Carcinoma and Other Tumors of the Cervix
The World Health Organization (WHO) 2014 classification recognizes three general categories of invasive carcinoma of the cervix: squamous cell carcinoma, adenocarcinoma, and “other epithelial tumors” (Table 1) (Kurman et al. 2014). The “other epithelial tumors” include adenosquamous carcinomas, adenoid basal and adenoid cystic carcinomas, undifferentiated carcinoma, as well as neuroendocrine tumors. The relative frequency of these different tumor types varies between countries; in general, squamous cell carcinoma is the most common histologic subtype accounting for 76–89% of invasive carcinomas. Adenocarcinoma and adenosquamous carcinoma comprise 10–24% of cervical cancers, and all other categories are relatively rare, adding up to less than 5% (de Sanjose et al. 2010).
Modified World Health Organization histological classification of invasive carcinomas of the uterine cervix
Squamous cell carcinoma
Endocervical, usual-type adenocarcinoma
Signet ring cell adenocarcinoma
Clear cell adenocarcinoma
Other epithelial tumors
Glassy cell carcinoma
Adenoid basal carcinoma
Adenoid cystic carcinoma
Small cell carcinoma
Large cell neuroendocrine carcinoma
2009 modification of FIGO staging of carcinoma of the uterine cervix
Cervical carcinoma confined to uterus (extension to the corpus should be disregarded)
Invasive carcinoma diagnosed only by microscopy; all macroscopically visible lesions, even with superficial invasion, are stage IB
Stromal invasion no greater than 3.0 mm in depth and 7.0 mm or less in horizontal spread
Stromal invasion more than 3.0 mm and not more than 5.0 mm with a horizontal spread of 7.0 mm or lessa
Clinically visible lesion confined to the cervix or microscopic lesion greater than IA2
Clinically visible lesion 4.0 cm or less in greatest dimension
Clinically visible lesion more than 4.0 cm in greatest dimension
Tumor invades beyond the uterus but not to pelvic wall or to lower third of the vagina
Without parametrial invasion
Clinically visible lesion ≤4.0 cm in greatest dimension
Clinically visible lesion >4 cm in greatest dimension
With parametrial invasion
Tumor extends to the pelvic wall and/or involves lower third of vagina and/or causes hydronephrosis or nonfunctioning kidneyb
Tumor involves lower third of vagina with no extension to pelvic wall
Tumor extends to pelvic wall and/or causes hydronephrosis or nonfunctioning kidney
The carcinoma has extended beyond the true pelvis or has involved (biopsy proven) the mucosa of the bladder or rectum. A bullous edema, as such, does not permit a case to be allotted to stage IV
Spread of the growth to adjacent organs
Spread to distant organs
Squamous Cell Carcinoma
Superficially Invasive Squamous Cell Carcinoma
The Lower Anogenital Squamous Terminology (LAST) Standardization Project for HPV-Associated Lesions introduced a new term – “superficially invasive squamous cell carcinoma” (SISCC) – for microscopic, clinically unapparent cervical squamous cell carcinoma (Darragh et al. 2012). SISCC is defined as a lesion that is not grossly visible, has an invasive depth of less than or equal to 3 mm from the basement membrane of the point of origin, and has a horizontal extent of less than or equal to 7 mm in maximum dimension, with margins negative for carcinoma. The presence of HSIL at the margin does not exclude a tumor from this category but should be noted. The presence of lymphatic invasion and whether there is multifocal invasion should be also included in the report (Darragh et al. 2012). SISCC represents the earliest phase of invasion from the background of high-grade squamous intraepithelial lesion (HSIL), and being an early lesion, it has a favorable prognosis.
In the past, early invasive carcinoma was termed “microinvasive carcinoma” (MICA). The Society of Gynecologic Oncology (SGO) in the United States defined it as a microscopic tumor that invaded to a depth of no greater than 3 mm beyond the base of the epithelium of origin, either surface or glandular, with no criterion for lesion width. Lymphatic invasion was not allowed, and margins were required to be negative for carcinoma (a requirement for the margin status with regard to the presence of SIL was not specified). The International Federation of Gynecology and Obstetrics (FIGO) had its own criteria and subdivided MICA into those that invaded no more than 3 mm in depth (stage IA1) and those that invaded more than 3 mm but not more than 5 mm in depth (stage IA2) and specified that horizontal extent could not exceed 7 mm (see Table 2). The presence of lymph-vascular invasion did not exclude a tumor from FIGO stage IA (Berek and Hacker 2010). Due to the variability in definitions of MICA, this terminology is not recommended, and the term SISCC as proposed by the LAST Project is currently endorsed. Clinical evidence demonstrates that lesions categorized as SISCC and without lymph-vascular invasion have an extremely low rate of lymph node metastases, recurrences, or risk of death. Such patients may be offered conservative treatment, usually conization or trachelectomy, while those with tumors measuring >3 mm in depth or with lymphatic invasion should be considered for more radical therapy (see subsection “Treatment”) (Berek and Hacker 2010; Eskander et al. 2015). Single institution studies report a rate of SISCC as 3.6–4.7% in patients treated with cone excisions after biopsy diagnosis of HSIL (Killackey et al. 1986; Matseoane et al. 1992).
The average age of patients with SISCC is 39–42 years. Most patients with SISCC are asymptomatic with a grossly normal cervix or “friable cervix” with abnormal capillaries prone to bleeding. Neither cytologic results nor colposcopic exam can accurately predict the presence of early invasion. In cytology, abnormally prominent nucleoli may be seen in dysplastic cells, suggesting an invasive process. However, the positive predictive value of cytology for the presence of early invasion is reported at just 27.3% (Andersen et al. 1995). In colposcopic examination, the early invasion is typically seen within an acetowhite area, consistent with the background of HSIL. Within this area an abnormal vascular pattern is observed consisting of vessels with irregular and haphazard distribution, increased intercapillary distance, marked variations of caliber, and abrupt changes in direction forming acute angles. The sensitivity of colposcopic exam for identification of SISCC is reported at 30–50%. As expected, colposcopy predicts invasion more accurately with increasing depth of invasive lesion (Berek and Hacker 2010).
The findings in HSIL associated with, or predictive of, superficial invasion include extensive involvement of surface epithelium by HSIL, extension of HSIL to deep endocervical crypts, lumenal necrosis, and intraepithelial squamous pseudomaturation. When such findings are present in a biopsy or cone specimen, extra attention should be paid to the possibility of invasion, and deeper sections may be obtained to thoroughly evaluate any suspicious areas.
Ultrastructural electron microscopy study of early invasion in cervical squamous cell carcinoma has demonstrated that there is a disappearance of basal lamina of the basement membrane and pseudopod-like cytoplasmic protrusions of the cancer cells are seen in a direct contact with the underlying stroma (Kudo et al. 1990). The protrusions of the invading neoplastic cells contain abundant cytoplasmic vesicles, 70–90 nm in size, some of which are seen open directly into the extracellular matrix of the stroma. These vesicles are not observed in adjacent HSIL, which is supported by the intact basement membrane. These findings suggest that the substances contained in the vesicles may play a role in basement membrane destruction. In addition, the neoplastic cells traversing the basement membrane gaps show accumulation of actin filaments in the pseudopod protrusions. These local aggregates of cytoskeletal structures, not observed in adjacent HSIL, are thought to facilitate an ameba-like movement of the cancer cells. Finally, neoplastic cells in the areas of superficial invasion show decreased number of desmosomal junctions, enabling cellular discohesion and migration (Kudo et al. 1990).
Measurement and Significance of Depth of Invasion
Percentage of pelvic node metastases, recurrence, and death of disease in relation to depth of invasion in early invasive squamous cell carcinoma. Number of reported patients shown in parenthesis
Depth of cervical stromal invasion
Lymph node metastasis
0.16% (4 out of 2508)
0.75% (9 out of 1196)
6.38% (35 out of 548)
0.45% (11 out of 2426)
1.12% (29 out of 2586)
4.09% (43 out of 1049)
Death of disease
0.12% (3 out of 2426)
0.50% (9 out of 1776)
1.88% (16 out of 850)
Measurement and Significance of Horizontal Extent
Residual invasive tumor in postconization hysterectomy specimens according to lateral extent of carcinoma showing up to 5 mm depth of invasion
Lateral extent of invasion (mm)
No. of patients
Percent (%) with residual disease in postconization hysterectomy specimen
Estimation and Significance of Third Dimension and Tumor Volume
Early invasive carcinoma extends longitudinally within the endocervical canal in cephalocaudal direction, and this measurement is assessed as horizontal tumor extent. But the lesion may also expand in the third dimension – circumferentially – and exceed the linear extent of 7 mm, a cutoff for superficially invasive carcinoma. In cases in which the invasive lesion spans adjacent blocks of tissue, it is recommended to estimate the third measurement of tumor circumference by multiplying the number of involved blocks by the estimated block thickness. It is important that such measurement should be only assessed from invasive areas that co-localize in contiguous sections. After determining the horizontal and circumferential tumor measurement, the larger of the two should be used to assign the tumor stage.
Burghardt and Holzer have introduced the concept of tumor volume as applied to SISCC and have reported no pelvic node metastases in patients with 420 mm3 of cancer or less, with the exception of one case in which vascular invasion was noted (Burghardt and Holzer 1977). In recent years, the concept of SISCC tumor volume has been emphasized by some authors. However, tumor volume estimation is not very accurate; it may require additional serial sectioning of the cone specimens and, therefore, is unlikely to become a routine laboratory method.
Assessment and Significance of Lymph-Vascular Space Invasion
Recommended treatment for stage IA cervical carcinoma
Margins and ECC negative
Cervical cone biopsy
Margins and ECC negative
Extrafascial hysterectomy with pelvic lymph node dissection
Cervical cone biopsy and laparoscopic pelvic lymph node dissection
Margins and ECC negative
Modified radical hysterectomy with pelvic lymph node dissection
Radical trachelectomy with pelvic lymph node dissection
IA1 and IA2
Margins or ECC positive for HSIL or invasion
Modified radical hysterectomy with pelvic lymph node dissection, if cone not feasible
Repeat cone biopsy
Assessment and Significance of Surgical Margin Involvement
Assessment of the margins is usually straightforward except for cases with marked cautery artifact of tissue edges. P16 or Ki-67 immunostaining may be helpful in such cases as both antigens survive thermal process. The status of cone margins is an important parameter in deciding the therapeutic approach to patients with SISCC. In most studies, women with cone margins positive for either SIL or invasive disease have a significantly higher rate of residual disease in the hysterectomy specimen (range 36–54%) than women with a negative cone margin (range 4–10%) (Yoneda et al. 2015; Jones et al. 1993a). A positive margin warrants re-excision or hysterectomy, as the residual invasion may be even deeper than that found in the cone biopsy specimen.
Of 265 cases with an original diagnosis of SISCC submitted to a group of reference pathologists of the Gynecologic Oncology Group (GOG), 37% (99 cases) were rejected due to overdiagnosis (superficial invasion was not identified by the expert panel) (Sedlis et al. 1979). Overdiagnosis typically results from overinterpretation of tangential epithelial sectioning in the setting of reactive, reparative changes which accompany severe cervicitis. Another error is overinterpretation of squamous epithelial nests (benign or dysplastic) entrapped deep in the cervical stroma in the area of a healing wound following cone excision. In both instances immunostaining for p16 may be helpful in delineating the contours of the lesion. Jagged, irregular borders of p16-positive epithelial nests favor superficially invasive carcinoma. Missed diagnosis of SISCC may be a result of obscuring dense inflammatory infiltrate or artifacts such as cautery and obscuring hemostatic agents such as ferric subsulfate (Monsel’s solution) or silver nitrate.
The therapy for stage IA cervical carcinoma is based on findings in the cone biopsy which include the depth of invasion, horizontal spread, presence of LVSI, as well as margin and endocervical curettage involvement (Table 5). The recommended therapy is further individualized with regard to the patient’s desire for future fertility and availability for follow-up. The data on risk factors for nodal metastases, recurrence, and death suggest that lesions with 3 mm or less stromal invasion without LVSI have only minimal potential for metastasis or recurrence; therefore, for such cases most centers recommend simple hysterectomy. However, women who desire to remain fertile can be managed with conization, provided that they accept minimal risk of recurrent disease and are available for long-term follow-up. For stage IA1 tumors with lymphatic invasion, the patient should receive lymph node dissection in addition to hysterectomy or cone excision. At present, radical hysterectomy with pelvic lymphadenectomy is considered the most appropriate therapeutic approach for most stage IA2 squamous cell carcinomas of the cervix. In women desiring to preserve fertility, radical trachelectomy (amputation of cervix) and laparoscopic bilateral pelvic lymphadenectomy can be performed (Berek and Hacker 2010; Eskander et al. 2015).
Invasive Squamous Cell Carcinoma
Cervical cancer is the fourth most common cancer in women worldwide, with an estimated 528,000 new cases in 2012. A large majority (close to 85%) of the global burden occurs in the regions with limited resources for screening. The highest-risk regions show age-standardized incidence rates over 30 per 100,000 women per year and include Eastern Africa (42.7 per 100,000), Melanesia (33.3 per 100,000), Southern Africa (31.5 per 100,000), and Middle Africa (30.6 per 100,000). The lowest rates are recorded in Australia and New Zealand (5.5 per 100,000), and Western Asia (4.4 per 100,000) (Globocan 2012). In the United States, in 2013, there were estimated 13,000 new cases of cervical cancer (all histologic types) corresponding to age-standardized rate of 7.5 per 100,000 (NCI 2017).
There is little doubt that cytologic screening has played a major role in reducing both the incidence and mortality from cervical cancer. In 1941 Dr. George Nicolaou Papanicolaou published a paper followed by a book in 1943 entitled Diagnosis of Uterine Cancer by the Vaginal Smear explaining how to use the cytology smear to screen for cervical neoplasia. Within a decade, in the early 1950s, the screenings with Pap smears began in the United States, and the cervical cancer rate has showed a continuous decline. In the 1940s, the incidence of invasive cervical cancer in the United States was approximately 30 per 100,000 women, which is similar to that currently seen in developing countries. After initiation of screening, the incidence was cut by half by 1975, when it had declined to 15 per 100,000, and then it was again cut by half to approximately 7.5 per 100,000 in the 2000s, mostly due to a decrease in squamous cell carcinoma. In 2003, the US Food and Drug Administration (FDA) approved the first HPV testing assay for use in conjunction with Pap testing. Overall, since the inception of Pap screening, there has been a 75% reduction in cervical cancer incidence in the United States. The reduction in incidence has been paralleled by the reduction in mortality. In 2013, in the United States there were estimated 4000 cervical cancer-related deaths, corresponding to 2.3 deaths per 100,000 (NCI 2017). The mortality was cut by half since 1975, when the death rate was 5.5 per 100,000 women per year (NCI 2017).
