Malignant diseases of the vagina are either primary vaginal cancers or metastatic from adjacent or distant organs. Primary vaginal cancers are defined as arising solely from the vagina with no involvement of the external cervical os superiorly or the vulva inferiorly. The importance of this definition lies in the different clinical approach in the treatment of upper and lower vaginal cancer. According to the International Federation of Gynecology and Obstetrics (FIGO), a vaginal lesion involving the external os of the cervix should be considered cervical cancer and treated as such; a tumor involving both vulva and vagina should be considered vulvar cancer. A patient with history of a preinvasive lesion or an invasive carcinoma arising from the cervix or the vulva requires that 5-10 years of disease-free interval have past before diagnosing a new vaginal lesion as primary vaginal carcinoma. This criterion is required to rule out recurrent cervical or vulvar disease.
About 80% of vaginal cancers are metastatic, primarily from the cervix or endometrium. Metastatic cancer from the vulva, ovaries, choriocarcinoma, rectosigmoid, and bladder are less common. These cancers usually invade the vagina directly. Cancers from distant sites that metastasize to the vagina through the blood or lymphatic system also occur, including colon cancer, renal cell carcinoma, melanoma, and breast cancer.
History of the Procedure
In 1946, Alexander Brunschwig published the first cases of pelvic exenteration. In his first series, 5 of 22 surgical patients died from the operation itself. The original procedure consisted of connecting the ureters to the colostomy. In 1950, Bricker modified the procedure by isolating a loop of ileum, closing one end, anastomosing the ureters to it, and bringing the patent end out as a stoma.1 Since then, several other modifications have improved the outcome of this procedure. Today, with vaginal reconstruction and continent vesicostomy, the procedure is widely accepted as a treatment in selective cases.
Frequency
Primary vaginal carcinoma is rare, constituting only 1-2% of all malignant gynecological tumors. It ranks fifth in frequency behind cancer of the uterus, cervix, ovary, and vulva. The age-adjusted incidence in the United States is 0.6 per 100,000 population. The strict criteria used in defining vaginal carcinoma contribute to this low incidence.
Etiology
The etiology of vaginal cancer has not been identified. Note that vaginal cancer is not histologically homogeneous; several types of lesions exist, each with its own characteristics, age predilection, aggressiveness, and prognosis (see the Table). This suggests that a single etiologic factor is unlikely. Although some histologic types of vaginal cancer have been associated with exposure to certain agents, so far no clear cause-and-effect relationship exists between any of those agents and vaginal carcinoma.
The strongest association is between squamous cell carcinoma and human papilloma virus (HPV) infection, which is similar to cervical carcinoma. HPV subtypes 16 and 18 have the highest oncogenic potential and are most commonly linked to dysplastic changes in the female genital tract. Because HPV is sexually transmitted, this association raises the question as to whether women who engage in high-risk sexual behaviors, such as sex with multiple partners, are at risk for developing vaginal cancer. Other associated infectious agents are herpes simplex virus (HSV) and Trichomonas vaginalis. In 2000, Lee and colleagues reported a case of rapidly progressive vaginal squamous cell carcinoma in a young woman with a 2-year history of human immunodeficiency virus (HIV) infection.2 They suggest that young women infected with both HIV and HPV are at increased risk for a more aggressive and less responsive vaginal cancer.
Another association that strengthens the link between HPV infection and vaginal cancer is the presence of a premalignant lesion in the vagina, known as vaginal intraepithelial neoplasia (VAIN). In 1991, Aho and coworkers reported that 5-9% of patients treated for VAIN progressed to invasive carcinoma.3 This suggests that VAIN may be a precursor to vaginal cancer even though the incidence of VAIN in the United States is 0.2-0.3 per 100,000 women, which is less than the incidence of diagnosed vaginal cancer. This is because of the fact that women with VAIN are usually asymptomatic and that screening for VAIN is not recommended for the general population. Still, the true malignant potential of VAIN needs to be identified.
A previous history of cervical intraepithelial neoplasia (CIN), invasive cervical carcinoma, or invasive vulvar carcinoma has also been associated with vaginal carcinoma. Several studies indicate that up to 30% of patients with primary vaginal carcinoma have a previous history of in situ or invasive carcinoma that was treated at least 5 years before diagnosis.
Long-term pessary use and chronic irritation of vaginal mucosa in women with procidentia have been associated with vaginal cancer. Other predisposing factors include cigarette smoking, immunosuppressive therapy, chemotherapy, and radiation therapy. Approximately 10% of women diagnosed with primary vaginal carcinoma have a previous history of irradiation to the pelvis. In a 2000 report, Carthew and colleagues demonstrated that tamoxifen, a chemotherapeutic drug, induced endometrial and vaginal cancer in rats in the absence of endometrial hyperplasia.4
In 1999, Pukkala and colleagues reported an association between low socioeconomic class in Finland and an increased incidence of cervical, endometrial, and vaginal cancer.5
Diethylstilbestrol (DES), a drug previously used in the first trimester to prevent pregnancy loss, has a strong association with clear cell adenocarcinoma of the vagina. Herbst and colleagues first observed this association in 19716, which led to the discontinuation of DES that same year. By 1987, the Registry for Hormonal Transplacental Carcinogenesis, established by Herbst and Scully, identified 524 women with clear cell adenocarcinoma, but only 60% had a history of DES exposure. Disease in the other 40% of patients with no history of DES exposure could be explained by recall bias or exposure to other unidentified factors. Women with in utero exposure to DES are at higher risk of developing adenocarcinoma than the general population. The estimated risk in these women is 1 in 1000.
