Dysplasia and carcinoma in situ are common lesions; most pathology laboratories receive several cases weekly. It is important to have well-defined terms and criteria for each type of these lesions, thereby promoting correct diagnoses and proper treatment. Unfortunately, as the literature shows us, there is confusion about how to interpret these lesions, in part,because of disagreement about how best to name them (e.g.,mild,mod-erate, severe dysplasia and carcinoma in situ versus cervical intraepithelial neoplasia, CIN 1-3), but also because the various lesions form a broad spectrum of biological and cytological changes with no sharp and precise limits. A two-tiered classification of low-grade and high-grade squamous lesions (LSIL and HSIL) introduced in the USA for reporting cervical cytology is mainly used in the USA by cytologists and cytopatholo-gists. Its introduction for use in histology parallel to the CIN classification is discussed (Schneider 2003; Crum 2003; for comparison see Table 3).
We have chosen to use the terms dysplasia and carcinoma in situ parallel to CIN 1-3, because these are the original names recommended by the WHO (Tavassoli and Devil-ee 2003) and most widely used. The differences between the two nomenclatures are, however,small. The WHO classification is a four-step division,the CIN a three-step. CIN 1 corresponds in general to mild dysplasia, CIN 2 to moderate dysplasia, and CIN 3 covers both severe dysplasia and carcinoma in situ (see Table 3).
The spectrum reaches from mild dysplasia to carcinoma in situ, changes that may be found in very small regions of the cervix, but virtually always in and around the transformation zone. They may involve the whole circumference of the cervical orifice to include smaller or larger parts of the endocervical mucosa. The dysplastic or neoplastic epithelium lies not only on the surface, but may also extend down into the endocervical glands.
It must be emphasized that often varying degrees of dysplasia and carcinoma in situ are found together in the same cervix. The prime diagnosis is made from the most advanced and severe change.
Adenocarcinoma in situ of the endocervical glandular epithelium is much less common. Nonetheless, it seems to be associated with dysplasia and carcinoma in situ of the reserve cell type (Fig. 140), and may arise from the same dysplastic primary lesion, derived from one parent cell clone (see p. 85).
Dysplastic and neoplastic lesions of the uterine cervix are in virtually all cases caused by persistent infections with HR-HPV types (Bosch et al. 2002; Burd 2003). Of the more than 100 types of HPV identified to date, mainly types 16 and 18, but also types 31,33,35, 39,45,51,52,56,58,59,68,73, and 82 are considered as potentially carcinogenic and are thus referred to as HR-HPV types (Crum et al. 1984; Munoz et al. 2003; deVelliers et al. 2004; see Table 4 and Fig. 100).
High-risk papillomaviruses are generally associated with preneoplastic or neoplastic anogenital lesions. Low-risk HPV types are usually found in benign epithelial lesions as, for example, genital warts (condylomata acuminata), but not in malignant lesions. For the group of probable HR-HPV types, the association is not yet confirmed; they have so far been observed only in incidental preneoplastic lesions.
26, 53, and 66
6,11,40,42, 43, 44,54,61,70,
56,58,59,68,73, and 82
72,81, and CP6108
Fig. 100. Relative distribution of individual HR-HPV types in cervical cancers in all world regions combined (in %, modified after Munoz et al. 2004)
Fig. 100. Relative distribution of individual HR-HPV types in cervical cancers in all world regions combined (in %, modified after Munoz et al. 2004)
Clearly HPV 16 and 18 account for most of the cases; however, there is some geographical variation in the prevalence of the various HPV types, as described in detail by Munoz et al. (2004).
Cervical carcinogenesis is thought to be a multistep event, with HPV as a necessary, but not sufficient carcinogenic agent (Walboomers et al. 1999). From various studies it could be concluded that high parity, smoking (Wyatt et al. 2001), immunodeficiency (Frisch et al. 2000) and, less consistently, long-term use of oral contraceptives are cofac-tors that may modulate the risk of progression from HPV infection to cervical cancer (Pater et al. 1988; Elson et al. 2000; Moodley et al. 2003; also reviewed in Castellsague and Munoz 2003). Immunological factors may play an important role in the natural surveillance of the early HPV infection, since immunosuppressed patients have significantly higher rates of persisting HR-HPV infections (Palefsky and Holly 2003).