The current incidence and mortality in the United States show racial differences, mainly due to uneven access and lack of participation in screening and treatment. The incidence is higher among African-American and Hispanic women as compared to Caucasian women, with 8.9 and 9.4 versus 7.5 cases per 100,000, respectively (NCI 2017). The cervical cancer deaths are the highest among African-American women (3.9 per 100,000), followed by American Indian (3.5 per 100,000), Hispanic (2.6 per 100,000), and Caucasian women (2.1 per 100,000) (NCI 2017).
Further reduction of invasive cervical carcinoma in the United States may be achieved by increasing participation in screening, which is expected to rise as a result of the Affordable Care Act and by HPV vaccination. Two preventive HPV vaccines, Gardasil and Gardasil 9, were approved by the FDA in 2006 and 2014, respectively. However, it will take several years to see the measurable impact of both the Affordable Care Act and HPV vaccination on cervical cancer rates.
Despite the advances in prevention, detection, and management of cervical cancer in developed countries, the tumor continues to pose significant healthcare problem worldwide. Globally, one in two women with cervical cancer dies as a result of this disease. In 2012 there were 266,000 cervical cancer deaths, of which 87% occurred in less developed regions (Globocan 2012).
The epidemiologic risk factors identified as significant for invasive cervical carcinoma can be divided into sexual behavior factors related to HPV acquisition, environmental factors such as tobacco smoking, hormonal factors such as oral contraceptive use, host factors like immunosuppression and genetic polymorphism, and viral factors related to different oncogenic potentials of different viral genotypes.
Infection with HPV shows the highest magnitude of relative risk/odds ratio values for development of cervical cancer, with association that is even stronger than that between tobacco smoking and lung carcinoma. The risk shows considerable variation depending upon the specific viral genotype. Infection with HPV 16 has the highest odds ratio (OR) in comparison to noninfection, for development of squamous cell carcinoma, with OR of 434.5 (95% CI: 278.2–678.7), followed by HPV 33, with OR of 373.5 (95% CI: 46.7–2985.8) and HPV 18, with OR of 248.1 (95% CI: 138.1–445.8). In contrast, OR is only 4.3 (95% CI: 0.5–38.4) for HPV 6, a low-risk oncogenic viral type (Munoz et al. 2003).
Women with invasive squamous cell carcinoma have epidemiologic characteristics similar to those with precursor lesions (see chapter “Precancerous Lesions of the Cervix”). The most important are sexual behavior risk factors which are related to acquisition of HPV infection. Like with SIL, the relative risk of invasive cervical carcinoma is significantly increasing with increasing number of sexual partners, number of pregnancies, early onset of sexual activity, and early age of first pregnancy (International Collaboration of Epidemiological Studies of Cervical Cancer 2009). There is also evidence that factors other than HPV are important in the development of invasive squamous cell carcinoma of the cervix. After controlling for sexual and reproductive factors, there continues to be a significant association between smoking and development of cervical squamous cell carcinoma, but not adenocarcinoma (Appleby et al. 2006). The magnitude of the risk is relatively small; current smokers show a 1.6-fold increased risk of squamous cell carcinoma of the cervix compared to never smokers (Appleby et al. 2006). The risk was observed to decline after cessation of smoking. In addition, long-term use of oral contraceptive can increase the risk of developing of both HSIL and invasive carcinoma (Appleby et al. 2007). The risk of invasive cancer increases with increasing duration of oral contraceptive use. In women who had used combined oral contraceptives for 5 or more years and are current users, the risk of cancer is 1.9-fold higher than that of never users. The risk declines after cessation of use and then, by 10 or more years, returns to the risk of never users. Patients with immunosuppression due to HIV infection have increased risk of HSIL and invasive carcinoma, with relative risk of cervical cancer increased by sixfold compared to general population (Dugue et al. 2013). In 1993, the CDC listed cervical cancer as an AIDS-defining disease. Immunosuppression following organ transplantation or treatment of autoimmune conditions results in increased cervical cancer risk with a range between two- and 30-fold (Dugue et al. 2013).
Since HPV infection in women is very common, with over 80% of women in Western countries being infected in their lifetime, but the development of cancer is a relatively rare event, it has been suggested that there may be a specific personal susceptibility to cervical cancer in some women, possibly on a genetic level. A number of host genetic factors including p53 polymorphism and major histocompatibility complex (MHC) alleles were investigated for their association with cervical cancer, but their precise roles have not been clarified.
The discovery and elucidation of HPV-related etiology of cervical carcinoma occurred over decades and involved epidemiologic, clinicopathologic, and molecular studies. Originally, it was the epidemiologic data that suggested an infectious agent in etiology of cervical cancer. In 1979, virologist Harald zur Hausen published a hypothesis that HPV may play a causative role in cervical cancer, and in 1983 he and his collaborators identified HPV 16 and HPV 18 in cervical tumors. For his role in the discovery, zur Hausen was awarded Nobel Prize in Medicine in 2008. Based on the strength of the molecular and epidemiologic evidence linking HPV to invasive squamous cell carcinoma of the cervix, 15 HPV types have been classified as carcinogenic (high-risk) to humans, including HPV 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 68, 73, and 82, and 3 types as probably carcinogenetic, including HPV 26, 53, and 66 (Munoz et al. 2003). For details of HPV classification and molecular mechanisms of HPV-driven tumor pathogenesis, see subsection “Classification of HPV and Association with Specific Types of Anogenital Lesions and Genomic Organization of HPV” in chapter “Precancerous Lesions of the Cervix.”
Except for rare subtypes of cervical adenocarcinoma, all cervical histologic tumor types are related to HPV infection, including squamous cell carcinoma, adenocarcinoma of usual type, adenosquamous carcinoma, and neuroendocrine tumors. Overall, HPV DNA is detected in 85–92% of cervical carcinomas of all histologic types (de Sanjose et al. 2010; Tjalma et al. 2013). Globally, HPV 16 is the most frequent type detected in invasive cervical cancer (61%), followed by HPV 18 (10%) and HPV 45 (6%) (de Sanjose et al. 2010). Five other types follow in the order: HPV 31 (4%), 33 (4%), 52 (3%), 35 (2%), and 58 (2%), and the remaining seven high-risk HPV types account for less than 1% of tumors, individually. The distribution of HPV types in invasive cervical cancers shows only minor geographic variations. Notably, type HPV 18 and related HPV 45 is seen more frequently in cancers in Africa (23% and 10%, respectively) at the expense of lower HPV 16 percentage (48%) (de Sanjose et al. 2010). HPV 18 and HPV 45 detection in tumor samples has been shown to correlate with younger patient age, faster progression to invasion, and higher cancer mortality (de Sanjose et al. 2010; Tjalma et al. 2013; Schwartz et al. 2001). The most comprehensive HPV preventive vaccine, Gardasil 9, is designed to prevent infection with seven high-risk HPVs including types 16, 18, 31, 33, 45, 52, and 58. It is expected that this vaccine should prevent over 80% of cervical cancer cases (90% of 90% HPV-positive tumors).
The failure to detect HPV DNA in a small subset of invasive squamous cell carcinomas appears to be caused by a number of factors including low quality of the histologic sample, presence of a HPV type other than those assayed for, and loss of part of the HPV genome after integration into the host DNA. However, when multiple molecular techniques are used to detect HPV DNA in cervical squamous cell carcinomas, the detection rate approaches 99.7% (Walboomers et al. 1999). For tumors in which multiple HPV types were detected in the whole tumor sections, it has been shown that after laser-capture microdissection of the tumor tissue, only one HPV type is detected in the tumor, whereas additional HPV types are detected in either adjacent uninvolved mucosal regions or in cellular debris. This observation leads to recognition of the phenomenon “one virus, one lesion” in which a specific, single HPV genotype can be mapped to a specific tissue location and identified as responsible for the neoplastic process (Quint et al. 2012).
The natural history of HPV infection in relation to cancer development is covered in detail in subsection “Development of Cervical Disease After HPV Infection” in chapter “Precancerous Lesions of the Cervix.” Here, we only mention in brief that most cervical HPV infections are asymptomatic without development of an intraepithelial lesion. In women in which HPV causes mucosal changes, the majority (approximately 85%) of lesions are so-called productive viral infections resulting in LSIL, and the majority of these lesions regress spontaneously within a time span of up to 2 years. Approximately 15% of cervical mucosal lesions are so-called transforming HPV infections resulting in HSIL, which have the potential of progression to invasive carcinoma. The time frame of progression and the rate of progression of HSIL to invasive carcinoma can be only estimated from indirect evidence, since follow-up studies are not possible. The time of progression is estimated from the comparison of peak ages of detection of HSIL and invasive cancer. For HPV 16, the median age of patients with HSIL is 34 years, compared to a median age of 49 years for patients with invasive squamous cell carcinoma. This observation suggests that HPV 16-positive HSIL has an average of 15 years to progress to invasive carcinoma. For HPV 18 the age gap is closer, averaging 9 years (median 38 years for HSIL and 47 years for invasive carcinoma) and very close for HPV 45, averaging 1 year (median 42 years for HSIL and 43 years for invasive carcinoma), whereas it spans over 20 years for HPV 31 and HPV 33 (Tjalma et al. 2013). The estimation of the rate of progression to invasion is more complicated, as discussed in detail in subsection “Behavior of SIL” in chapter “Precancerous Lesions of the Cervix.” A meta-analysis of studies correlating cervical cytology with follow-up outcome reported a rate of 1.44% (95 CI: 0%, 3.95%) of invasive cancer at 24 months following HSIL diagnosis on Pap (Melnikow et al. 1998). All these numbers apply to the general population. The progression rate is higher in immunosuppressed individuals (Petry et al. 1994).
Molecular genetic studies have provided evidence that HPV inactivates some of the crucial mechanisms regulating the cellular mitotic cycle. As a consequence, the virus causes a cascade of uncontrolled genetic events which may lead to malignant transformation of the host cell. In order to replicate, HPV has to induce DNA synthesis in host cells. Since HPV replicates only in maturing, already nonproliferating squamous cells, the virus has to reactivate the mitotic cycle in these cells. Experimental molecular genetic studies have shown that HPV activates the cell cycle by interfering with the function of two guardians of the cell cycle, retinoblastoma protein (pRB) and p53 (McLaughlin-Drubin and Munger 2009).
Analysis of chromosomal copy number alterations in cervical carcinomas using comparative genomic hybridization demonstrates gains at chromosome 3q in approximately half of cancer cases and losses at 3p and 11q in one-third of tumors (Thomas et al. 2014). Chromosome arm 3q encodes for the telomerase RNA component (TERC) gene and phosphatidylinositol 3-kinase, catalytic subunit α gene (PIK3CA). Both genes play an important role in cell growth and apoptosis; 3q gains lead to overexpression of both gene products resulting in increased cell growth and decreased apoptosis (Wang et al. 2014; Ma et al. 2000). Loss of chromosome arm 3p may lead to deletion of fragile histidine triad (FHIT), a tumor suppressor gene, which main function is induction of apoptosis (Butler et al. 2002).
Mutational analysis of cervical tumor samples demonstrates high frequency of activating mutations of PIK3CA and KRAS genes. A study of 80 tumor samples using high-throughput genotyping platform interrogating for 1250 known mutations in 139 cancer genes demonstrated validated mutations in 60% of tumors (Wright et al. 2013). The highest mutation rate (31.3%) was identified in PIK3CA, followed by KRAS (8.8%) and epidermal growth factor receptor (EGFR) (3.8%). PIK3CA mutations were identified in both adenocarcinomas and squamous cell carcinomas (25% and 37.5%, respectively). However, KRAS mutations were identified only in adenocarcinomas (17.5%). No associations between HPV 16 or HPV 18 and somatic mutations were identified (Wright et al. 2013). A whole-exome sequencing analysis of 79 cervical squamous cell carcinomas in comparison to paired normal samples showed recurrent mutations of EP300 gene (E1A binding protein p300) in 16% tumors, FBXW7 (F-box and WD repeat domain containing 7) in 15% tumors, PIK3CA (phosphatidylinositol 3-kinase, catalytic subunit α) in 14% tumors, HLA-B (major histocompatibility complex, class I, B) in 14% tumors, MAPK1 (mitogen-activated protein kinase 1) in 6% tumors, PTEN (phosphatase and tensin homologue) in 6% tumors, NFE2L2 (nuclear factor, erythroid 2-like 2) in 4% tumors, and STK11 (serine/threonine kinase 11) in 4% tumors. In 24 analyzed adenocarcinomas, the significantly mutated genes included PIK3CA (16%), ELF3 (E74-like factor 3) (13%), KRAS (8%), and CBFB (core-binding factor, beta subunit) (8%) (Ojesina et al. 2014). Understanding the molecular underpinnings of cervical cancer will be the basis for design of future targeted therapies.
Cervical carcinoma can occur anytime between the second and tenth decade of life, with age distribution showing a bell curve with peak at 50 years (NCI 2017; Tjalma et al. 2013). The mean and median age of women with cervical carcinoma is 50–51 and 49 years, respectively (de Sanjose et al. 2010; NCI 2017; Tjalma et al. 2013). Patients with invasive squamous cell carcinoma are on average 15 years older than patients with HSIL (Tjalma et al. 2013). The median age at death is 57 years (NCI 2017).
The presenting symptoms of patients with invasive carcinoma of the cervix depend on the size and stage of the lesion. In countries without effective screening programs, the patients present with bulky, late-stage disease. Nearly all these patients have clinically visible cancers, and nearly all have abnormal vaginal bleeding or serosanguinous discharge due to tumor infiltration of the cervical stroma. The bleeding may be postcoital, intermenstrual, or postmenopausal. Patients with tumors extending from the cervix into the parametria present with classic triad of sciatica, lower extremity edema, and renal failure, due to obstruction of the ureters resulting in hydronephrosis. Weakness, pallor, weight loss, rectal pain, and hematuria are other symptoms and signs of either locally advanced or metastatic disease.
In the United States, most patients present with stage I disease (52.7%) (Jessup et al. 1996). Patients with early invasive carcinoma are either asymptomatic or complain of intermittent spotting or bleeding following intercourse or douching. Primary presentation correlates with disease stage and survival. Patients who present with an abnormal Pap smear have a 3-year disease-free survival of 96% as compared to 51% for those presenting with vaginal bleeding and 21% for those presenting with pain (Pretorius et al. 1991). Cervical cancers are staged based on lesion size on cone biopsy, physical examination, and imaging studies. If the physical examination indicates stage >IB1, CT or MRI of the abdomen and pelvis is typically done to identify metastases, estimate tumor volume, and determine parametrial involvement. If MRI and CT are not available, cystoscopy, sigmoidoscopy, and IV urography, when clinically indicated, may be used for staging.
