina, causes hydronephrosis or nonfunctioningkidney

IIIA--involves lower third of vagina

IIIB--extends to pelvic wall and/or causes hydronephrosis or nonfunctioning kidney

Stage IV--extends beyond true pelvis or clinically involves the mucosa of bladder or rectum

IVA--invades mucosa of bladder or rectum and/or extends beyond true pelvis

IVB--distant metastasis

Accepted for publication December 11, 2008.

References

(1.) Ferlay J, Bray F, Pisani P, Parkin DM. GLOBOCAN 2002: Cancer Incidence, Mortality and Prevalence Worldwide. IARC CancerBase No. 5, version 2.0. Lyon, France: IARC; 2004.

(2.) US Cancer Statistics Working Group. United States Cancer Statistics: 2004 Incidence and Mortality. Atlanta, GA: Department of Health and Human Services, Centers for Disease Control and Prevention, and National Cancer Institute; 2007.

(3.) Munoz N, Bosch FX, de Sanjose S, et al. Epidemiologic classification of human papillomavirus types associated with cervical cancer. N Engl J Med. 20036;348(6):518-527.

(4.) Robertson JH, Woodend B, Elliott H. Cytological changes preceding cervical cancer. J Clin Pathol. 1994;47(3):278-279.

(5.) Schiffman M, Solomon D. Findings to date from the ASCUS-LSIL Triage Study (ALTS). Arch Pathol Lab Med. 2003;127(8):946-949.

(6.) Ismail SM, Colclough AB, Dinnen JS, et al. Observer variation in histopathological diagnosis and grading of cervical intraepithelial neoplasia. BMJ. 1989; 298(6675):707-710.

(7.) Crum CP, Egawa K, Fu YS, et al. Atypical immature metaplasia (AIM): a subset of human papillomavirus virus infection of the cervix. Cancer. 1983; 51 (12):2214-221 9.

(8.) Duggan MA, Akbari M, Magliocco AM. Atypical immature cervical metaplasia: immunoprofiling and longitudinal outcome. Hum Pathol. 2006;37(11): 1473-1481.

(9.) Iaconis L, Hyjek E, Ellenson LH, Pirog EC. p16 and Ki-67 immunostaining in atypical immature squamous metaplasia of the uterine cervix: correlation with human papillomavirus detection [published correction appears in Arch Pathol Lab Med. 2008;132(1):13]. Arch Pathol Lab Med. 2007;131(9):1343-1349.

(10.) Geng L, Connolly DC, Isacson C, Ronnett BM, Cho KR. A typical immature metaplasia (AIM) of the cervix: is itrelated to high-grade squamous intraepithelial lesion (HSIL)? Hum Pathol. 1999;30(3):345-351.

(11.) Ma L, Fisk JM, Zhang RR, Ulukus EC, Crum CP, Zheng W. Eosinophilic dysplasia of the cervix: a newly recognized variant of cervical squamous intra epithelial neoplasia. Am J Surg Pathol. 2004;28(11):1474-1484.

(12.) Jaworski RC, Pacey NF, Greenberg ML, Osborn RA. The histologic diagnosis of adenocarcinoma in situ and related lesions of the cervix uteri: adenocarcinoma in situ. Cancer. 1988;61:1171-1181.

(13.) Ostor AG, Duncan A, Quinn M, Rome R. Adenocarcinoma in situ of the uterine cervix: an experience with 100 cases. GynecolOncol. 2000;79:207-210.

(14.) Denehy TR, Gregori CA, Breen JL. Endocervical curettage, cone margins, and residual adenocarcinoma in situ of the cervix. Obstet Gynecol. 1997;90: 1-6.

(15.) Poynor EA, Barakat RR, Hoskins WJ. Management and follow-up of patients with adenocarcinoma in situ of the uterine cervix. Gynecol Oncol. 1995; 57:158-164.

(16.) Andersen ES, Arffmann E. Adenocarcinoma in situ of the uterine cervix: a clinicopathologic study of 36 cases. Gynecol Oncol. 1989;35:1-7.

(17.) Bertrand M, Lickrish GM, Colgan TJ. The anatomic distribution ofcervical adenocarcinoma in situ: implications for treatment. Am J Obstet Gynecol. 1987; 157:21-25.

(18.) Muntz HG, Bell DA, Lage JM, Goff BA, Feldman S, Rice LW. Adenocar cinoma in situ of the uterine cervix. Obstet Gynecol. 1992;80:935-939.