Ostor, in his widely cited review paper, reports that CIN 1 progresses to CIN 3 in 10% of cases, and that the likelihood of progression of CIN 3 to invasive cancer is >12%. The percentage of definitive regression, however, seems unpredictable, since CIN3 lesions, once histologically verified, were most likely to be excised in total or erased by surgical or chemical damage (Ostor 1993). These observations are in good agreement with the concept that the vast majority of HR-HPV infections remain transient, reflecting a period of active virus replication, and then obviously are defeated by the natural immune response of the infected women (Fig. 101).
Viral gene expression, and particularly that of the viral oncogenes E6 and E7, is in most instances restricted to the intermediate or superficial cell layers. Apparently, control mechanisms suppress their expression in the basal or parabasal cell layers, i.e., in those epithelial cells that retain the capacity to replicate. This mechanism prevents a concomitant expression of viral and cellular genes in genome replication and proliferation (zur Hausen 1994). If these negative regulating mechanisms lose their function, concomitant expression of viral oncogenes and cellular replication may occur. This in-
Progression Invasion n
Fig. 101. Natural history of cervical HPV infections (modified after Schiffman and Kjaer, 2003)
duces severe chromosomal instability due to dysregulation of the mitotic spindle apparatus and results in major numerical and structural alterations of the host cell chromosomes (Duensing and Munger 2004).
Histomorphologically, CIN 1 lesions with this deregulated type of viral oncogene expression can initially not be differentiated in H & E stains from the early CIN 1 lesions that still retain the negative regulatory features and are able to suppress viral oncogene expression in the basal and parabasal cell layers. Immunohistochemistry using anti-pi6INK4a-directed antibodies, however, allows us to monitor the expression of the viral oncogenes in the basal and parabasal cells, since the presence of the E7 oncoprotein in replicating epithelial cells results in strong overexpression of the p16INK4a gene product (see p. 93 and Fig. 107c). This concept of the shift from regulated to deregulated viral gene expression is reflected by the clinical observation that CIN 1 lesions with diffuse pi6INK4a gene expression (due to deregulated viral oncogene expression) have a significantly higher risk for progression to higher grades of dysplasia or even invasive carcinomas in comparison to lesions that still retain the regulated pattern of viral gene expression and thus do not display diffuse expression of pi6INK4a in the basal and parabasal cell layers (Negri et al. 2004).
Precursors of cervical cancer may develop from the basal layer of regenerating ecto-cervical squamous epithelium, from proliferating columnar epithelial cells of the endo-cervical glands, or, most frequently, from hyperplastic reserve cells beneath the endo-cervical glandular epithelium, located in the transformation zone. Since proliferative activity with high mitotic rates is known to be a prerequisite of cancerogenesis, those cells undergoing proliferation will be most vulnerable to carcinogenic agents. In addition to cellular damage by mechanical or chemical irritation, cellular proliferation here is largely initiated by excessive (mainly exogenous) hormonal stimulation of target cells (see p. 49; Moreno et al. 2002). The bipotential capacity of hyperplastic reserve cells to differentiate into glandular (mucinous) as well as squamous epithelial cells (see Fig. 140a; Tase et al. 1989) is in agreement with the findings that coexisting adenocarcino-mas and squamous cell carcinomas are derived from one parent cell clone (see below).