Cervical cancer is the most common gynecologic cancer in pregnancy, but the incidence is low (1.5–12 per 100,000 pregnancies) (Jones et al. 1996). The mean age of pregnant women with invasive cervical cancer is 32 years, which is considerably lower than that of women in the general population with invasive cervical cancer. Women who are pregnant usually present with early-stage tumors; in one series 83% were stage I (Jones et al. 1996). Most women present with abnormal Pap findings as cervical cytology exam is a routine part of the initial prenatal evaluation. One-third of patients are diagnosed during first, second, and third trimester each, respectively (Jessup et al. 1996). The treatment of pregnant patients depends on the clinical stage and gestational age. Approximately 37% of patients carry the pregnancy to fetal maturity (Jessup et al. 1996). In general, prognosis is not altered by pregnancy.
The gross appearance of invasive squamous cell carcinoma varies widely. Early lesions may be focally indurated, ulcerated, or present as a slightly elevated, granular area that bleeds readily. Colposcopic examination usually reveals atypical, tortuous vessels varying widely in size and configuration. Approximately 98% of early carcinomas are localized within the transformation zone, with variable degrees of encroachment onto the neighboring native portio. Advanced tumors are endophytic or exophytic. Endophytic carcinomas are ulcerated or nodular; they tend to develop within the endocervical canal and frequently invade deeply into the cervical stroma to produce an enlarged, hard, barrel-shaped cervix. In some patients with endophytic carcinomas, the cervix appears grossly normal. The exophytic varieties of cervical carcinoma have a polypoid or papillary, friable appearance.
The current WHO classification divides invasive squamous cell carcinomas into two groups, keratinizing and nonkeratinizing (Kurman et al. 2014). In addition, it lists separately rare histologic variants of squamous cell carcinoma, namely, basaloid, warty, verrucous, papillary, squamotransitional, and lymphoepithelioma-like carcinoma. The most common tumor types are keratinizing and nonkeratinizing, and clinical references pertaining to cervical carcinoma refer mainly to these two tumor types.
Keratinizing and Nonkeratinizing Carcinoma
Keratinizing carcinomas are characterized by the presence of well-differentiated squamous cells that are arranged in nests that vary in size and configuration. The defining feature of keratinizing carcinomas is the presence of keratin pearls within the epithelium (Fig. 7). Keratin pearls are composed of clusters of squamous cells that have undergone keratinization and are arranged in a concentric nest. The neoplastic squamous cells not forming keratin pearls frequently have abundant eosinophilic cytoplasm and prominent intracellular bridges. The nuclei are often enlarged, but mitotic figures are not numerous.
The most commonly used grading system for squamous cell carcinoma divides tumors into three groups, well differentiated (grade 1), moderately differentiated (grade 2), and poorly differentiated (grade 3). Most squamous cell carcinomas are moderately differentiated (grade 2), followed by poorly differentiated (grade 3) and well differentiated (grade 1).
Tumors are graded as well differentiated (grade 1) when the cells appear mature, with abundant eosinophilic cytoplasm (Fig. 7). Typically, there are keratin pearls in the center of neoplastic nests. Individual cell keratinization (dyskeratosis) characterized by intense cytoplasmic eosinophilia may also be present. The cells are tightly packed and have well-developed intercellular bridges. The nuclei are large, irregular, and hyperchromatic. Mitotic figures are present most notably at the periphery of the advancing epithelial nests. The stroma is often infiltrated by chronic inflammatory cells, and occasionally a foreign body giant cell reaction is observed.
Tumors are graded as moderately differentiated (grade 2) squamous cell carcinomas when the neoplastic cells are more pleomorphic than in grade 1 tumors, have large irregular nuclei, and have less abundant cytoplasm. The cellular borders and intercellular bridge appear indistinct. Keratin pearl formation is rare, but individual cell keratinization is seen in the center of nests of tumor cells. Mitotic figures are more numerous than in grade 1 carcinomas.
Poorly differentiated (grade 3) squamous cell carcinomas are diagnosed when the tumors are composed of cells showing little, if any, squamous maturation (Fig. 8). The nuclear to cytoplasmic ratio is high, the cytoplasm is scant, and indistinct and the cellular bridges are not seen. Tumor cells have hyperchromatic oval nuclei with coarse chromatin. Mitoses are abundant, and areas of central necrosis in tumor nests are present. Poorly differentiated tumors are occasionally composed of large, highly pleomorphic cells with giant, bizarre nuclei and abnormal mitotic figures.
Although in some studies tumor grade was associated with patient’s survival, most studies have failed to confirm that histopathologic grade influences clinical outcome (Zaino et al. 1992). A study by the Gynecologic Oncology Group (GOG) evaluated a number of different tumor-grading systems including those proposed by Warren, Reagan, and Broder in surgically treated stage IB cervical cancers. Although there was good reproducibility between observers, none of the grading systems had prognostic significance. Nuclear grade, degree of keratinization, mitotic activity, pattern of infiltration, and degree of lymphoid response all lacked prognostic significance (Zaino et al. 1992).
Immunohistochemical Staining and HPV In Situ Hybridization
Cervical squamous cell tumors are positive for a range of cytokeratins, including CK7, as well as CK 4, 5, 6, 8, 13, 14, 16, 17, 18, and 19 (Smedts et al. 1992). Cervical squamous cell carcinomas are in general negative for PAX-8, in contrast to adenocarcinomas which are positive in 87% of cases (Tacha et al. 2011; Ozcan et al. 2011). Cervical squamous tumors are also positive for p63 and its isoform, p40, in close to 100% of cases (Nemejcova et al. 2015). In a study of 250 invasive carcinomas, a strong, diffuse p63 expression was present in 97% of squamous cell carcinomas, including 91% of small cell nonkeratinizing squamous cell carcinomas. In contrast, small cell neuroendocrine carcinomas either did not show p63 staining or only had focal expression (<30% of tumor cells) (Wang et al. 2001). In addition, cervical squamous tumors show positivity for CEA in 90.6% cases, D2-40 in 53.6% cases, estrogen receptors in 10.6% cases, and progesterone receptors in 1.2% cases (Nemejcova et al. 2015).
Spread and Metastases
Squamous cell carcinoma of the cervix spreads principally by direct local invasion of adjacent tissues and lymphatics and less commonly through blood vessels. Initially, the tumor grows by extending along tissue planes into the paracervical and parametrial areas and into the broad and uterosacral ligaments. Ultimately, lateral spread may reach the bony pelvis, encompassing and obstructing one or both ureters. Direct extension may also involve the uterine cavity and vagina, with extension into the urinary bladder and rectum, resulting in vesicovaginal and rectovaginal fistulas.
The spread of cervical cancer via lymphatics occurs relatively early and is present in 25–50% of patients with stage IB and II carcinomas. The most common sites of lymph node metastases are the internal iliac, obturator, external iliac, and common iliac lymph nodes. Later in the course of the disease, extension to the lateral sacral, para-aortic, and inguinal nodes can occur. Isolated invasion of the sacral, external iliac, and hypogastric nodes is occasionally observed. Distant lymph node metastases above the diaphragm including the supraclavicular lymph nodes are uncommon but may occur in advanced disease. Hematogenous dissemination is the least common metastatic pathway of cervical carcinoma. Blood-borne metastases to the lung, liver, bone, heart, skin, and brain may be seen in advanced, stage IV tumors.
Ureteral obstruction caused by tumor invasion of the ureteral wall or by compression due to tumor in periureteral lymphatics leads to hydroureter, hydronephrosis, hydronephrotic renal atrophy, pyelonephritis, and loss of renal function. Obstruction of both ureters results in uremia and used to be a leading cause of death from cervical cancer. Other major causes of death in order of frequency include peritonitis caused by obstruction and bowel perforation, respiratory failure associated with pulmonary metastasis, massive edema, hemorrhage, cardiac failure, massive venous thrombosis, pulmonary embolism, and complications of radiation therapy.
Differential Diagnosis of Primary and Metastatic Tumors
Histologically, the lesions most commonly confused with invasive squamous cell carcinomas are squamous metaplasia or HSIL with extensive endocervical gland involvement, gestational decidual reaction with degenerative features, trophoblastic lesions such as placental site nodules, placental site trophoblastic tumor, and epithelioid trophoblastic tumor.
Endocervical glands involved by either squamous metaplasia or HSIL have rounded, smooth contours and lack the irregular margins, scalloping, and stromal desmoplasia (described in detail in subsection “Superficially Invasive Squamous Cell Carcinoma”). A decidual reaction with degenerative features lacks mitotic activity and proliferative activity as assessed with a Ki-67 immunostain and lacks positivity for cytokeratin. Placental site nodules appear as well-circumscribed nodules or plaques containing intermediate trophoblastic cells. These cells lack mitotic activity and are arranged in nests embedded in hyaline material. In contrast, cervical cancer cells show infiltration of the stroma. It can be challenging to differentiate epithelioid trophoblastic tumor and placental site trophoblastic tumor called collectively intermediate trophoblastic tumors (ITTs) from cervical squamous cancer. All of these entities may show a histologic pattern of infiltrating single cells and small tumor nests with large polygonal cells with eosinophilic cytoplasm. Immunohistochemical markers such as p16 and p63 may show overlapping positivity between trophoblastic tumors and cervical cancer; however ITTs do not have diffuse p16 or diffuse p63/p40, whereas they show HLA-G as well as HSD3B1 staining, so they can usually be distinguished with these markers (Kalhor et al. 2009; Mao et al. 2008). Other useful markers for identifying trophoblastic tumors include hCG and hPL, with only rare positivity of these markers seen in cervical squamous cell carcinoma (Kalhor et al. 2009).
While evaluating sections of squamous cell carcinoma, it is important to carefully search of any evidence of glandular differentiation, i.e., an adenosquamous component. Tumors dominated by the squamous component can be misclassified as purely squamous rather than as adenosquamous, and metastases from such adenosquamous carcinomas may be comprised of only the glandular component. This can lead to potential erroneous interpretation of a glandular tumor as a new/independent primary adenocarcinoma rather than as a metastasis from a previously diagnosed squamous cell carcinoma that was in fact an adenosquamous carcinoma.
Occasionally, it may be difficult to distinguish poorly differentiated squamous cell carcinomas composed of small, basaloid cells from small cell carcinoma of the neuroendocrine type, lymphoma, and malignant melanoma. Small cell carcinomas of the neuroendocrine type typically invade the stroma diffusely as individual cells or as small discohesive nests and show extensive crush artifact. They frequently form rosettes or trabeculae, and the cells characteristically have smudged, intensely hyperchromatic nuclei and lack nucleoli. In contrast, poorly differentiated squamous cell carcinomas invade as cohesive nests. In difficult cases immmunohistochemical analysis of p63 and p40 expression can be used to confirm squamous differentiation (Wang et al. 2001), and neuroendocrine markers such as chromogranin, synaptophysin, and CD56 can confirm the diagnosis of a neuroendocrine tumor. Malignant melanomas and lymphomas are p63-negative and cytokeratin-negative.
Squamous cell carcinomas that contain large amounts of cytoplasmic glycogen can sometimes be confused with clear cell carcinomas (Fig. 7). Like clear cell carcinomas, cells in these tumors have clear cytoplasm and distinct cell membranes. However, squamous cell carcinoma with clear cytoplasm lacks the characteristic hobnail cells and the papillary or tubulocystic areas that are typical of clear cell carcinomas. In squamous cell carcinoma, a careful search of multiple sections will usually detect areas with unambiguous squamous differentiation.
In metastatic squamous tumors of uncertain primary, immunohistochemical analysis of p16 expression can be useful for determining that a metastatic squamous cell carcinoma is likely of anogenital tract rather than pulmonary origin because high-risk HPV-related anogenital tract squamous cell carcinomas have diffuse p16 expression, whereas lung squamous cell carcinomas do not (Pereira et al. 2011). Although diffuse p16 expression serves as a surrogate marker for the presence of high-risk HPV in cervical tumors, it is important to remember that similar p16 expression can occur in some types of non-HPV-related neoplasms. When available, high-risk HPV detection using sensitive methods such as polymerase chain reaction or amplified in situ hybridization is useful for confirmation that diffuse p16 expression is indicative of a high-risk HPV-related tumor and supportive of cervical/anogenital tract origin in the appropriate clinical setting (Weichert et al. 2009).
Prognostic Significance of Histopathologic Findings
Stage is the most important prognostic factor in cervical carcinoma. Histologic typing and grading have little influence on survival within any stage. The most significant pathologic prognostic factors in women with stage IB and IIA squamous cell carcinoma are tumor size, depth of invasion, presence of LVSI, and nodal status. Consequently, the pathology report should state the size of the tumor in three dimensions, depth of invasion, presence or absence of LVSI, as well as presence or absence of parametrial involvement and resection margin involvement.
Reported 10-year disease-free survival rates are 90% for IB tumors <2 cm, 76% for 2.1–4 cm, 61% for 4.1–5 cm, and 47% for >5 cm. Similar relation between tumor size and survival is seen in stage IIA, with the survival rates of 93%, 63%, 39%, and 59%, respectively (Perez et al. 1998). In the GOG series of patients with stage IB squamous cell carcinoma treated surgically, women whose tumors invaded the inner third of the cervical stroma had a 98% 5-year progression-free survival, whereas for those whose tumors invaded the outer third, the progression-free survival was only 63% (Zaino et al. 1992). In the same GOG series, the 5-year progression-free survival was 70% in tumors with LVSI compared to 83% in cases without LVSI (Zaino et al. 1992). In most series of surgically treated stage IB cancers, LVSI has proven to be somewhat less important than tumor size and depth of invasion. In a series of 978 stage IB–IIA patients, the median survival was 5.3 years for those with negative nodes and decreased to 3.2 and 1.3 years for those with positive pelvic and para-aortic lymph nodes, respectively (Averette et al. 1993). Parametrial extension correlates with tumor recurrence and decreased survival. In 69 radical hysterectomies performed for clinical stage IB and IIA carcinomas, parametrial involvement was observed histologically in 31% of IB tumors, 63% of IB2, and 58% of IIA tumors. Survival was significantly higher for the patients with no parametrial extension compared with those with positive parametrial findings (100% vs. 78%) (Benedetti-Panici et al. 2000).
Prognostic Significance of HPV Genotype
Several studies showed a significant correlation between detection of HPV 18 in tumor tissue and poor prognosis in early-stage cervical cancer (Schwartz et al. 2001). In a study of 399 stage IB–IV cancers, detection of HPV 18 DNA in tumor sample was associated with an approximately sixfold increase in the risk of death in stage IB and IIA disease. This association was seen in all histologic tumor types assessed together, as well as in squamous cell carcinomas evaluated separately. There was no difference in survival based on HPV type in patients with advanced cancer stages IIB–IV (Schwartz et al. 2001).
Recent studies have suggested that patients infected with multiple HPV types have a significantly shorter cancer-specific survival and do not respond as well to radiation treatment as patients with a single HPV type. The confounding factor in this observation may be the presence of HIV infection, as HIV positivity is associated with detection of multiple HPV types in anogenital tumor samples and poor response to treatment (Meyer et al. 2013).