(19.) Ostor AG, Pagano R, Davoren RA, Fortune DW, Chanen W, Rome R. Adenocarcinoma in situ of the cervix. Int JGynecol Pathol. 1984;3:179-190.

(20.) Qizilbash AH. In situ and microinvasive adenocarcinoma of the uterine cervix: a clinical, cytologic and histologic study of 14 cases. Am J Clin Pathol. 1975;64:155-170.

(21.) Shin CH, Schorge JO, Lee KR, Sheets EE. Conservative management of adenocarcinoma in situ of the cervix. Gynecol Oncol. 2000;79:6-10.

(22.) Tobon H, Dave H. Adenocarcinoma in situ ofthe cervix: clinicopathologic observations of 11 cases. Int J Gynecol Pathol. 1988;7:139-151.

(23.) Wolf JK, Levenback C, Malpica A, Morris M, Burke T, Mitchell MF. Adenocarcinoma in situ of the cervix: significance of cone biopsy margins. Obstet Gynecol. 1996;88:82-86.

(24.) Zaino RJ. Symposium part I: adenocarcinoma in situ, glandular dysplasia, and early invasive adenocarcinoma of the uterine cervix. Int J Gynecol Pathol. 2002;21(4):314-326.

(25.) Goldstein NS, Mani A. The status and distance of cone biopsy margins as a predictor of excision adequacy for endocervical adenocarcinoma in situ. Am J Clin Pathol. 1998;109:727-732.

(26.) Ostor AG, Pagano R, Davoren RAM, Fortune DW, Chanen W, Rome R. Adenocarcinoma in situ of the cervix. Int J Gynecol Pathol. 1984;3:179-190.

(27.) Park JJ, Sun D, Quade BJ, et al. Stratified mucin-producing intraepithelial lesions of the cervix: adenosquamous or columnar cell neoplasia? Am J Surg Pathol. 2000;24(10):1414-1419.

(28.) Kurman RJ, ed. Blaustein's Pathology of the Female Genital Tract. 5th ed. New York, NY: Springer-Verlag; 2002:235-241.

(29.) Young RH, Clement PB. Pseudoneoplastic glandular lesions of the uterine cervix. Semin Diagn Pathol. 1991;8(4):234-249.

(30.) Lesack D, Wahab I, Gilks CB. Radiation-induced atypia of endocervical epithelium: a histological, immunohistochemical and cytometric study. Int J Gynecol Pathol. 1996;15(3):242-247.

(31.) Oliva E, Clement PB, Young RH. Tubal and tubo-endometrioid metaplasia of the uterine cervix: unemphasized features that may cause problems in differential diagnosis: a report of 25 cases. Am J Clin Pathol. 1995;103(5):618-623.

(32.) Jaworski R. Endocervical glandular dysplasia, adenocarcinoma in situ, and early invasive (microinvasive) adenocarcinoma of the uterine cervix. Semin Diagn Pathol. 1990;7:190-205.

(33.) Brown LJ, Wells M. Cervical glandular atypia associated with squamous intraepithelial neoplasia: a premalignant lesion? J Clin Pathol. 1986;39:22-28.

(34.) Casper G, Ostor A, Quinn M. A clinico pathologic study of glandular dysplasia of the cervix. Gynecol Oncol. 1997;64:166-170.

(35.) Gompel C, Silverberg S. Pathology in Gynecology and Obstetrics. Philadelphia, PA: JB Lippincott;1994.

(36.) Ioffe OB, Sagae S, Moritani S, Dahmoush L, Chen TT, Silverberg SG. Proposal of a new scoring scheme for the diagnosis of noninvasive endocervical glandular lesions. Am J Surg Pathol. 2003;27(4):452-460.

(37.)Rock JA, Jones HW, eds. Te Linde s Operative Gynecology. In press.

(38.) Delgado G, Bundy B, Zaino R, Sevin BU, Creasman WT, Major F. Prospective surgical-pathological study of disease-free interval in patients with stage IB squamous cell carcinoma ofthe cervix: a Gynecologic Oncology Group study. Gynecol Oncol. 1990;38(3):352-357.

(39.) Urabe A, Matsumoto T, Kimura M, Sonoue H, Kinoshita K. Grading system of lymphatic invasion according to D2-40 immunostaining is useful for the prediction of nodal metastasis in squamous cell carcinoma of the uterine cervix. Histopathology. 2006;49(5):493-497.

(40.) Zaino RJ, Ward S, Delgado G, et al. Histopathologic predictors of the behavior of surgically treated stage IB squamous cell carcinoma of the cervix: a Gynecologic Oncology Group study. Cancer. 1992;69(7):1750-1758.