The expression of the viral oncogenes in the replicating basal, parabasal and reserve cell layers results in continuous damage to the host cell genome and ultimately in shifts of the overall DNA content of the host cells. This is reflected by the increasing degree of aneuploidy in these cells. Among those proliferating E6-E7-expressing cell clones, further modified cell clones are selected that stepwise lose their capacity to differentiate in mature squamous epithelia. In consequence the number of cells kept in the active cell cycle increases. This is histopathologically reflected by transition from CIN 1 to CIN 2 and 3 and thickening of the p16INK4a-positive cell layer (Figs. 110,111). At the end of the process, the predominant cell clone has lost its differentiation capacity completely and the lesion impresses as carcinoma in situ with full thickness of replicating diffusely p16INK4a-positive cells (Fig. 117). This process of increasing aneuploidization is characterized by increasing chromosomal instability accompanied by continuous recombination of DNA fragments. In frame of these ongoing recombination events, integration of viral genome fragments may occur. The viral genome fragments are predominantly integrated in fragile sites of the host cell genome; however, they do not prefer a specific locus in the human genome (Wentzensen et al. 2004). The majority (up to 80-100%) of carcinomas in situ and invasive cervical carcinomas contain genomically integrated
HPV. Integration of papillomavirus genomes creates unique molecular fingerprints in the genome and unambiguously allows us to identify HPV transformed cells derived from one parent clone (Luft et al. 2001). Using this fingerprinting technique, coexisting adenocarcinomas and squamous cell carcinomas could be identified as being derived from one parent cell clone. This demonstrates that even completely divergent histomor-phological structures in different parts of a carcinoma still may be derived from one pre-existing parental cell clone.
Some HPV types, as, for example, the genotype HPV 18 and 45, are found particularly frequently in adenocarcinomas, small cell carcinomas and also in carcinomas of relatively young patients. The genomes of these virus types are found to be integrated in a very high percentage of respective cancers. This may suggest that the damage of the host cell genome conferred by these virus types may be particularly extensive. This is further supported by the clinical observation that the average age of patients with HPV 18 or 45 related carcinomas is about 10 years less than that of patients with cancers associated with less aggressive HR-HPV types, such as 31,33,35 or 58, for example. Previous studies have shown that the mean age of patients with invasive carcinoma associated with HPV 16 is 49 years, as contrasted with 37 years for patients with invasive carcinoma associated with HPV 18 (Kurman et al. 1988). Similar observations were made by Walker et al. (1989),who observed rapid recurrences in 45% of patients with carcinomas containing HPV 18 compared with 16% of patients with tumors associated with HPV 16. According to the speed of progression, differences could be observed between several HR-HPV types, e.g., HPV 16 and 18 (Lörincz et al. 1992). The odds ratio indicating the risk of progression to carcinoma for the individual HPV types is different, being highest for HPV 18. Several other reports suggest that the overall survival of patients having cancer associated with HPV 18 or 45 is less in comparison to patients with cancers related to HPV 31 or 33 (Rose et al. 1995; Burger et al. 1996; Lombard et al. 1998; Schwartz et al. 2001). It is therefore of paramount importance to pursue appropriate precautions to prevent an invasive carcinoma from evolving once the diagnosis of a preinvasive lesion is made, by assessing all risk factors and diagnostic tools (see also Trunk et al. 2005).
For routine use, the difference of odds ratios between LR and HR types is clinically more significant, however, than the difference between odds ratios for the individual HR types, warranting testing for risk groups (LR versus HR) rather than for individual HPV types in clinical routine (Munoz et al. 2003).
Histopathology and Immunohistochemistry
Dysplasia and Carcinoma In Situ (CIN 1-3; Figs. 102-140)
The development of dysplasia and carcinoma in situ is a continuum, extending from slight to severe cytological atypia with a gradual loss of epithelial stratification; an increase in nuclear changes, and an increase in the number of atypical mitoses. Because the changes merge into one another, it may be difficult in some instances to differentiate between mild and moderate dysplasia, moderate and severe dysplasia, or severe dyspla-sia and carcinoma in situ. On the other hand, the distinction between a mild dysplasia and an irregular regenerative or reparative epithelium should not prove difficult. The nuclei of regenerative epithelium may be enlarged, slightly irregular, yet the chromatin pattern normal. The epithelial stratification may be barely altered. On the other hand, the nuclei of dysplastic epithelium are dyskaryotic from the beginning, enlarged, and irregular with chromatin clumping. In addition, abnormal mitotic figures, loss of basal polarity and multinucleated cells may be present. Depending upon the cell of origin and/or the direction of cellular differentiation, structural variations can be observed between squamous and reserve cell intraepithelial neoplasias.