Treatment and Prognosis
The three basic therapeutic modalities for squamous cell carcinoma are surgery, radiation, and combinations of radiation and surgery or radiation with concurrent chemotherapy. Recently targeted therapies have also shown some promise. Treatment options for stage IA tumors are shown in Table 5. In stage IB and IIA patients, virtually identical results are seen with radiation therapy versus radical hysterectomy plus bilateral pelvic lymphadenectomy followed by tailored adjuvant therapy. For stage IIB to stage IVA, the treatment involves chemoradiation with brachytherapy, and for stage IVB the treatment consists of chemotherapy. Chemotherapy for cervical cancer includes many regimens, most of which use cisplatin either alone or in combination with other drugs.
For patients who desire to preserve their fertility and who present with stage IA2 and IB1 tumors, particularly those measuring <2 cm in diameter, radical trachelectomy combined with laparoscopic pelvic lymphadenectomy has been widely accepted treatment option. The 5-year disease-specific survival in women with radical trachelectomy as compared to radical hysterectomy is 92% versus 91%, respectively (Diaz et al. 2008). The pregnancy rate after trachelectomy is reported as 88%, with 66% of pregnancies resulting in live births (Diaz et al. 2008).
The majority of tumor recurrences appear within 2 years after initial therapy. Chemotherapy is used for patients with metastatic or recurrent cancer previously treated with surgery or radiation therapy. Response rates for cisplatin-based therapy in these recurrent or late-stage patients range from only 20% to 30%, and overall survival is less than 10 months. Overall, of patients who die of disease, 85% succumb within 3 years of diagnosis. Since effective treatment options are limited for advanced cervical cancer, therapies targeting cervical cancer signaling pathways are currently being investigated.
The 5-year survival for treated stage I patients is 95%, 60–80% for stage II, 37% for stage III, and < 20% for patients with stage IV disease (Jessup et al. 1996). The survival rates are reduced even in patients with low-stage disease when metastases to lymph nodes are present and the survival rates correlate with the number of positive nodes.
Clinical Trials of Targeted Therapy
Numerous studies have investigated the immunohistochemical expression of potential therapeutic targets in squamous cell carcinomas of the cervix, including expression of cyclooxygenase 2 (COX-2), epidermal growth factor receptor (EGFR, HER1), and vascular endothelial growth factor (VEGF). Despite biologic promise, two phase II trials testing specific COX-2 inhibitors as radiosensitizers in locally advanced cervical cancer demonstrated increased toxicity with no change in therapeutic effect (Gaffney et al. 2007). EGFR inhibitors such as cetuximab, a chimeric monoclonal antibody that binds to EGFR, also failed to show value in clinical trials (de la Rochefordiere et al. 2015). Intratumoral protein levels of VEGF are increased in cervical cancer as compared to normal cervical tissue, and in phase II clinical trials, bevacizumab (Avastin®), a recombinant humanized anti-VEGF monoclonal antibody, was shown to be active as second- and third-line treatment in patients with recurrent cervical cancer (Monk et al. 2009).
Frequent molecular alterations of PI3K pathway in cervical cancer led to phase I clinical trial of matched therapy targeting activated PI3K/AKT/mTOR pathway. The results showed significantly longer progression-free survival in treatment group (6 mo) compared to non-matched therapy (1.5 mo). The detection of PIK3CA mutations in squamous tumors was associated with a significantly longer overall survival (median, 9.4 months) compared to absence of PIK3CA mutations (Hou et al. 2014).
A most recent approach is aimed at T cell immune checkpoint inhibitors. By blocking inhibitory molecules, these agents are meant to enhance anticancer T cell immune responses. Several clinical trials are currently underway including phase II trial for patients with cervical, vulvar, or anal cancer using pembrolizumab (Keytruda®), a PD-1 antibody; a phase I/II trial for patients with viral-associated cancers, including cervical, vaginal, and vulvar cancer using nivolumab (Opdivo®), also a PD-1 antibody; and a phase I study for patients with locally advanced cervical cancer using chemoradiation followed by ipilimumab (Yervoy®), an anti-CTLA-4 antibody (Institute, C.R. 2018).
Basaloid carcinoma is listed as a separate tumor subtype in the recent WHO classification; however, no systematic study of this tumor type in the cervix is available, and publications include rare case reports and one opinion paper (Grayson and Cooper 2002). Given that cervical carcinomas with basaloid features are always or nearly always high-risk HPV-positive and that there is a spectrum of basaloid morphology found in less differentiated cervical squamous cell carcinomas, basaloid carcinoma may not really be a separate entity (Meyer et al. 2013). Due to lack of publications with significant follow-up, it is unclear whether the behavior of cervical basaloid tumors differs significantly from that of conventional cervical squamous cell carcinomas of similar clinical stage.
Warty (Condylomatous) Carcinoma
Warty (condylomatous) carcinoma is a variant of squamous cell carcinoma of the cervix that has prominent exophytic (condylomatous) as well as infiltrative growth and cytologically showing squamous maturation and keratinization with conspicuous koilocytic changes. The presence of infiltrative invasion and marked cytologic atypia distinguishes this tumor from condyloma and verrucous carcinoma. The tumor is very rare in the cervix and more common in the vulva. Although clinical experience with these tumors is very limited, they appear to behave less aggressively than conventional squamous cell carcinomas, based on a single reported series of nine patients, all of whom presented with low-stage disease (IA to IIA) and who were disease-free following surgical treatment (Cho et al. 1998).
Verrucous carcinoma is a rare variant of squamous cell carcinoma. In the female genital tract, the tumor is more common in the vulva, but rare cases have been described also in the cervix. In general, verrucous carcinomas of the anogenital tract are negative for HPV (del Pino et al. 2012), and cytologically these tumors lack hallmarks of HPV infection (atypia, koilocytes); however, occasionally low-risk HPVs are detected (Yorganci et al. 2003). In single case reports, patients treated with surgery alone had uneventful follow-up (Yorganci et al. 2003).
Papillary Squamous Cell Carcinoma and Squamotransitional Carcinoma
Cervical tumors with a papillary appearance are rare variants of carcinoma of the cervix. A comprehensive study from the Armed Forces Institute of Pathology of 32 cases of papillary cervical carcinomas reported that these tumors can be subdivided into three groups based on their histologic appearance: predominately squamous, predominantly transitional, and mixed squamous and transitional (Koenig et al. 1997). The current WHO classification separates papillary tumors with squamous morphology into category of papillary squamous cell carcinoma and those with transitional morphology into category of squamotransitional carcinoma (Kurman et al. 2014). The distinction between these two categories is subjective, and they are likely to represent a spectrum of the same tumor type. Transitional morphology may be an appearance of less mature squamous differentiation. Many reports combine these two categories together under the name of papillary squamotransitional carcinoma. Both papillary squamous cell carcinomas and squamotransitional carcinomas have identical immunohistochemical staining patterns with cytokeratins 7 and 20. Both these tumors are positive for cytokeratin 7 and negative for cytokeratin 20, which is typical of cervical squamous cancers and unlike transitional cell tumors of the urinary tract (Koenig et al. 1997). Both papillary squamous cell carcinoma and squamotransitional carcinoma were reported to be positive for HPV 16 (Lininger et al. 1998).
A first large series was described by Randall et al., who reported that these tumors had a histologic resemblance to papillary carcinoma of the urinary tract (Randall et al. 1986). Of nine patients with follow-up, four died of disease (stage II to IV). In a subsequently reported series of 12 patients, 3 patients with advanced stages (IIB and IIIB) died of disease (Koenig et al. 1997). No apparent differences were observed in the clinical behavior between papillary squamous and squamotransitional tumor subtypes (Koenig et al. 1997). Mirhashemi et al. ((2003) recorded only 1 death in 12 patients with papillary carcinoma. These reports suggest that papillary squamotransitional tumors may have similar behavior to conventional squamous cell carcinomas.
Papillary squamotransitional carcinoma can be mistaken for a papillary HSIL on a superficial biopsy. Because papillary squamotransitional carcinoma is capable of acting aggressively, it is important that a cone biopsy is carefully evaluated for invasion whenever a papillary HSIL is diagnosed. These tumors can also be mistaken for squamous papilloma or condyloma acuminatum especially in cases with more mature squamous differentiation. Evaluation of the underlying stroma for evidence of invasion is paramount for accurate diagnosis.
Lymphoepithelioma-like carcinomas (LELC) are a distinctive subset of squamous cell carcinomas of the cervix that are typically well circumscribed and composed of undifferentiated cells surrounded and infiltrated by a dense stromal inflammatory infiltrate. These tumors are reported to have a better prognosis than typical squamous cell carcinoma. LELC represent only 0.7% of all primary cervical malignancies in Western countries but up to 5.5% of cervical cancers in some series from Asia (Tseng et al. 1997). The pathogenesis of cervical LELC depends on the patient’s ethnic background. The Epstein-Barr virus (EBV) has been suggested as a causative agent and has been detected in 75% of cervical LELC cases in Asian women (Tseng et al. 1997). In contrast, HPV 16 and HPV 18, but not EBV, are detected in cervical LELC in Caucasian women in Western countries (Bais et al. 2005; Noel et al. 2001).
Adenocarcinomas of the cervix comprise a heterogeneous group of neoplasms that display a variety of histologic patterns (see Table 1). Because these different differentiation patterns are sometimes admixed, the histologic classification of these tumors is based on the predominant cell type. If additional histologic components comprise at least 10% of the tumor, some authors recommend classifying the tumor according to the predominant pattern and listing the individual components as part of the diagnosis.
The most common cervical adenocarcinoma type is the endocervical adenocarcinoma, also referred to as usual type. Less prevalent tumor types include intestinal, endometrioid, clear cell, and gastric adenocarcinomas. Although some studies have reported cervical endometrioid adenocarcinomas to be as prevalent as the usual-type adenocarcinomas, when using strict diagnostic criteria, the endometrioid tumor subtype accounts for less than 10% of all cervical adenocarcinomas (Young and Clement 2002; Pirog et al. 2000).
While the absolute incidence of invasive squamous cell carcinoma has been decreasing in the Western countries due to effective screening, the incidence of adenocarcinoma in situ and invasive adenocarcinoma has been increasing (Smith et al. 2000; Vizcaino et al. 1998). This trend for adenocarcinoma may be related to a general increase in HPV-related infections and lower sensitivity of the Pap test for detection of neoplastic glandular lesions. Adenocarcinomas develop higher in the endocervical canal and may be present deep in the endocervical clefts which can result in inadequate cytologic sampling. Between 1973 and 1996, the overall age-adjusted incidence of invasive cervical cancer in the United States decreased by 36.9%. At the same time, the age-adjusted incidence of adenocarcinoma increased by 29.1% (Smith et al. 2000). As a result, the ratio of invasive squamous cell carcinoma to adenocarcinoma has been shifting. In the 1950s and 1960s, approximately 95% of all invasive cervical carcinomas were classified as squamous cell carcinomas and only 5% as adenocarcinomas (Mikuta and Celebre 1969). However, starting in the 1970s, there was a gradual shift where the ratio of squamous cell carcinomas decreased to 75–80%, while the remaining 20–25% of the cases included various types of adenocarcinomas, adenosquamous carcinomas, and other epithelial tumors (Vizcaino et al. 1998). However, worldwide and in the developing countries, adenocarcinomas still account for a minor fraction of less than 10% of all invasive carcinomas (de Sanjose et al. 2010).
Pathogenesis and Epidemiologic Risk Factors
Different histologic subtypes of cervical adenocarcinoma arise along different pathogenetic pathways. The most common tumor subtypes, namely, endocervical, intestinal, endometrioid, and villoglandular, accounting together for approximately 90% of all adenocarcinomas, are related to infection with high-risk HPVs (Pirog et al. 2000). Rare histologic subtypes of adenocarcinoma such as gastric (including minimal deviation adenocarcinoma), clear cell, serous, and mesonephric are not related to HPV infection, and their proposed pathogenesis is described in the respective subsections (Pirog et al. 2000; Park et al. 2011).
In HPV-related adenocarcinomas, only a narrow spectrum of HPV genotypes is detected. Three HPV types, HPV 16, 18, and 45, account for over 90% of HPV-positive cases and individually are detected in the following order – HPV 16, 41–50% cases; HPV 18, 32–37% cases; and HPV 45, 2–12% cases (de Sanjose et al. 2010; Pirog et al. 2000). While HPV 16 dominates in squamous cell carcinomas, HPV 18 is frequently found in association with adenocarcinomas, adenosquamous carcinomas, and neuroendocrine tumors (de Sanjose et al. 2010).
Many of the epidemiologic risk factors for the development of adenocarcinoma of the cervix are the same as those described for invasive squamous cell carcinoma, such as multiple sexual partners, young age at first intercourse, and history of sexually transmitted diseases (Brinton et al. 1993; Castellsague et al. 2006). Use of oral contraceptives for more than 5 years has been shown to increase the risk of adenocarcinoma by 1.5-fold, while use of an intrauterine contraceptive device has been found to be protective (Castellsague et al. 2006). Tobacco smoking is associated with increased risk of squamous cell but not adenocarcinoma of the cervix (Castellsague et al. 2006).
Clinical Presentation and Gross Findings
In the Western countries most patients with invasive adenocarcinoma are asymptomatic and have early-stage tumors detected with the Pap test. Of the patents who did not participate in screening, or when early lesions are not sampled and missed with Pap screening, the most common presenting symptom is abnormal vaginal bleeding, occurring in about 75% of patients. Occasionally women present with mucoid vaginal discharge or with pelvic pain. The majority of invasive cervical adenocarcinomas arise in the transformation zone. On gross examination the patients have either fungating, polypoid, or papillary mass, or the cervix is diffusely enlarged or nodular. In 15% patients no gross lesions are visible. Although the majority of patients with grossly unapparent tumors have early-stage tumors, some have deep invasion because the carcinoma developed higher within the endocervical canal.
Early Invasive Adenocarcinoma
It is important to emphasize that recognizing early invasive, stage IA adenocarcinoma is often problematic, and less is known about the behavior of early invasive adenocarcinoma compared to squamous cell carcinoma. When referring to early invasive adenocarcinoma, four histologic tumor types are considered, namely, endocervical, intestinal, endometrioid, and villoglandular. No systematic information is available for early invasive rare tumor subtypes including gastric, clear cell, serous, and mesonephric adenocarcinoma.
Most patients with early invasive adenocarcinoma have coexisting adenocarcinoma in situ and some also have HSIL. The irregular distribution and architecture of the normal endocervical crypts in the cervical stroma makes it difficult to differentiate between adenocarcinoma in situ and early stromal invasion. In addition, in cases with admixed adenocarcinoma in situ and superficially invasive adenocarcinoma, the depth of the invasion may be difficult to measure.