Although 59% of women with vaginal cancer had a prior hysterectomy, in a 1986 report, Herman and colleagues demonstrated that when age and prior cervical cancer are controlled for, risk of vaginal cancer is not increased following hysterectomy for benign disease.7 Note that hysterectomy by itself is not a risk factor, rather women who underwent hysterectomy were poorly monitored.
In a 2004 publication, Hellman et al in Sweden reviewed 341 cases of primary carcinoma of the vagina from 1956-1996 and suggested that the etiology of vaginal cancer may be age related.8 In younger women, the disease occurred in the upper part of the vagina and seemed to be related to cervical dysplasia and HPV infection, while in older patients, the tumors were exophytic. There was significant correlation with late menarche, suggesting hormonal factors and trauma to the vagina as probable etiologies.
Pathophysiology
The presence of different stages of histologic differentiation—VAIN, carcinoma in situ, possible microinvasive carcinoma, and invasive cancer—suggests a continuum of transformation from less malignant to more invasive, which is similar to the continuum described for cervical cancer. As reported by Ikenberg et al in 19909 and Ostraw et al in 198810, identification of HPV DNA in squamous cell cancer cells by in situ hybridization (21%) and southern blot hybridization (56%) strongly suggests an association with HPV infection and a possible role of HPV in the pathogenesis of squamous cell vaginal carcinoma.
On the other hand, the significant association with a previous history of cervical or vulvar cancer suggests that the entire genital tract is at risk for squamous cell carcinoma once malignancy has occurred anywhere along the tract; this is a phenomenon postulated by Marcus and is known as the field effect. HPV infection could explain this phenomenon because HPV is associated with cervical, vaginal, and vulvar disease. Koyamatsu et al did a comparative analysis of the presence of HPV types 16 and 18 by polymerase chain reaction (PCR) and expression of p53 gene and Ki-67 antigen using immunohistochemistry in cervical, vaginal, and vulvar cancer.11 They suggested that in cervical cancer, HPV 16 and 18 played a common causal role, and in vulvar cancer, p53 gene mutations were the main carcinogenic cause, while vaginal cancer has transitional characteristics between cervical and vulvar cancer. There was no significant difference in overexpression of Ki-67 antigen among the 3 cancers.
Another explanation for this association is that an occult residual disease, such as VAIN, is trapped within the vaginal cuff posthysterectomy and goes unnoticed until it develops into invasive carcinoma. This possibility illustrates the theory of the field effect and HPV infection because HPV has also been linked to VAIN. It also partially explains why women with previous hysterectomy go unnoticed until they present with advanced-stage vaginal carcinoma.
The third possibility is radiation carcinogenesis.
The pathogenesis by which DES might play a role in inducing clear cell adenocarcinoma is unclear. In 1972, Forsberg and colleagues12 proposed the possibility of estrogen-induced maturation arrest of the müllerian ducts, and in 1984 Robboy and colleagues13 suggested that atypical vaginal adenosis and atypical cervical ectropion of the tuboendometrial type might act as the precursors of clear cell adenocarcinoma of the vagina and cervix.
Most vaginal cancers occur in the upper third of the vagina. Reports are contradictory as to whether the anterolateral wall or the posterior wall is the more frequent site. Reports suggesting that the upper posterior wall is the most common site favor the hypothesis that irritating substances, such as vaginal secretions and semen, pool in the posterior fornix and cause chronic irritation, which could lead to induction of a carcinogenic process.
The proximity of the bladder anteriorly and the rectum posteriorly to the vagina predisposes these organs to direct invasion by the tumor. Lymphatic dissemination follows the lymphatic drainage of the vagina. The middle-to-upper vagina communicates superiorly with the lymphatics of cervix and drains into the pelvic obturator node, the internal and external iliac chains, then to the para-aortic nodes. The distal third of vagina drains to the inguinal node then the pelvic node. Posterior wall lymphatics communicate with rectal lymphatics and drain to the inferior gluteal, sacral, and rectal nodes. Hematogenous dissemination to distant sites includes the lungs, liver, bone, and skin. A submucosal lesion suggests that the malignancy is metastatic via the vaginal lymphatics.
Clinical
Duration of symptoms averages 6-12 months before diagnosis, with a range of 0-11 years. Delay in diagnosis of vaginal carcinoma is not uncommon, and this is partially because of disease rarity and delay in relating patient symptoms to a vaginal origin. As expected, the longer the delay, the more advanced the cancer once the diagnosis is made, which results in a poorer outcome.
Painless vaginal bleeding is the most common symptom, accounting for 65-80% of all presentations. Bleeding is postmenopausal in about 70% of patients, which is consistent with the peak age of 60 years for squamous cell carcinoma, the most common type. Menorrhagia, intermenstrual bleeding, and postcoital bleeding have also been reported.
Vaginal discharge occurs in 30% of patients, while 20% of patients report urinary symptoms, which are caused by an anterior lesion compressing or invading the bladder, the urethra, or both. This causes bladder pain, dysuria, urgency, and hematuria.
About 15-30% of patients present with pelvic pain. Posterior lesions compress or invade the rectosigmoid, which causes tenesmus or constipation.
Only 10% of patients report a vaginal mass or vaginal prolapse. In 2000, Eltabbakh and coworkers reported a single patient who presented with a cystic pelvic mass arising from the posterior vaginal wall that mimicked an ovarian neoplasm.14
About 10-27% of patients are asymptomatic; diagnosis is made during routine pelvic examination. These patients tend to be caught at a much earlier stage than those presenting with symptoms, and their prognosis is much better.

Share This