In mild and moderate dysplasia of squamous cell type (CIN 1 and 2) (Figs. 102-107), epithelial stratification is only partly lost, with an increasing basal hyperplasia and a general thickening (Figs. 102,105) or slightly papillomatous change (Figs. 103,104) of the epithelial layer. Basal polarity and cellular orientation are gradually lost. The nuclei become irregular, hyperchromatic with an abnormal, coarsely granular chromatin pattern. Mitoses increase in number, primarily in the basal and middle epithelial layers. The upper epithelial layers may (Fig. 106) or may not (Fig. 105) show koilocytosis, as a result of viral replication following HPV infection. In mild or moderate dysplasia, the presence of koilocytes may indicate replication of LR-HPV types, e.g., 6 or 11, which will remain episomal; such a dysplasia is virtually always reversible. If, however, the infection is with HR-HPV types, the viral DNA may become integrated into the cellular genome. This, besides mere deregulated expression of the viral E6 and E7 oncogenes in basal or parabasal cells, may further contribute to malignant progression of cervical t t
Fig. 107b. Same as Fig. 107a. Positive nuclear signals in koilocytes located in intermediate cellular layers
precancer (Klaes et al. 1999) and initiate malignant change. The dysplastic change starts at the squamocolumnar junction and is usually located there, originating from the basal layer of the regenerated ectocervical epithelium.
■ Differential Diagnosis. It may be possible to distinguish between reversible and irreversible koilocytic dysplasia at this early stage by identifying the risk type of infective virus (Campion et al. 1986; Fig. 107a and b, 124). If a LR type of HPV is detected, the lesion is to be considered as reversible; if a HR type is identified, it is not possible to make the distinction between reversible and irreversible dysplasia on the assessment of HPV type alone (refer to Fig. 101). The higher the grade of dysplasia, the higher the ratio of HR-HPV infections.
If the decision is not possible, histological (Winkler et al. 1984; Crum et al. 1985) or immunohistochemical (Dallenbach-Hellweg and Lang 1991) distinction should be attempted (see p.8f). Dysplasias infected with LR-HPV have dyskaryotic polyploid nuclei but normal mitoses and express cytokeratin 13 only (Fig. 104). Those infected with HR-HPV types have often aneuploid nuclei and atypical mitoses and often, but not always (Fig. 115) show a coexpression of cytokeratins 13,8,18, and CEA. Such coexpression of intermediate filaments and CEA may indicate that the dysplasia is most likely of reserve cell type (see below). The distinction is clinically important for deciding how to treat the patient further. According to recent studies (Wang et al. 2004; Negri et al. 2004),
pi6INK4a, a cyclin-dependent kinase inhibitor, may be a prognostic marker (see Fig. 107c). Negri showed that CIN 1 lesions with diffuse pi6INK4a staining had a significantly higher tendency to progress to a high-grade lesion than pi6INK4a-negative cases.
In severe dysplasia with squamous differentiation (CIN 3) (Figs. 108-112), loss of epithelial stratification is almost complete. Nuclear changes (enlargement, chromatin clumping, polymorphism, hyperchromasia) and the number of mitoses are considerably increased. Since koilocytes can only develop in maturing cells, they will be less numerous than in mild or moderate dysplasia. The dysplastic epithelium is usually quite high, may be covered by a thick layer of parakeratosis (Fig. 112), and may form broad papillae that extend downwards into the underlying stroma (Fig. 109) or into the mouths of endocervical glands (Fig. 108).
Carcinoma in situ with squamous differentiation (CIN 3) [8070/2] (Figs. 113-117) shows a complete loss of stratification. The entire epithelium consists of poorly differentiated neoplastic cells, which contain disorganized, large, atypical, hyperchromatic nuclei surrounded by little cytoplasm. Atypical mitoses are frequent in all layers; koilocytes are not present. This neoplastic epithelium originates at the squamocolumnar junction and from there may grow out to replace large areas of the ecto- and endocervical surface epithelium. It is usually covered by a layer of atypical parakeratosis. Its spread into glands is much less pronounced than that of carcinoma in situ with reserve cell differentiation.