After confirming the presence of invasion, the depth of invasion and horizontal extent of the invasive component should be measured. In most cases, the depth of invasion is measured from the epithelial surface rather than from the point of origin in AIS, which may be impossible to determine. For tumors in which the in situ and invasive components are intimately associated, it is recommended that the entire tumor thickness, including both the in situ and invasive components, should be measured because it can be difficult or even impossible to determine where the invasion begins (Ostor 2000). In exophytic/papillary tumors, the entire tumor thickness should be reported. The horizontal dimension should be reported as the greatest continuous span of the invasive area measured parallel to the surface. Similarly, for tumors in which the in situ and invasive components are intimately associated, it is recommended that the entire horizontal extent, including both the in situ and invasive components, be measured. For cases with multifocal invasion, the depth/thickness and horizontal extent/width of each focus should be measured individually and reported, and the FIGO stage should be determined based on the measurement of the largest focus. In cases in which the invasive lesion spans adjacent blocks of tissue and co-localizes in the contiguous sections, it is recommended that the third measurement of circumference be estimated by multiplying the number of involved blocks by the approximate block thickness. In addition to tumor size, involvement of the margins and presence or absence of LVSI should be reported.
Clinical Behavior and Treatment
Early invasive adenocarcinoma accounts for 12% of adenocarcinomas (Ostor 2000). The average age of the patients is 39 years, approximately 6 years before the peak age for frankly invasive adenocarcinoma (Ostor 2000). The overall prognosis for early invasive adenocarcinoma, both stages IA1 and IA2, is excellent. In a meta-analysis totaling 436 cases of early invasive adenocarcinoma defined as 5 mm or less of invasion, only 2% (5 of 219) of patients undergoing lymphadenectomy had nodal metastases (Ostor 2000). A more recent review of the literature including 943 cases confirmed a low risk of adverse outcome in stage IA adenocarcinoma (Smith et al. 2002). In stage IA1 versus IA2, the incidence of lymph node metastases (1.45% vs. 1.73%), recurrence (1.54% vs. 1.96%), and death (0.85% vs. 1.12%) is not statistically different (Smith et al. 2002). In the same study there were no differences in recurrence and death due to disease between patients treated conservatively (with conization or simply hysterectomy) and those undergoing radical hysterectomy. Search of the US Surveillance, Epidemiology, and End Results database of all women with stage IA1 and IA2 cervical carcinoma diagnosed from 1988 to 2005 was used to compare treatment and outcomes of women with adenocarcinomas versus squamous cell carcinomas. Of 3987 women identified in the database, 988 were adenocarcinomas (24.8%). Survival for stage IA1 and IA2 adenocarcinomas was similar to that of women with squamous cell tumors of corresponding stage. For both stages of IA1 and IA2 adenocarcinomas, survival was similar in conization- and hysterectomy-treated patients: for patients with stage IA1 adenocarcinomas, 5-year survival was 96.9% (95% CI, 94.0–98.4%) in hysterectomy-treated patients as compared to 98.8% (95% CI, 91.5–99.8%) in patients treated with conization, and in stage IA2 adenocarcinomas, 5-year survival for women was similar for hysterectomy- (98.2%; 95% CI, 88.1–99.7%) and conization-treated patients (97.8%; 95% CI, 95.1–99.0%) (Spoozak et al. 2012). The authors concluded that prognosis in early invasive adenocarcinoma did not differ from the excellent prognosis in the squamous counterpart and that conization appeared to be adequate treatment for early invasive adenocarcinoma. The same conclusions were drawn from smaller studies with literature meta-analysis (Baalbergen et al. 2011). The recommendations from these studies were that patients with early invasive adenocarcinoma could be treated identically to patients with squamous tumors, and for stage IA1 and IA2 adenocarcinoma, a conization with negative margins was considered adequate treatment. In cases with LVSI, an additional lymphadenectomy was advised. For patients with stage IA2 adenocarcinoma with LVSI, a trachelectomy or radical hysterectomy with lymph node dissection was recommended (Baalbergen et al. 2011).
Few studies have examined margin status at conization and residual disease in a subsequent hysterectomy specimen. Approximately one-third of patients undergoing conization for early invasive adenocarcinoma had margin involvement (Poynor et al. 2006). Among the patients who underwent conization and had negative margins, none had residual disease in the subsequent hysterectomy specimen, while half of the patients with positive margins had residual tumor in hysterectomy (Poynor et al. 2006).
Endocervical Adenocarcinoma, Usual Type
Adenocarcinoma of endocervical type is the most common type of adenocarcinoma arising from the mucinous endocervical epithelium, and for that reason it is also referred to as “usual type” (Young and Clement 2002). The tumor accounts for approximately 75% of all cervical adenocarcinomas (Pirog et al. 2000). The average age of patients is 43–45 years (range 22–80) (Pirog et al. 2000; Holl et al. 2015). The precursor lesion is adenocarcinoma in situ, endocervical type. HPV is detected in approximately 90% of cases (Pirog et al. 2000; Holl et al. 2015).
The usual-type adenocarcinomas of the cervix are graded on the basis of architectural features, in a manner similar to that used for endometrial adenocarcinomas. Well-differentiated tumors are defined as those in which <5% of the tumor volume is composed of solid sheets of cells, the remainder of the tumor being glandular; in moderately differentiated tumors, 6–50% of the tumor is composed of solid sheets of cells; and in poorly differentiated tumors, >50% of the tumor is solid. Unlike in squamous cell carcinoma, the histologic grade of adenocarcinoma was shown to correlate with patients’ prognosis.
For staging purposes, measurement of tumor size should be obtained during gross examination. In microscopic tumors, the measurement of the depth of invasion can be problematic due to difficulty in establishing the point of origin of invasion (see discussion in section “Early Invasive Adenocarcinoma”). In most cases, the depth of invasion is measured from the epithelial surface. The presence of lymph-vascular space invasion should be reported.
Immunohistochemical Staining and In Situ Hybridization
Well-differentiated adenocarcinoma must be distinguished from benign conditions. A deceptive microcystic variant of usual-type endocervical adenocarcinoma can simulate type B tunnel clusters. Recognition of this variant can be hindered not only by the benign appearing architecture but also by the presence of largely denuded cystic glands. Additional difficulty may be encountered when, on occasion, benign endocervical glands, nabothian cysts, and mesonephric duct remnants are found extending deep in the stroma, sometimes even to the outer third of the cervical wall. Sometimes, desmoplasia around benign glands may develop as a result of a nonneoplastic process, such as a reaction to gland rupture and mucin extravasation. High-power examination of cytologic features is necessary to determine the benign nature of such glands. In contrast to usual endocervical adenocarcinoma, most benign glandular lesions retain expression of hormone receptors (with the exception of lobular endocervical glandular hyperplasia) and have low proliferation indices as assessed with immunostaining for Ki-67. Mesonephric remnants can be identified by their expression of GATA-3 (Howitt et al. 2015). The criteria for differentiation of invasive adenocarcinoma from adenocarcinoma in situ are described above in subsection “Early Invasive Adenocarcinoma.”
Biomarkers for distinguishing endocervical and low-grade endometrial adenocarcinomas
Endocervical adenocarcinoma (HPV-related)
Endocervical adenocarcinoma (non-HPV/gastric)
Low-grade endometrial adenocarcinoma (endometrioid, mucinous)
Wild typeb or aberrant/mutation typec
Negative to focally positived
Biomarkers for distinguishing endocervical and high-grade endometrial adenocarcinomas
Endocervical adenocarcinoma (HPV-related)
High-grade endometrial endometrioid carcinoma
Often: Wild typeb
Some: Aberrant/mutation typec
The vast majority of endocervical adenocarcinomas are high-risk HPV-related and exhibit diffuse, strong p16 expression due to complex molecular mechanisms resulting in p16 overexpression. In addition, endocervical adenocarcinomas often lack hormone receptor (ER/PR) expression. In contrast, endometrial endometrioid carcinomas are etiologically unrelated to high-risk HPV and display heterogeneous/patchy p16 expression ranging from limited to extensive, with the more positive tumors still usually having less than 80% of the tumor cells positive (nondiffuse pattern), with scattered negative foci or interspersed individual negative tumor cells present (Yemelyanova et al. 2009). Only a small subset of endometrioid carcinomas of all grades has diffuse p16 expression. Some endometrial endometrioid carcinomas with prominent mucinous/metaplastic-type differentiation can have more extensive to occasionally diffuse p16 expression, but the staining intensity is usually not as strong as is seen in high-risk HPV-related endocervical adenocarcinomas, and some negative patches are usually present if the sample is not too small (Ronnett, unpublished observations). Endometrioid carcinomas typically express hormone receptors (both ER and PR), but some tumors, particularly but not exclusively high-grade carcinomas, can lose expression. Therefore, analysis of p16 expression, alone or in combination with ER/PR, distinguishes high-risk HPV-related endocervical adenocarcinomas from low-grade endometrial endometrioid carcinomas in most cases (Yemelyanova et al. 2009).
Some pathologists find CEA and vimentin to be of some value in the distinction of high-risk HPV-related (usual-type) endocervical adenocarcinomas from endometrial endometrioid carcinomas, but others do not. Most endometrial endometrioid carcinomas are vimentin-positive and CEA-negative, and most HPV-related usual-type endocervical adenocarcinomas are vimentin-negative and CEA-positive. However, in practice use of these markers is problematic for several reasons: (1) expression of these markers can be focal; (2) CEA staining can be difficult to interpret because squamous elements (commonly seen in endometrial endometrioid adenocarcinoma) can be positive, and in addition, there can be apical/glycocalyceal staining in endometrioid adenocarcinomas, while some usual-type endocervical adenocarcinomas are negative; (3) it may be difficult to ascertain whether vimentin expression is actually within glands versus in closely apposed stroma; and (4) tumors with mucinous differentiation usually express CEA, regardless of their origin, and may be vimentin-negative.
Expression of p16 is also useful for identifying endocervical adenocarcinoma metastatic to the ovary. These ovarian metastases have a propensity to simulate primary ovarian endometrioid and mucinous tumors (atypical proliferative (borderline) tumors and carcinomas). Primary ovarian endometrioid and mucinous tumors, with few exceptions, are characterized by generally patchy p16 expression (or lack of expression), whereas metastatic high-risk HPV-related endocervical adenocarcinomas are diffusely/strongly positive. In the absence of a known primary endocervical adenocarcinoma, demonstrating high-risk HPV by in situ hybridization in metastatic lesions can be used for arriving at a definitive diagnosis (Fig. 24).
Metastatic adenocarcinoma to the cervix usually occurs in the setting of a patient with a known, widely metastatic primary lesion and is histologically characterized by a lack of surface involvement and widespread lymph-vascular involvement. In assessing whether a carcinoma is of primary endocervical origin or is metastatic in the cervix, the pathologist should evaluate the following morphologic features: (1) neoplastic growth pattern, (2) coexistent in situ changes, (3) cell type, and (4) immunohistochemical characteristics. Transition between in situ and invasive carcinoma provides evidence for a primary origin and is found in approximately 50% of primary cervical adenocarcinomas. Pax-8 positivity may be used to confirm primary müllerian origin of the tumor.
Prognostic Risk Factors
A recent study (Roma et al. 2016) found a correlation between the histologic pattern of invasion in usual-type adenocarcinoma and risk of lymph node metastasis and recurrences. In that study the tumors were subclassified into three groups based on their pattern of growth (Silva classification). The tumors with well-demarcated glands and lacking destructive stromal invasion or lymphatic invasion were classified as pattern A. Tumors with diffuse destructive stromal infiltration or solid tumors were classified as pattern C. Tumors with early destructive invasion from well-demarcated glands were classified as pattern B. The study consisted of 352 cases, and all pattern A and B tumors as well as 83% of pattern C tumors were stage I. No lymph node metastases or recurrences were seen in any of pattern A cases regardless of their stage (IA1, IA2, or IB) or depth of invasion. Metastases and recurrences were recorded in 4.4% and 1.2% of tumors with pattern B and 23.8% and 22.1% of tumors with pattern C, respectively. Based on these results, the authors proposed that patients with pattern A growth might be spared lymph node dissection, patients with pattern B growth could be considered for sentinel node sampling, and patients with pattern C were the only group requiring lymph node dissection (Roma et al. 2016).
Other reported prognostic indicators for cervical adenocarcinoma include tumor size, depth of invasion, involvement of lymph-vascular spaces, parametrial involvement, stage, age, and presence or absence of lymph node metastases (Eifel et al. 1990).
Clinical Behavior and Treatment
Adenocarcinoma of the cervix spreads in a fashion similar to squamous cell carcinoma, and, in general, both squamous and adenocarcinomas are treated similarly. The most commonly used therapeutic modalities for stage I and II adenocarcinoma are radiation alone, radiation with concurrent chemotherapy, radiation followed by simple hysterectomy, or radical surgery. Only a few studies have directly compared the therapeutic results achieved with invasive squamous cell and adenocarcinomas over the same time period and from the same institution; and these studies have produced conflicting data. Some studies have found that the overall 5-year survival rates are lower for adenocarcinoma (48–65%) than for squamous cell carcinoma patients (68%) (Eifel et al. 1990). Other comparison studies, as well as several population-based studies, have failed to confirm that prognosis and survival are affected by histologic type (Anton-Culver et al. 1992). Therefore, the prognosis of cervical adenocarcinoma relative to squamous carcinoma remains a controversial issue.
Endometrioid Adenocarcinoma of the Cervix
It is not clear whether endocervical adenocarcinomas with endometrioid-type differentiation are truly a distinct entity. Tumors with this appearance in which high-risk HPV is detected are most consistent with mucin-depleted forms of usual-type endocervical adenocarcinoma and are thus not a distinct entity. According to the literature on tumors reported as this type, the average age of patients is 50 years (Pirog et al. 2000). In different series, endometrioid carcinomas accounted for 7–50% of endocervical adenocarcinomas (Young and Clement 2002). Studies have also reported that HPV DNA is identified in 78–100% of endometrioid adenocarcinomas that are arising from the cervical squamocolumnar junction (Pirog et al. 2000; Jones et al. 2013). However, endometrioid adenocarcinomas arising from upper endocervix and lower uterine segment are typically HPV-negative (Holl et al. 2015). These observations indicate that there are problems in classification related to difficulty in distinction of true primary endocervical adenocarcinomas of usual type with mucin depletion from true primary endometrial endometrioid carcinomas. Use of ancillary techniques (discussed above) can resolve this problem. Most experts agree that true endometrioid adenocarcinomas of the cervix are rare, and only tumors with scant eosinophilic cytoplasm without apparent intracytoplasmic mucin should be classified as endometrioid (Young and Clement 2002).
The pattern of immunostaining of cervical endometrioid adenocarcinoma is similar to that of usual-type adenocarcinoma, with tumor cells demonstrating strong and diffuse p16 positivity and often negative staining for hormone receptors (Jones et al. 2013).