(41.) Sedlis A, Bundy BN, Rotman MZ, Lentz SS, Muderspach LI, Zaino RJ. A randomized trial of pelvic radiation therapy versus no further therapy in selected patients with stage IB carcinoma of the cervix after radical hysterectomy and pelvic lymphadenectomy: a Gynecologic Oncology Group Study. Gynecol Oncol. 1999;73(2):177-183.

(42.) Stehman FB, Bundy BN, DiSaia PJ, Keys HM, Larson JE, Fowler WC. Carcinoma of the cervix treated with radiation therapy: a multi-variate analysis of prognostic variables in the Gynecologic Oncology Group. Cancer. 1991;67(11): 2776-2785.

(43.) Shingleton HM, Orr JW, eds. Cancer of the Cervix. Philadelphia, PA: JB Lippincott; 1995.

(44.) Sedlis A, Sall S, Tsukada Y, et al. Microinvasive carcinoma of the uterine cervix: a clinical-pathologic study. Am JObstet Gynecol. 1979;133(1):64-74.

(45.) Burghardt E, Pickel H, Haas J, Lahousen M. Prognostic factors and operative treatment of stages IB to IIB cervical cancer. Am J Obstet Gynecol. 1987; 156(4):988-996.

(46.) Kaspar HG, Dinh TV, Doherty MG, Hannigan EV, Kumar D. Clinical implications of tumor volume measurementin stage I adenocarcinoma of the cervix. Obstet Gynecol. 1993;81(2):296-300.

(47.) Reich O, Pickel H. Multifocal stromal invasion in microinvasivesquamous cell carcinoma of the cervix: how to measure and stage these lesions. Int JGy necol Pathol. 2002;21(4):416-417.

(48.) Ostor AG. Early invasive adenocarcinoma of the uterine cervix. Int J Gy necol Pathol. 2000;19(1):29-38.

(49.) Tavassoli FA, Devilee P, eds. Tumours of the Breast and Female Genital Organs. Lyon, France: IARC Press; 2003. World Health Organization Classification of Tumours.

(50.) Ishikawa M, Fujii T, Masumoto N, et al. Correlation of p16INK4A overexpression with human papillomavirus infection in cervical adenocarcinomas. Int J Gynecol Pathol. 2003;22(4):378-385.

(51.) Riethdorf L, Riethdorf S, Lee KR, Cviko A, Lconing T, Crum CP. Human papillomaviruses, expression of p16, and early endocervical glandular neoplasia. Hum Pathol. 2002;33(9):899-904.

(52.) Ansari-Lari MA, Staebler A, Zaino RJ, Shah KV, Ronnett BM. Distinction of endocervical and endometrial adenocarcinomas: immunohistochemical p16 expression correlated with human papillomavirus (HPV) DNA detection. Am J Surg Pathol. 2004;28(2):160-167.

(53.) Keating JT, Cviko A, Riethdorf S, et al. Ki-67, cyclin E, and p16INK4 are complimentary surrogate biomarkers for human papillomavirus-related cervical neoplasia. Am J Surg Pathol. 2001;25:884-891.

(54.) Horree N, Heintz AP, Sie-Go DM, van Diest PJ. p16 is consistently expressed in endometrial tubal metaplasia. Cell Oncol. 2007;29(1):37-45.

(55.) O'Neill CJ, McCluggage WG. p16 expression in the female genital tract and its value in diagnosis. Adv Anat Pathol. 2006;13(1):8-15.

(56.) Dabbs DJ, Sturtz K, Zaino RJ. The immunohistochemical discrimination of endometrioid adenocarcinomas. Hum Pathol. 1996;27(2):172-1 77.

(57.) Alkushi A, Irving J, Hsu F, et al. Immunoprofile of cervical and endometrial adenocarcinomas using a tissue microarray. Virchows Arch. 2003;442(3):271 277.

(58.) Bulten J, van der Laak JA, Gemmink JH, Pahlplatz MM, de Wilde PC, Hanselaar AG. MIB1, a promising marker for the classification of cervical intraepithelial neoplasia. J Pathol. 1996;1 78(3):268-273.

(59.) Pirog EC, Baergen RN, Soslow RA, et al. Diagnostic accuracy of cervical low-grade squamous intraepithelial lesions is improved with MIB-1 immunostaining. Am J Surg Pathol. 2002;26(1):70-75.