Mild, moderate or severe dysplasia with reserve cell differentiation (CIN 1-3) [8077/2] (Figs. 118-133) develops from the reserve cell layer of the endocervical epithelium and is usually preceded by a reserve cell hyperplasia. The gradual increase in cytological and nuclear atypicality corresponds closely to the various degrees of squamous-type dysplasia. Because the reserve cells are bipotential, irregular maturation towards mucin formation, often monocellular (Fig. 122), clear cell change (Figs. 118,123), or maturation towards keratinization (Fig. 119) may be seen. Koilocytosis is often present and nearly always caused by infection with HR-HPV types (Fig. 124). Topographically, reserve cell dysplasia is not always concentrated at, but rather above the squamocolumnar junction, and may also arise anywhere within the endocervical mucosa (surface epithelium and glands; Figs. 118, 119).
In rare instances reserve cell dysplasia may form papillary structures resembling a pedunculated papilloma of the ectocervix (Figs. 130,131). Since an invasive carcinoma may develop at their base, the papillomatous growths should be excised in total and examined carefully (Randall et al. 1986). Because of the histological resemblance of its epithelial proliferation to transitional metaplasia, this lesion has been called transitional cell papilloma (Albores-Saavedra and Young 1995). Its negative reaction for CK 20, however, casts some doubts on this interpretation and appears more in favor of a reserve cell (Mullerian) origin (see also Koenig et al. 1997).
The carcinoma in situ with reserve cell differentiation (CIN 3) [8077/2] (Figs. 134-140) also presents a complete loss of stratification with densely arranged atypical nuclei. These are generally elongated and usually smaller than those of the squamous type. Atypical mitotic figures such as three-group metaphases are frequently observed (Fig. 136,137). Koilocytes are not seen,because no mature cells are present in the upper layers. The very sparse cytoplasm of the tumor cells may show incomplete differentiation in
Fig. 140. a Carcinoma in situ, reserve cell differentiation. Positive immunohistochemical reaction for CEA (upper right corner) and with coexisting adenocarcinoma in situ (lower left corner), demonstrating that both precancerous lesions derive from one parent cell clone (s. p. 85f). b Carcinoma in situ, reserve cell differentiation, with coexisting adenocarcinoma in situ.Van Gieson
Fig. 140. a Carcinoma in situ, reserve cell differentiation. Positive immunohistochemical reaction for CEA (upper right corner) and with coexisting adenocarcinoma in situ (lower left corner), demonstrating that both precancerous lesions derive from one parent cell clone (s. p. 85f). b Carcinoma in situ, reserve cell differentiation, with coexisting adenocarcinoma in situ.Van Gieson some areas, forming either keratin or mucin or presenting cytoplasmic clearing (Fig. 138). In contrast to that with squamous differentiation, carcinoma in situ with reserve cell differentiation often grows into glands, and in the noninvasive stage may even spread out over large areas of the endocervical mucosa (Fig. 139).
■ Differential Diagnosis. The two types of dysplasia and carcinoma in situ with squamous or with reserve cell differentiation can be distinguished in most cases by their topographical localization and spread (pattern of growth), and by the variation in their nuclear structure, as pointed out above. In some instances, both types develop simultaneously in neighboring areas. There are, however, intermediate lesions in which it is difficult to make a distinction (Figs. 108,109,113,114). In those instances immunohisto-chemistry will be of considerable help, since both types have different compositions of intermediate filaments in their cytoplasm: the dysplasias and carcinoma in situ with squamous differentiation react positively with cytokeratin 13 (Fig. 115), but negatively with cytokeratin 8 and 18 and with CEA. The in situ lesions with reserve cell differentiation, on the contrary, show a coexpression of cytokeratin 13,8 and 18 and are also positive with CEA (Fig. 140a). The distinction between the two is clinically important, since the reserve cell type, which is most frequently observed in young women, progresses more rapidly and frequently recurs in the infected mucosa of the endocervix, despite its complete removal by conization. Such recurrences are explained by the fact that normal epithelium around the carcinoma in situ may already be infected by HPV at the time of conization (Colgan et al. 1989). Distinction of carcinoma in situ from transitional cell metaplasia is obvious by the regular nuclear structure of the metaplastic lesion (see p. 34).