Differentiating cervical from uterine endometrioid adenocarcinoma may require use of immunostains (ER-/PR-/diffuse p16+ result supports the diagnosis of cervical primary). Rare cases may show mixed/overlapping immunostaining patterns. Careful assessment of imaging studies may be of help. Primary uterine corpus tumors are usually bulky tumors that have invaded the myometrium by the time they extend to the cervix and therefore cause uterine enlargement. In contrast, primary cervical adenocarcinomas often cause cervical enlargement in the absence of uterine enlargement. In addition, multiple sections may reveal atypical endometrial hyperplasia in primary endometrial tumors and either adenocarcinoma in situ or HSIL or foci with features of a typical endocervical carcinoma in primary endocervical tumors. In certain cases establishing the correct diagnosis may require knowing the HPV status of the tumor. If a tumor appears endometrioid and is truly HPV-negative and present in the lower uterine segment, then it is likely truly endometrial.
Intestinal Adenocarcinoma of the Cervix
The tumor accounts for approximately 8% of cervical adenocarcinomas. The average age of patients is 47 years (range 26–69) (Pirog et al. 2000). HPV DNA is identified in 83% of tumors (Pirog et al. 2000). The precursor lesion is adenocarcinoma in situ, intestinal type. Recent reports have described cases of intestinal adenocarcinoma in situ which were HPV-negative and occurred in older patients (Talia et al. 2014).
Intestinal adenocarcinoma displays immunohistochemical profile similar to adenocarcinoma of usual type (Saad et al. 2009). In addition, the malignancy may show an enteric immunophenotype with at least focal positivity for CDX2 and/or CK20 (Saad et al. 2009; Park et al. 2009). It has been observed, however, that CK7 positivity in these lesions is stronger than the staining for enteric markers (Park et al. 2009).
The differential diagnosis includes direct extension or metastasis from colonic adenocarcinoma. The presence of large, garland-shaped glands with incomplete lining and intraluminal necrosis is highly suggestive of a metastasis from the gastrointestinal tract. In equivocal cases immunohistochemistry should be used; negative immunostaining for PAX-8 and CK7 and positive enteric markers are indicative of spread from colonic adenocarcinoma.
Villoglandular Adenocarcinoma of the Cervix
Villoglandular adenocarcinoma is a well-differentiated variant of endocervical, endometrioid, or intestinal adenocarcinoma that occurs predominantly in young women and has an excellent prognosis (Jones et al. 1993b). The tumor accounts for approximately 3–6% of all adenocarcinomas. The entity was established as a separate diagnostic category based on a prognosis that is more favorable than that of usual-type endocervical adenocarcinoma. The average age of patients is reported at 33–41 years (range 21–61) with most patients younger than 40 years (Jones et al. 1993b). HPV DNA is identified in 100% of villoglandular adenocarcinomas (Jones et al. 2000).
The papillae can be either long and thin or thick and short. Small papillary tufts composed entirely of epithelial cells of the type characteristic of serous carcinomas of the ovary are absent. Beneath the papillary surface, the infiltrating portion of the tumor is composed of irregular branching glands that are typically surrounded by only a minimal desmoplastic response. In the majority of cases, the tumor is superficially invasive, although deep invasion with extension into the uterine corpus may occur. The pattern of immunostaining is similar to that of usual-type endocervical adenocarcinoma (Jones et al. 2000).
The diagnosis of villoglandular adenocarcinoma may pose a challenge on biopsy. The differential diagnosis of this tumor includes papillary endocervicitis, papillary adenofibromas of the cervix, and müllerian papillomas. All three of these lesions lack the degree of cellular atypia and mitotic activity that is apparent in villoglandular adenocarcinomas. The müllerian papilloma is a lesion of children, whereas the average age of patients with villoglandular adenocarcinomas is in the fourth and fifth decade. In contrast to villoglandular adenocarcinoma, adenofibroma of the cervix and müllerian papilloma have a more prominent stromal component. Villoglandular adenocarcinoma must also be distinguished from usual adenocarcinoma with papillary architecture. Usual endocervical adenocarcinoma demonstrates irregular, thick fibrovascular cores and higher degree of cytologic atypia. Examination of multiple sections typically uncovers the conventional appearance with infiltrative glands or cribriform growth. While the rare serous carcinoma of the cervix is theoretically in the differential diagnosis, some doubt its existence and believe such tumors are more likely secondary to a serous carcinoma of endometrial or tubal/ovarian origin. Immunohistochemical analysis of p53 expression is useful for this rare situation, as villoglandular adenocarcinoma has non-aberrant (heterogeneous) p53 expression similar to usual-type endocervical adenocarcinoma, whereas serous carcinomas have aberrant p53 expression (either diffuse overexpression or complete absence of expression (“null”)).
Prognosis and Treatment
In most of the cases published to date, the clinical outcome of patients with villoglandular adenocarcinomas has been excellent. In the two largest series, all patients, including those who were treated by cone biopsy, were alive and well with no evidence of recurrent disease after 7–77 months of follow-up (Jones et al. 2000). Conservative treatment should be considered only if the tumor is superficial and does not involve lymph-vascular spaces and there is no disease on the cone margins. In a series of cases from Japan, the presence of lymph-vascular space involvement was associated with lymph node metastases (Kaku et al. 1997). Rarely, villoglandular carcinomas may be mixed with other types of carcinomas. The authors are aware of two such cases in which lymph node metastasis has occurred (unpublished observations).
Gastric Adenocarcinoma of the Cervix, Including Minimal Deviation Adenocarcinoma
Gastric adenocarcinoma of the cervix is a recently described diagnostic entity (Kojima et al. 2007) that was included in the updated WHO 2014 tumor classification. The tumor spans a spectrum of morphologic appearances with the common feature of expression of gastric-type mucins. Minimal deviation adenocarcinoma with mucinous morphology was identified to be a part of that spectrum, and this tumor is now considered to represent a well-differentiated form of gastric adenocarcinoma (Kojima et al. 2007). The true incidence of gastric adenocarcinoma is not currently known as this lesion has been previously misclassified as usual, intestinal, or clear cell adenocarcinoma, but it is thought to be rare in Western countries, while in Japan it may account for over 20% of all cervical adenocarcinomas (Kojima et al. 2007). The pathogenesis of gastric adenocarcinoma is still being investigated. Notably, gastric adenocarcinoma, including minimal deviation type, is not related to HPV infection (Park et al. 2011). Some cases were reported to be associated with Peutz-Jeghers syndrome, an autosomal dominant disorder caused by germline mutation of STK11, a serine threonine kinase gene (Gilks et al. 1989). In addition, somatic mutations of STK11 are reported in over a half of the sporadic cases (Kuragaki et al. 2003). Mutation of p53 gene is suspected in proportion of cases based on mutational-type immunostaining pattern of some tumors (Carleton et al. 2016). The putative precursor lesion is lobular endocervical gland hyperplasia (LEGH), a benign proliferation of endocervical glands with gastric phenotype. The average age of the patients is between 42 and 50 years (range 25–84) (Kojima et al. 2007; Karamurzin et al. 2015), which is similar to that of patients with usual adenocarcinoma. Many patients present with profuse watery vaginal discharge or enlarged, barrel-shaped cervix. Cytologic screening has low sensitivity for minimal deviation adenocarcinoma, and HPV testing has no role; therefore many patients present with high-stage tumors. The prognosis for patients with gastric adenocarcinoma, including that of minimal deviation adenocarcinoma, is significantly worse in comparison to usual type (Gilks et al. 1989). The reported overall 5-year disease-specific survival for gastric versus usual adenocarcinoma is 30–42% vs. 74–91%, respectively (Park et al. 2011; Karamurzin et al. 2015). The tumors have a predisposition for metastatic spread to the ovaries, abdominal cavity, and extraperitoneal sites (Karamurzin et al. 2015). In some cases the metastases may deceptively mimic primary ovarian mucinous tumors, including cystadenomas and atypical proliferative/borderline tumors, due to a maturation phenomenon in which the ovarian metastases can be even more differentiated than the primary cervical tumor. Sometimes the primary cervical neoplasm is identified only when staging surgery is performed.
Minimal deviation adenocarcinoma of gastric type is an extremely well-differentiated variant of gastric adenocarcinoma in which the neoplastic epithelium shows a high degree of maturation. These tumors were originally referred to as “adenoma malignum.” Because of their close cytologic resemblance to normal endocervical glands, in 1975 Silverberg and Hurt introduced the term “minimal deviation adenocarcinoma” (MDA) for these lesions (Silverberg and Hurt 1975). Minimal deviation adenocarcinomas are uncommon tumors and account for only 1–3% of all cervical adenocarcinomas. The characteristic microscopic features of MDA include low-grade cytology but markedly atypical architecture (Figs. 39 and 40). The glands are lined by epithelium that usually has minimal, if any, nuclear atypia, but the nuclei are larger in comparison to adjacent benign glands. The nuclei are rounded, slightly vesicular with conspicuous nucleoli. Desmoplasia may be present surrounding the angular outpouchings or in the deep portion of the tumor. However, in some cases, large areas of invasive tumor may be devoid of any stromal reaction. The presence of glands adjacent to thick-walled blood vessels is helpful in determining that stromal invasion is present. The most reliable criterion is the haphazard arrangement of glands that extend beyond the level of normal endocervical glands. Minimal deviation adenocarcinoma often involves more than two-thirds of the thickness of the cervical stroma, while the normal endocervical crypts and tunnels do not extend as deep. Because the depth of penetration of the glands is an essential histologic feature of minimal deviation adenocarcinoma, in most cases, the diagnosis cannot be made on a superficial cervical biopsy, but instead requires either a cone biopsy or hysterectomy specimen.
The cytoplasm of gastric adenocarcinoma characteristically shows immunopositivity for gastric mucin markers MUC-6 and HIK-1083 (Carleton et al. 2016; Mikami et al. 2004), however, in some cases such staining may be only focal. The staining has to be interpreted in the proper context because benign endocervical glands may be positive for HIK-1083 and MUC-6 in 2% and 8% of cases (Mikami et al. 2004). Similar to usual-type adenocarcinoma, gastric adenocarcinoma is negative for ER and PR (Carleton et al. 2016) and demonstrates cytoplasmic positivity for CEA (Carleton et al. 2016). The staining for p16 is typically negative (Fig. 45b), correlating with the negative HPV status; however, rare cases may show strong, blocklike p16 positivity (Carleton et al. 2016). Less than a half of cases show mutation-type p53 staining with either diffuse strong positivity (Fig. 45c), or a complete lack of p53 expression (null pattern) (Carleton et al. 2016). In addition, gastric adenocarcinomas are reported to express CK7, PAX-8, CA19.9, and hepatocyte nuclear factor-1 beta (Carleton et al. 2016).
Gastric minimal deviation adenocarcinoma has to be differentiated from benign conditions, of which the most important are LEGH, deep nabothian cysts, endocervical tunnel clusters, endocervicosis of the cervical wall, diffuse laminar endocervical glandular hyperplasia and adenomyomas of endocervical type. Lack of ER/PR staining distinguishes this tumor from all of the benign conditions except for LEGH (Carleton et al. 2016). Similar to gastric adenocarcinoma, LEGH shows immunopositivity for gastrin mucin markers and is negative for hormone receptors (Mikami et al. 2004). Differentiation of minimal deviation adenocarcinoma from non-invasive LEGH may be extremely difficult. A haphazard gland arrangement, extension deep into the cervical stroma and a desmoplastic response are all in favor of minimal deviation adenocarcinoma. Desmoplasia can be better visualized with smooth muscle actin (SMA) immunostaining of the stromal cells. Positive SMA staining has been described in cervical stromal cells immediately adjacent to foci of tumor invasion, but not around LEGH (Mikami et al. 2005). In addition, minimal deviation adenocarcinoma was reported to be negative for PAX-2, with approximately 40% showing mutation type staining for p53, whereas LEGH exhibited positive PAX-2 staining and non-aberrant (“wild-type”) p53 expression (Carleton et al. 2016; Mikami et al. 2009; Rabban et al. 2010). Among the malignant lesions, gastric adenocarcinoma may be difficult to distinguish from usual-type and clear cell adenocarcinomas. Expression of MUC-6 (HIK-1083 is not routinely available outside of Japan) and a negative p16 stain favor gastric differentiation, whereas usual-type adenocarcinoma has diffuse p16 expression and lacks MUC-6 expression. In contrast to gastric adenocarcinoma, clear cell carcinomas typically show variable positivity for p16 and lack MUC-6 expression (Park et al. 2011). Interestingly, both clear cell carcinomas and gastric adenocarcinomas demonstrate cytoplasmic positivity for hepatocyte nuclear factor-1 beta (Carleton et al. 2016) and napsin A (unpublished observation).
Gastric-type cervical adenocarcinomas show a predisposition for metastatic spread to the ovaries, abdominal cavity, and extraperitoneal sites, and the initial presentation may be with metastases which must be differentiated from metastatic gastrointestinal and pancreatobilliary tract adenocarcinomas. The most useful immunostain in such cases is PAX-8, as 68% of primary cervical gastric adenocarcinomas were shown to express this marker (Carleton et al. 2016).
Signet Ring Cell-Type Adenocarcinoma of the Cervix
Primary signet ring cell tumors of the cervix in the pure form are exceedingly rare and more often are admixed with intestinal, gastric, or endocervical-type mucinous adenocarcinomas. A proportion of signet ring cell adenocarcinomas found in the cervix are metastatic lesions, most commonly from a primary gastric carcinoma (Imachi et al. 1993). Only few case reports of primary cervical tumors are available (Sal et al. 2016). Tumors arising in a background of usual-type adenocarcinoma were shown to be HPV-positive (Sal et al. 2016).
The tumor shows characteristic infiltration by clusters or single cells distended by a mucin vacuole. Of the reported cases, the tumors were positive for p16, CK7, and CEA (Sal et al. 2016). A metastasis from primary gastric, breast, colonic, or appendiceal carcinoma has to be ruled out using immunohistochemical analysis in correlation with clinical findings and imaging studies.
Clear Cell Adenocarcinoma of the Cervix
Clear cell adenocarcinomas account for 2–7% of cervical adenocarcinomas and comprise a heterogeneous group of malignancies. Clear cell adenocarcinoma has a bimodal age distribution. The first peak occurs between 17 and 37 years of age (mean age 26 years); in addition, rare cases of clear cell adenocarcinoma in children have been reported (Hanselaar et al. 1997; Liebrich et al. 2009). The second peak occurs in women who are 44–88 years of age (mean age 71 years). The pathogenesis of clear cell adenocarcinomas of the cervix is not well understood. With rare exceptions, clear cell adenocarcinomas are negative for HPV DNA (Park et al. 2011; Holl et al. 2015; Ueno et al. 2013). In the past, cases occurring in younger patients were linked to DES exposure in utero (Robboy et al. 1984). In these patients the tumors were developing on the ectocervix, rather than in the endocervical canal. More recent data on patients born outside the DES exposure period still show a bimodal age distribution with some cases still occurring in young, virginal women without exposure to either DES or HPV (Liebrich et al. 2009). It is thought that clear cell adenocarcinomas may develop from either adenosis of the ectocervix, cervical endometriosis, or cervical tubo-endometrioid metaplasia (Hiromura et al. 2009). Immunostaining analysis of tumors from a group of older patients showed increased expression of EGFR (75% cases) and HER2 (25% cases) and loss of PTEN expression (50% cases). In addition, 58% of cases demonstrated expression of p-AKT, and 50% of cases had expression of p-mTOR, suggesting involvement of the PI3K-AKT pathway (Ueno et al. 2013). The prognosis of surgically treated patients with stage IB–IIB clear cell carcinomas without exposure to DES is similar to patients with non-clear cell adenocarcinomas (Thomas et al. 2008).