(60.) Young RH, Clement PB. Endocervical adenocarcinoma and its variants: their morphology and differential diagnosis. Histopathology. 2002;41:185-207.

(61.) McCluggage WG, Maxwell P, McBride HA, Hamilton PW, Bharucha H. Monoclonal antibodies Ki-67 and MIB1 in the distinction of tuboendometrial metaplasia from endocervical adenocarcinoma and adenocarcinoma in situ in formalin-fixed material. Int J Gynecol Pathol. 1995;14(3):209-216.

(62.) Pirog EC, Isacson C, Szabolcs MJ, Kleter B, Quint W, Richart RM. Prolif erative activity of benign and neoplastic endocervical epithelium and correlation with HPV DNA detection. Int J Gynecol Pathol. 2002;21(1):22-26.

(63.) Chen Y, Miller C, Mosher R, et al. Identification of cervical cancer markers by cDNA and tissue microarrays. Cancer Res. 2003;63:1927-1935.

(64.) Malinowski DP. Multiple biomarkers in molecular oncology: molecular diagnostics applications in cervical cancer detection. Expert Rev Mol Diagn. 2007;7:117-131.

(65.) Santin AD, Zhan F, Bignotti E, et al. Gene expression profiles of primary HPV 16- and HPV 18-infected early stage cervical cancers and normal cervical epithelium: identification of novel candidate biomarkers for cervical cancer diagnosis and therapy. Virology. 2005;331:269-291.

(66.) Kelly D, Kincaid E, Fansler Z, Rosenthal DL, Clark DP. Detection of cervical high-grade squamous intraepithelial lesions from cytologic samples using a novel immunocytochemical assay (ProEx C). Cancer. 2006;108:494-500.

(67.) Shroyer KR, Homer P, Heinz D, Singh M. Validation of a novel immunocytochemical assay for topoisomerase II-alpha and minichromosome maintenance protein 2 expression in cervical cytology. Cancer. 2006;108:324-330.

(68.) Shi J, Liu H, Wilkerson M, et al. Evaluation ofp16NK4a, minichromosome maintenance protein 2, DNA topoisomerase II alpha, ProE(X) C, and p16INK4a/ ProE(X) C in cervical squamous intraepithelial lesions. Hum Pathol. 2007;38: 1335-1344.

(69.) Badr RE, Walts AE, Chung F, Bose S. BD ProEX C: a sensitive and specific marker of HPV-associated squamous lesions of the cervix. Am J Surg Pathol. 2008; 32:899-906.

(70.) Pinto AP, Schlecht NF, Woo TYC, Crum CP, Cibas ES. Biomarker (ProEx C, p16INK4a, and MiB-1) distinction of high-grade squamous intraepithelial lesion from its mimics. Mod Pathol. 2008;21:1067-1074.

(71.) Cooper K, Herrington CS, Stickland JE, Evans MF, McGee JO. Episomal and integrated human papillomavirus in cervical neoplasia shown by non-isotopic in situ hybridisation. J Clin Pathol. 1991;44:990-996.

(72.) Evans MF, Mount SL, Beatty BG, Cooper K. Biotinyl-tyramide-based in situ hybridization signal patterns distinguish human papillomavirus type and grade of cervical intraepithelial neoplasia. Mod Pathol. 2002;15:1339-1347.

(73.) Guo M, Gong Y, Deavers M, et al. Evaluation of a commercialized in situ hybridization assay for detecting human papillomavirus DNA in tissue specimens from patients with cervical intraepithelial neoplasia and cervical carcinoma. J Clin Microbiol. 2008;46(1):274-280.

(74.) Dabic MM, Hlupic L, Babic D, Jukic S, Seiwerth S. Comparison of polymerase chain reaction and catalyzed signal amplification in situ hybridization methods for human papillomavirus detection in paraffin-embedded cervical preneoplastic and neoplastic lesions. Arch Med Res. 2004;35:511-516.

Kay J. Park, MD; Robert A. Soslow, MD

From the Department of Pathology, Memorial Sloan-Kettering Cancer Center, New York, New York.

The authors have no relevant financial interest in the products or companies described in this article.

Presented in part at the Surgical Pathology of Neoplastic Diseases course, Memorial Sloan-Kettering Cancer Center, NewYork, NewYork, May 12-16, 2008.

Reprints: Kay J. Park, MD, Department of Pathology, Memorial Sloan Kettering Cancer Center, 12 75 York Ave, New York, NY 10065 (e-mail: parkk@mskcc.org).
Gale Copyright: Copyright 2009 Gale, Cengage Learning. All rights reserved.