Adenocarcinoma In Situ [8140/2] (Figs. 141-149)
In adenocarcinoma in situ the normal columnar epithelium of the endocervical glands may be replaced by two types of abnormal epithelial cells. In the first type, one sees a pseudostratified or even multilayered atypical glandular epithelium with enlarged, elongated, hyperchromatic nuclei surrounded by a sparse undifferentiated cytoplasm. Mitoses are frequent. Mucin formation is absent or minimal (uniform type of Gloor and Ruzicka 1982; Figs. 141,142). The glandular lumina are preserved, but may show intra-glandular budding or bridging (Fig. 144). This atypical change is usually focal and may be limited to only one part of the gland (Fig. 141). In the second type, seen less frequently, the glands may be lined by strikingly disorganized, irregular cells with enlarged depolarized, pleomorphic nuclei with disordered chromatin. The cytoplasm is clear, foamy, and PAS negative (pleomorphic type of Gloor and Ruzicka; Figs. 145,146). Occasional goblet cells may be found. These two types of ACIS may replace smaller or larger portions of normal glandular epithelium of the endocervical mucosa. Occasionally, the surface epithelium of the endocervix may also be involved (Fig. 147).
Adenocarcinoma in situ of the endocervix was once a rare lesion (Friedell and McKay 1953; Abell and Gosling 1962; Krimmenau 1966; Barter and Waters 1970; Sachs et al. 1975; Werner and Waidecker 1975), but has since increased in frequency (Jaworski et al. 1988; Hemminki et al. 2002; Wang et al. 2004), like invasive adenocarcinoma (see p. 136), and, like this, contains HPV 18 in a high percentage (Farnsworth et al. 1989; Tase
Fig. i43a,b. Adenocarcinoma in situ. a H&E; b p16INK4a. Diffuse, continuous positivity for p16INK4a in atypical epithelium, negative reaction in normal endocervical epithelium et al. 1989). Since only a few glands may be affected and be focally distributed, they may be overlooked unless serial sections of a cone biopsy are examined. If the lesion is overlooked or incompletely excised, it will progress to invasive adenocarcinoma (Boon et al. 1981; Wells and Brown 1986; Östör et al. 2000). Since both glandular and squamous cells of the endocervix may originate from the subcolumnar reserve cell (Alva and Lauchlan 1975; Christopherson et al. 1979; Boon et al. 1981; Tase et al. 1989),their precancerous and neoplastic changes are often closely related. Consequently, they often show an identical or a very similar expression of intermediate filaments (Fig. 140). When dealing with a dysplasia or carcinoma in situ of the endocervix, it is essential to search carefully for a coexistent ACIS.
■ Differential Diagnosis. Adenocarcinoma in situ can be distinguished from adenomatous and microglandular hyperplasia of the endocervix immunohistochemically: in most cases ACIS reacts positively with p16INK4a (Fig. 143b) and CEA (Fig. 148), whereas adenomatous and microglandular hyperplasias react negatively (compare also Hurlimann and Gloor 1984; Gloor and Hurlimann 1986; Cameron et al. 2002; Negri et al. 2003). Distinction from invasive adenocarcinoma is not possible from a small biopsy, but rather requires examination of larger areas of the endocervical mucosa. In contrast to invasive carcinoma, ACIS is limited to the glands; a stromal response is lacking, and normal glands are constantly admixed with neoplastic glands (Östör et al. 1984).
Was this article helpful?
Learning About 10 Ways Fight Off Cancer Can Have Amazing Benefits For Your Life The Best Tips On How To Keep This Killer At Bay Discovering that you or a loved one has cancer can be utterly terrifying. All the same, once you comprehend the causes of cancer and learn how to reverse those causes, you or your loved one may have more than a fighting chance of beating out cancer.