Cervical clear cell carcinomas are positive for hepatocyte nuclear factor-1 beta and napsin A similar to their ovarian and uterine counterparts (Park et al. 2011). The staining for both of these markers however is not entirely specific, as a proportion of gastric adenocarcinomas also shows positivity with both stains. Despite the negative HPV status, the tumors may show positivity for p16 with a spectrum from focal to diffuse. Clear cell carcinomas are negative for CEA, ER, and PR, and the majority show non-aberrant (“wild-type”) p53 expression (Park et al. 2011; Ueno et al. 2013).
A metastasis from clear cell adenocarcinoma of the endometrium or ovary has to be excluded based on tumor location on imaging studies or histologic findings. In addition, clear cell adenocarcinoma has to be differentiated from the HPV-related clear cell variant of squamous cell carcinoma and clear cell adenosquamous carcinoma of the cervix, as well as HPV-negative gastric and mesonephric adenocarcinomas. The clear cell variant of squamous cell carcinoma shows a diffuse solid growth pattern, whereas clear cell adenocarcinoma displays a spectrum of architectural patterns within the same lesion. Clear cell adenosquamous carcinoma is composed of solid sheets of clear cells in addition to glandular spaces lined by columnar cells which are positive for mucicarmine (Fujiwara et al. 1995). Clear cell adenocarcinoma with tall, columnar cells may be difficult to distinguish from gastric adenocarcinoma on routine sections, and immunostains for p16, hepatocyte nuclear factor-1 beta, and napsin A may not be helpful due to some overlap in expression patterns; however, CEA was reported to be negative in clear cell carcinomas while showing variable expression in gastric tumors (Park et al. 2011). Clear cell adenocarcinoma with a tubulocystic pattern and flattened epithelium must be differentiated from mesonephric adenocarcinomas, which typically express GATA-3 staining.
In a small biopsy or curettage sample, the tumor may be difficult to distinguish from benign conditions such as florid or solid microglandular hyperplasia, florid Arias-Stella reaction, atypical oxyphilic metaplasia, and radiation-induced atypia. In all these conditions, lack of significant proliferative activity and expression of hormone receptors favors the benign processes. Clinical history of use of progestagenic hormones, current or recent pregnancy, or history of radiation can help to support the respective benign diagnoses in the differential diagnosis.
Serous Adenocarcinoma of the Cervix
Primary cervical serous adenocarcinoma is exceedingly rare. The majority of serous tumors involving the cervix are malignancies spreading directly from the endometrium or metastases from a primary ovarian, tubal, or peritoneal serous adenocarcinoma. Patients with primary cervical serous cancer range in age from 27 to 79 years (mean 52 years) (Zhou et al. 1998; Nofech-Mozes et al. 2006). Of the few cases reported in the literature, only rare tumors were positive for HPV DNA (Park et al. 2011). It is speculated that cervical serous adenocarcinoma may be caused by p53 gene mutations, similar to its endometrial counterpart (Nofech-Mozes et al. 2006). In the largest reported series, this histologic variant has been associated with a poor prognosis when diagnosed at an advanced stage, but the outcome for patients with stage I tumors is similar to that of patients with cervical adenocarcinomas of the usual type (Zhou et al. 1998).
Serous adenocarcinoma of the cervix is histologically identical to serous adenocarcinomas arising in the endometrium. The diagnosis of primary serous cervical adenocarcinoma should be made only if spread from other gynecologic sites has been excluded. These tumors are composed of papillary tufts and complex papillae lined by cells with pleomorphic, hobnailed, high-grade nuclei. Alternatively, they infiltrate the stroma with glands with slit-like lumina and lined by markedly atypical cells, some with “smudgy” nuclei. High mitotic activity may be seen. In the reported series of serous adenocarcinomas, almost half of the cases were mixed with another histologic type, most commonly with low-grade villoglandular adenocarcinoma (Zhou et al. 1998). The tumors are positive for p16 while negative for ER and PR. Aberrant (“mutation-type”) p53 expression was seen in roughly half of the reported cases (Nofech-Mozes et al. 2006).
Mesonephric Carcinoma of the Cervix
Mesonephric adenocarcinoma is a rare cervical tumor which develops from the mesonephric duct remnants located deep in the lateral cervical stroma. Mesonephric duct remnants are detected in up to 20% of cervices removed during routine hysterectomy, and adenocarcinomas can rarely develop in these remnants. Patients range in age from 34 to 72 years without apparent peak age, and most present with abnormal vaginal bleeding (Clement et al. 1995; Silver et al. 2001). In a recent study of targeted next-generation sequencing, 81% of mesonephric adenocarcinomas had either a KRAS (n = 12) or NRAS (n = 1) mutation. Mutations in chromatin remodeling genes (ARID1A, ARID1B, or SMARCA4) were present in 62% of cases. No mutations of PIK3CA or PTEN genes were identified. In addition, 1q gain was found in 75% of cases (Mirkovic et al. 2015). HPV DNA is not detected in this tumor type (Kenny et al. 2012). Stage I mesonephric carcinomas seem to have a more indolent behavior than other types of cervical adenocarcinoma (Clement et al. 1995). However, several high-stage tumors have had an aggressive course, and several tumors with a sarcomatoid component have metastasized (Silver et al. 2001).
Mesonephric carcinomas are very rare and in the past were confused with clear cell carcinomas of the cervix. In contrast to the superficial location of cervical clear cell adenocarcinomas, true mesonephric adenocarcinomas develop deep in the lateral wall of the cervix, in a site corresponding to the location of mesonephric duct remnants. Therefore, they often extend into the outer third of the cervical wall.
Cytologically, the tumors are composed of relatively uniform columnar or cuboidal cells with scant to moderate amount of dark eosinophilic cytoplasm. The nuclei are oval, hyperchromatic with mostly stippled chromatin showing minimal to moderate atypia. Marked nuclear atypia is not seen. The mitotic index is highly variable and may range from 1 to 50 mitoses per 10 HPFs (Silver et al. 2001). Mesonephric hyperplasia of the lobular or diffuse pattern is seen adjacent to the invasive lesions in the majority of cases.
The immunohistochemical profile of mesonephric carcinomas is similar to that of normal mesonephric duct remnants. Epithelial markers including CK7 and EMA are positive. The tumors are characterized by expression of PAX-8 (diffuse/strong), GATA-3 (variable extent and intensity), calretinin (may be focal), and CD10 (in the apical portion, luminal edge of the cells). Tumors typically have patchy p16 expression and are negative for ER, PR, and CK20 (Howitt et al. 2015; Silver et al. 2001; Kenny et al. 2012).
Mesonephric adenocarcinoma has a wide differential diagnosis. On the benign spectrum, it has to be differentiated from hyperplasia of mesonephric remnants. The presence of cytologic atypia and variable architectural patterns favors adenocarcinoma. In addition, assessment of proliferative activity with a Ki-67 immunostain may be helpful, as hyperplasia was reported to show positivity in only 1–2% of cells, as compared to 5–36% of cells in adenocarcinoma (Silver et al. 2001). Mesonephric carcinomas with a prominent spindle cell component have to be distinguished from cervical carcinosarcoma. The carcinomatous component in the latter entity is often squamous or basaloid in contrast to tubules and glands present in mesonephric carcinomas with a prominent spindle cell component. The ductal variant of the tumor has to be differentiated from endometrioid adenocarcinoma. The presence of other distinct architectural patterns, intraluminal eosinophilic secretions, adjacent benign mesonephric remnants, and the pattern of ER-/PR-/calretinin+/GATA-3+ staining confirm the diagnosis of mesonephric adenocarcinoma. The tubular pattern of mesonephric adenocarcinoma may simulate clear cell carcinoma, while the papillary pattern may mimic either clear cell or serous adenocarcinoma; however, in contrast to these two latter tumors, mesonephric adenocarcinoma typically shows lower nuclear grade, adjacent mesonephric remnants, and expression of GATA-3 (Howitt et al. 2015).
Other Epithelial Tumors
A rare tumor variant of adenosquamous carcinoma in which at least 70% of tumor cells have vacuolated, clear cytoplasm containing large amounts of glycogen has been referred to as clear cell adenosquamous carcinoma (Fujiwara et al. 1995). The cohesive sheets of tumor cells are frequently subdivided by connective tissue septa, which can have a prominent lymphocytic infiltrate that produces a lobulated appearance. The tumors demonstrate focal gland formation and stain positively with a mucin stain such as mucicarmine. In some clear cell adenosquamous carcinomas, there are spindle-shaped cells suggesting squamous differentiation. Clear cell adenosquamous carcinomas need to be distinguished from clear cell carcinomas and glassy cell carcinoma of the cervix. Unlike clear cell carcinomas, clear cell adenosquamous carcinomas lack papillary or tubulocystic areas and hobnail cells. Clear cell adenosquamous carcinoma is associated with HPV 18 and has an aggressive clinical course. In one series, 7 of the 11 patients died of their disease, including 3 of 5 patients with stage IB disease (Fujiwara et al. 1995).
Glassy Cell Carcinoma
Glassy cell carcinomas are most often positive for HPV 18 (Pirog et al. 2000), and the average age at diagnosis is 45 years (Guitarte et al. 2014). The tumor has an aggressive clinical course (median survival 25 months), with a poor response to radiation and surgery. More recently patients treated with neoadjuvant chemotherapy have been reported to have longer survival (Guitarte et al. 2014).
Adenoid Basal Tumors, Including Epithelioma and Carcinoma
The more generic term adenoid basal tumor is recommended for tumors with adenoid basal differentiation encountered in excisional specimens with positive margins, with more specific designation as epithelioma versus carcinoma deferred to evaluation of a completely excised tumor (Russell and Fadare 2006). Final designation as adenoid basal epithelioma is reserved for completely excised pure low-grade tumors. Pure adenoid basal tumors of the low-grade epithelioma type appear to have a benign clinical course (Brainard and Hart 1998; Russell and Fadare 2006). Carcinomas with adenoid basal differentiation, which are typically mixed with other carcinomatous components, are expected to behave in a similar fashion as tumors with those other types of differentiation (squamous, adenoid cystic) of similar stage.
Adenoid Cystic Carcinoma
Adenoid cystic carcinoma of the cervix is a very rare cervical tumor. It is most often seen in patients between the sixth and seventh decades (mean age of 71 years in the largest series) and is more common in African-American women (Grayson et al. 1999). Adenoid cystic carcinomas mixed with usual squamous cell carcinoma or small cell carcinoma are etiologically related to high-risk HPV and are characterized by diffuse p16 expression. Pure cervical adenoid cystic carcinomas are exceedingly rare and appear to be unrelated to high-risk HPV and are characterized by nondiffuse p16 expression (Xing et al. 2016). Origin from reserve cells has been proposed for adenoid cystic and adenoid basal tumors (Grayson et al. 1999).
Adenoid cystic carcinoma behaves aggressively with frequent local recurrences or metastatic spread. In a review of 43 cases, overall survival for patients with stage I disease was 56% at 3–5 years (Prempree et al. 1980). While some studies have reported a higher survival rate, it appears to be considerably lower than other types of stage I cervical cancer. Adenoid cystic carcinoma of the cervix should be differentiated from adenoid basal carcinoma of the cervix.
Mucoepidermoid carcinoma (MEC) is not recognized as separate entity in the recent WHO classification of cervical tumors but is included within the category of adenosquamous carcinoma. Morphologic criteria for diagnosing cervical MEC are identical to those of salivary gland MEC and include the presence of three cell types – epidermoid, intermediate, and mucin-producing – and lack of recognizable glands. In mucoepidermoid carcinoma, the squamous component consists of large cell nonkeratinizing or focally keratinizing areas and the mucin-producing cells that are frequently localized in the center of the squamous nests. The mucinous component includes goblet or signet ring-type cells that contain mucicarminophilic, periodic acid-Schiff (PAS)-positive, diastase-resistant mucopolysaccharides. The mucin from these cells is extruded into the intercellular spaces or fibrous stroma, where it may collect in small or large lakes. The tumor is distinguished from the conventional adenosquamous carcinoma by the lack of gland formation.
Rearrangements of CRTC1 and MAML2 genes have been described in cervical MEC (Lennerz et al. 2009). Prior to that, translocation involving these two genes was reported in minor salivary gland MEC. None of the analyzed cervical adenosquamous carcinomas harbored rearrangements of either locus suggesting that cervical MEC is distinct from conventional cervical adenosquamous carcinoma.
Neuroendocrine tumors of the cervix are relatively rare. The terminology that has been used to describe these tumors has varied over the last 25 years resulting in difficulties in interpreting clinicopathologic studies. Because the morphologic features of cervical neuroendocrine tumors are similar to those in the lung, the current WHO classification for cervical tumors is the same as the one used for the lung. There are two main tumor categories: low-grade neuroendocrine tumors including carcinoid tumor and atypical carcinoid tumor and high-grade neuroendocrine tumors including large cell neuroendocrine carcinoma and small cell neuroendocrine carcinoma. High-grade neuroendocrine tumors are associated with high-oncogenic-risk HPVs (Wang and Lu 2004; Grayson et al. 2002) .
The exact cellular origin of neuroendocrine tumors of the cervix is unknown. Small numbers of argyrophil cells, a potential precursor for neuroendocrine tumors, are present in the exocervical epithelium as well as endocervical epithelium.
Typical and Atypical Carcinoid Tumor
Histologic features used for distinguishing neuroendocrine tumors of the cervix
Typical carcinoid tumor
Trabecular, insular, sheetlike
Atypical carcinoid tumor
Trabecular, insular, sheetlike
Large cell neuroendocrine carcinoma
Sheets, organoid trabecular, cord-like
Small cell carcinoma
Sheets, nests, trabecular, cord-like
Typical and atypical carcinoid tumors are uncommon. Some of the tumors that have been previously described as well-differentiated cervical carcinoid tumors appear to be adenocarcinomas of the cervix that focally resemble carcinoid tumors of the intestinal tract and have neuroendocrine differentiation. It should be noted that neurosecretory granules can be identified in many carcinomas of the cervix if a diligent search is made. Although early reports suggested that these tumors had a relatively good prognosis, more recent reports demonstrate that these tumors can act in a malignant fashion with local and distant metastasis (Gardner et al. 2011). To date, none of the published cases have been associated with the carcinoid syndrome.
Large Cell Neuroendocrine Carcinoma
Atypical carcinoids and large cell neuroendocrine carcinomas can be distinguished based on mitotic activity, nuclear atypia, and degree of necrosis (Table 8) (Albores-Saavedra et al. 1997). It is more difficult to differentiate between large cell neuroendocrine carcinoma and poorly differentiated cervical adenocarcinomas or squamous cell carcinomas. It is important that trabecular and insular growth patterns be looked for in poorly differentiated cervical tumors and that stains for neuroendocrine markers be used whenever there is an indication of neuroendocrine differentiation. It should be cautioned, however, that occasional typical cervical adenocarcinomas and adenosquamous carcinomas can stain focally with neuroendocrine markers or contain occasional argyrophilic cells. In contrast, large cell neuroendocrine carcinomas have evidence of neuroendocrine differentiation by routine light microscopy and show more diffuse expression of neuroendocrine markers. Finally, large cell neuroendocrine carcinoma has to be distinguished from malignant melanoma. The presence of melanin pigment and immunoreactivity for S100, HMB-45, and Sox10 facilitates the distinction between these tumors.
Several studies have explored the role of HPV in large cell neuroendocrine carcinomas. In the largest series, HPV 16 was detected by in situ hybridization or polymerase chain reaction in 58% of cases, and HPV 18 was detected in 16% of cases (Grayson et al. 2002). Large cell neuroendocrine carcinomas are highly aggressive neoplasms. In the review of 31 published cases, 65% of patients died of disease within 3 years of diagnosis (Grayson et al. 2002).
Small Cell Carcinoma
Small cell carcinomas of the cervix are histologically identical to their counterparts at other sites such as the lung. In most series these tumors account for 1–2% (range 0.5–5%) of all cervical tumors (Abeler et al. 1994). The age of the patients ranges from the second to ninth decade, with mean and median age in the fifth decade (Abeler et al. 1994). Most patients present with abnormal vaginal bleeding and have an obvious mass on pelvic examination. In rare cases, patients present with abdominal symptoms related to ovarian metastases. The number of patients presenting with abnormal cytologic examination is smaller than in patients with squamous cell carcinoma. This results from lack of an in situ component and rapid growth of the tumor (Ambros et al. 1991).
Immunohistochemical and Molecular Genetic Findings
Neuroendocrine dense-core granules can be detected using Grimelius stains or electron microscopy in most cases. Although small cell carcinomas have been associated with ectopic ACTH, insulin, and gastrin production, clinical symptoms related to ectopic hormone production are uncommon. By immunohistochemistry, neuroendocrine markers such as chromogranin or synaptophysin are present in many cases; however, staining might be very focal. Small cell carcinomas show variable expression of cytokeratins, epithelial membrane antigen, as well as a variety of hormones and polypeptides including ACTH, calcitonin, serotonin, gastrin, substance P, VIP, and somatostatin (Ueda et al. 1989). In addition, the tumors show diffuse p16 expression and TTF-1 positivity (McCluggage et al. 2010). Virtually all cervical small cell carcinomas have been associated with high-risk HPV types 18 and 16, with type 18 being the most prevalent, detected in 82–100% of the cases (Wang and Lu 2004).
Differentiation between small cell carcinoma and nonkeratinizing squamous carcinoma with small cells can be difficult. The diagnosis of small cell carcinoma should be reserved for tumors composed of small cells in which squamous or glandular differentiation is absent or minor. Histologically, cells of nonkeratinizing squamous carcinoma with small cells resemble those of high-grade SIL and lack the nuclear molding and extensive crush artifact present in most small cell carcinomas. Small cell carcinomas invade the stroma diffusely in trabeculae and poorly defined nests. In contrast, nonkeratinizing squamous carcinomas with small cells invade the stroma in discrete nests. In an individual case, immunohistochemistry for neuroendocrine markers may not be helpful because 40% of nonkeratinizing squamous carcinomas with small cells are positive for neuroendocrine markers and 40% of small cell carcinomas are positive for cytokeratins (Ambros et al. 1991). Nuclear staining for p63 confirms squamous differentiation in nonkeratinizing small cell carcinomas, while neuroendocrine-type small cell carcinomas are negative for this marker (Wang et al. 2001). Due to the presence of high-risk HPV in both tumor types, p16 expression is similar. Expression of thyroid transcription factor 1 (TTF-1) has been reported in cervical small cell carcinomas and can cause diagnostic problems in tumors that metastasize to the lungs. Immunohistochemical staining with antibodies against leukocyte common antigen and neuroendocrine markers can be useful for differentiating small cell carcinoma from lymphoproliferative disorders.
Clinical Behavior and Treatment
Small cell carcinoma of the cervix is a highly aggressive tumor. Lymph-vascular invasion is present in 90% of cases and is often extensive (Abeler et al. 1994). The prognosis of small cell carcinoma of the cervix is worse than that of stage-comparable, poorly differentiated squamous carcinoma (Ambros et al. 1991; Zivanovic et al. 2009). Five-year survival rate is reported at 14% (Abeler et al. 1994). Combined modality treatments with chemotherapy and radiotherapy are currently used in management of small cell carcinoma. In a recent series, patients with early-stage disease treated with platinum-based chemotherapy in addition to radiation had significantly better overall and disease-free survival when compared to patients who did not receive chemotherapy as part of their initial treatment (Zivanovic et al. 2009).
Mesenchymal and Mixed Epithelial-Mesenchymal Tumors
Primary cervical mixed epithelial and mesenchymal tumors include malignant müllerian mixed tumor (MMMT) and müllerian adenosarcoma. Cervical MMMT are less common than their much more common uterine counterparts, and in contrast to uterine MMMT, cervical MMMTs are related to infection with high-risk HPV (Grayson et al. 2001). Both cervical and uterine tumors usually occur in postmenopausal women, and both typically form polypoid or pedunculated masses. The mean age of patients was 65 years in the largest published series of cervical tumors (Clement et al. 1998). Histologically, cervical MMMTs differ in their appearance from MMMTs arising in the uterus. The carcinomatous component in cervical MMMT is often a basaloid tumor composed of anastomosing densely cellular trabeculae of small cells with scant cytoplasm and peripheral palisading. Other epithelial patterns include typical squamous cell carcinoma and endometrioid adenocarcinoma. Adenoid basal and adenoid cystic components have also been reported in several cases (Mathoulin-Portier et al. 1998). The sarcomatous element is typically homologous and frequently has the appearance of a fibrosarcoma or endometrial stromal sarcoma. The sarcomatous element is frequently high-grade and may have myxoid change.
Extension of uterine MMMT to the cervix is in the differential diagnosis of cervical MMMT. The correct diagnosis is based on the dominant location of the tumor, the appearance of the carcinomatous component, and the detection of high-risk HPV. In a study of eight patients with cervical MMMT, HPV DNA was detected by polymerase chain reaction in all cases. Using in situ hybridization, HPV 16 DNA was detected in both the epithelial and sarcomatous components in three cases (Grayson et al. 2001). Although the number of reported cases is small, cervical MMMTs may have a better prognosis than their uterine counterparts (Clement et al. 1998).
Only a small number of müllerian adenosarcomas of the cervix have been reported (Jones and Lefkowitz 1995). Adenosarcomas occur in women between the ages of 14 and 67 years, with a mean age of 38 years (Jones and Lefkowitz 1995). Women typically present with vaginal bleeding or recurrent cervical polyps. Microscopically, tumors usually demonstrate thick papillae covered with a typical endocervical-type epithelium. The appearance of the sarcomatous component can vary considerably. In some tumors, it consists of mitotically active, plump spindle cells that form periglandular cuffs and a cambium layer under the surface epithelium. At least two mitotic figures per 10 high-power fields are required to make a diagnosis of adenosarcoma, but in most cases the mitotic index exceeds 4 per 10 high-power fields. In other tumors, the stromal component contains foci that are more embryonic in appearance, with small, undifferentiated round cells that are mitotically active. Stains for desmin and myogenin as well as Ki-67 are useful to identify such foci as rhabdomyoblastic differentiation which can occur in adenosarcomas of both endometrial and cervical origin. Heterologous sarcomatous elements, including strap cells (skeletal muscle differentiation), lipoblasts, cartilage, and osteoid, can be present.
The differential diagnosis includes adenofibroma, atypical endocervical polyp, adenomyoma of the cervix, and MMMT. Adenofibroma is also a biphasic lesion, but both the epithelium and stroma are benign. Atypical endocervical polyps show increased stromal cellularity and reactive nuclear atypia, but these changes are often focal and mitotic activity is absent. Adenomyomas can be distinguished from adenosarcoma by the presence of a well-defined bland myomatous component. The prognosis of cervical adenosarcoma is usually good after surgical therapy, although several patients have died of disease or developed recurrent tumor (Jones and Lefkowitz 1995).
Primary malignant melanoma is among the least common of the malignant tumors that arise in the cervix (Clark et al. 1999). Common presenting signs include vaginal bleeding, frequently of short duration. In most instances, the lesion is pigmented and dark brown. The diagnosis of primary melanoma of the cervix is based on the histologic demonstration of junctional changes in the squamous epithelium and the absence of similar lesions elsewhere in the body. Morphologically, it is identical to melanoma arising in the skin and extragenital mucous membranes; it frequently contains intracytoplasmic melanin pigment granules. Some tumors are amelanotic and have to be distinguished from undifferentiated carcinoma. Rarely cervical malignant melanoma can be composed of clear cells. Immunohistochemical staining for HMB 45, Melan-A (MART-1), S100 protein, and Sox10 and absence of epithelial markers are helpful to exclude clear cell carcinoma. Spindle cell malignant melanoma has to be distinguished from leiomyosarcoma or malignant peripheral nerve sheet tumor. In contrast to malignant melanoma, leiomyosarcoma expresses smooth muscle markers and is negative for melanocytic markers. The cell pigmentation, nesting, and the presence of an atypical epidermal or junctional component, together with diffuse, strong reactivity for S100 protein and positivity for other melanocytic markers, help to differentiate melanoma from malignant peripheral nerve sheet tumor. The prognosis of primary malignant melanoma is poor, with only 25% survival rate for patients with stage I disease (Clark et al. 1999).
Primary choriocarcinoma and epithelioid trophoblastic tumors in the cervix are rare. The gross and microscopic appearance, as well as the clinical course, is identical with those found in the uterine corpus. Primary cervical germ cell tumors have been described: these include both the mature teratomas and yolk sac tumors. There are also case reports of primitive neuroectodermal tumors (PNET) of the cervix (Snijders-Keilholz et al. 2005). These tumors appear to be identical to PNETs occurring at other sites and in some cases have expressed the restricted surface antigen MIC-2, showed positive staining for CD99, and contained the EWS/FLI-1 chimeric mRNA transcript characteristic of PNET/Ewing sarcoma family (Masoura et al. 2012).
Direct extension from a pelvic tumor is the most common source of cervical involvement by secondary carcinoma, often originating in the endometrium, rectum, or bladder. Intrapelvic and intragenital, lymphatic, or vascular metastases to the cervix occur less frequently. These lesions are usually associated with ovarian carcinoma and endometrial adenocarcinoma and less commonly with transitional cell carcinoma of the bladder. Another lesion that has a relatively high rate of cervical metastasis is choriocarcinoma. Sarcomas of the uterine corpus may also involve the cervix. Metastases to the cervix from distant primary sites are rare, the most common being the gastrointestinal tract (the colon and stomach), ovary, and breast (Perez-Montiel et al. 2012). Instances of metastatic carcinoma from the kidney, gallbladder, pancreas, lung, thyroid, and malignant melanoma have also been described. On occasion, metastases may occur primarily as cervical involvement and pose a differential diagnostic problem. Unusual gross appearance or histologic patterns, e.g., signet ring cell carcinoma or clear cell carcinoma, provide a clue to the possibility of origin from a distant primary site.
- Appleby P et al (2006) Carcinoma of the cervix and tobacco smoking: collaborative reanalysis of individual data on 13,541 women with carcinoma of the cervix and 23,017 women without carcinoma of the cervix from 23 epidemiological studies. Int J Cancer 118(6):1481–1495PubMedCrossRefPubMedCentralGoogle Scholar
- Berek JS, Hacker NF (2010) Berek & Hacker’s gynecologic oncology. Wolters Kluwer/Lippincott Williams & Wilkins Health, Philadelphia, p 1 online resource (xv, 895 pages)Google Scholar
- Darragh TM et al (2012) The Lower Anogenital Squamous Terminology Standardization Project for HPV-Associated Lesions: background and consensus recommendations from the College of American Pathologists and the American Society for Colposcopy and Cervical Pathology. J Low Genit Tract Dis 16(3):205–242PubMedCrossRefGoogle Scholar
- Eskander RN, Bristow RE, SpringerLink (Online Service) (2015) Gynecologic oncology a pocketbook. Springer, New YorkGoogle Scholar
- Institute, C.R. 2018Google Scholar
- International Collaboration of Epidemiological Studies of Cervical Cancer (2009) Cervical carcinoma and sexual behavior: collaborative reanalysis of individual data on 15,461 women with cervical carcinoma and 29,164 women without cervical carcinoma from 21 epidemiological studies. Cancer Epidemiol Biomarkers Prev 18(4):1060–1069CrossRefGoogle Scholar
- Kenny SL et al (2012) Mesonephric adenocarcinomas of the uterine cervix and corpus: HPV-negative neoplasms that are commonly PAX8, CA125, and HMGA2 positive and that may be immunoreactive with TTF1 and hepatocyte nuclear factor 1-beta. Am J Surg Pathol 36(6):799–807PubMedCrossRefPubMedCentralGoogle Scholar
- Kurman RJ, International Agency for Research on Cancer, World Health Organization (2014) WHO classification of tumours of female reproductive organs, 4th edn. World Health Organization classification of tumours. International Agency for Research on Cancer, Lyon, 307 pGoogle Scholar
- Mikami Y et al (2005) Immunophenotypic alteration of the stromal component in minimal deviation adenocarcinoma (‘adenoma malignum’) and endocervical glandular hyperplasia: a study using oestrogen receptor and alpha-smooth muscle actin double immunostaining. Histopathology 46(2):130–136PubMedCrossRefGoogle Scholar
- NCI 2017Google Scholar
- Tambouret R, Clement PB, Young RH (2003) Endometrial endometrioid adenocarcinoma with a deceptive pattern of spread to the uterine cervix: a manifestation of stage IIb endometrial carcinoma liable to be misinterpreted as an independent carcinoma or a benign lesion. Am J Surg Pathol 27(8):1080–1088PubMedCrossRefPubMedCentralGoogle Scholar
- Yemelyanova A et al (2009) Endocervical adenocarcinomas with prominent endometrial or endomyometrial involvement simulating primary endometrial carcinomas: utility of HPV DNA detection and immunohistochemical expression of p16 and hormone receptors to confirm the cervical origin of the corpus tumor. Am J Surg Pathol 33(6):914–924PubMedCrossRefGoogle Scholar