CHAPTER 32: Vaginal Cancer

Cancer found in the vagina is most likely metastatic disease. Primary vaginal carcinoma is rare and makes up only 3 percent of all gynecologic malignancies (Siegel, 2015). This low incidence reflects the infrequency with which primary carcinoma arises in the vagina and the strict criteria for its diagnosis. According to International Federation of Gynecology and Obstetrics (FIGO) staging criteria, a vaginal lesion that involves adjacent organs such as the cervix or vulva, by convention, is deemed primary cervical or vulvar, respectively (Pecorelli, 1999). The most common histologic type of primary vaginal cancer is squamous cell carcinoma, followed by adenocarcinoma (Platz, 1995).

During embryogenesis, the müllerian ducts fuse caudally to form the uterovaginal canal (Chap. 18). The canal’s distal portion forms the proximal vagina, whereas the distal vagina arises from the urogenital sinus. The uterovaginal canal is lined by columnar epithelium, which is subsequently replaced by squamous cells migrating cephalad from the urogenital sinus. These squamous cells stratify, and the vaginal epithelium matures and thickens. Underlying this epithelium, muscularis and adventitial layers surround the vaginal tube.

With vaginal cancer, local extension and lymphatic invasion are common patterns of spread. The lymphatic channels that drain the vagina form extensive, complex, and variable anastomoses. As a result, any node in the pelvis, groin, or anorectal area may drain any part of the vagina. Of these, the external, internal, and common iliac lymph nodes are the primary sites of vaginal lymphatic drainage. Thus, pelvic lymphadenectomy, which samples these nodal groups, is commonly performed during primary surgical excision of proximally located vaginal cancers. Alternatively, lymphatic vessels of the posterior vagina may empty into the inferior gluteal, presacral, or perirectal nodes, and those of the vagina’s distal third may drain to the superficial and deep inguinal lymph nodes (Frank, 2005).

Hematogenous spread of vaginal cancer is less frequent, and venous drainage consists of the uterine, pudendal, and rectal veins, which drain into the internal iliac vein. Arterial blood supply to the vagina comes primarily from internal iliac artery branches, which include the uterine, vaginal, middle rectal, and internal pudendal arteries (Fig. 38-12).

According to estimates for 2015, 4070 new cases of vaginal cancer will be diagnosed in the United States, and there will be 910 deaths (Siegel, 2015). The overall incidence is 0.45 cases per 100,000 women but is notably lower in whites (0.42 cases) compared with black and Hispanic women (0.73 and 0.56 cases, respectively). Vaginal cancer rates increase with age and peak among women ≥80 years. The median age at diagnosis is 58 (Watson, 2009). Of histologic forms, squamous cell carcinoma accounts for 70 to 80 percent of all primary vaginal cancer cases (Beller, 2003; Platz, 1995).

Risks

Squamous cell cancer of the vagina arises within its stratified nonkeratinized epithelium (Fig. 32-1). As with other cancers of the lower reproductive tract, human papillomavirus (HPV) has been closely linked with squamous cell vaginal cancer (Chap. 30). A systematic review found an HPV prevalence of 65 percent in samples from invasive vaginal cancer, and a 93-percent prevalence in high-grade vaginal dysplasia lesions. HPV 16 was the most common type and was present in 55 percent of vaginal cancer samplings (Smith, 2009). A retrospective cross-sectional study involving 31 countries found similar results (Alemany, 2014).

Because of this association with HPV infection, vaginal in situ and invasive squamous cell carcinomas share risk factors similar to those for cervical cancer. Some of these include multiple lifetime sexual partners, early age at first intercourse, and current cigarette smoking. Women with a vulvar or cervical cancer history are also at increased risk. This last association may stem from the field effect of HPV affecting multiple lower genital tract epithelia or may result from direct tumor spread.

Vaginal intraepithelial neoplasia (VaIN) is a precursor to invasive vaginal cancer, and approximately 2 to 3 percent of patients with VaIN will progress to invasive cancer (Dodge, 2001; Ratnavelu, 2013). The quadrivalent HPV vaccine is effective in preventing VaIN 2 and 3 associated with HPV 16 or 18 (Joura, 2007). It is possible that use of this vaccine will decrease invasive vaginal cancer rates in the future.

Diagnosis

Vaginal bleeding is the most common complaint associated with vaginal cancer, although pelvic pain and vaginal discharge also may be noted. Less frequently, lesions involving the anterior vaginal wall may lead to dysuria, hematuria, or urgency. Alternatively, constipation may result from posterior wall masses. Most vaginal cancers develop in the upper third of the vagina. Moreover, of those with cancers, women who have had a prior hysterectomy are significantly more likely to have lesions in the upper vagina (70 percent) than those without prior hysterectomy (36 percent) (Chyle, 1996).

During pelvic evaluation in all women, the vagina is inspected as the speculum is inserted or removed. If a gross lesion is found, vaginal cancer usually can be diagnosed by punch biopsy in the office. Biopsy may be obtained with Tischler biopsy forceps (Fig. 29-16). An Emmett hook, one type of skin hook, may be useful to elevate and stabilize vaginal tissue during biopsy. Monsel paste is applied as needed for hemostasis. If a gross lesion is not detectable, vaginoscopy can guide directed biopsy. Bimanual examination assists in determining the tumor size, and rectovaginal examination is especially important for posterior wall lesions.

Once cancer is diagnosed, no specific laboratory testing other than that used generally for preoperative preparation such as complete blood count and serum chemistry panel is required.

Staging and Classification

Vaginal cancer staging is similar to that for cervical cancer and is completed clinically by physical examination and with the assistance of cystourethroscopy and/or proctosigmoidoscopy depending on tumor location. Chest radiography aids the search for metastatic disease (Table 32-1 and Table 32-2). If needed, general anesthesia can permit a more detailed pelvic examination for accurate staging. Proctosigmoidoscopy to a depth of at least 15 cm can detect local bowel invasion, whereas cystourethroscopy assists identification of bladder or urethral involvement.

Computed tomography (CT) scanning, magnetic resonance (MR) imaging, and fluorodeoxyglucose-positron emission tomography (FDG-PET) may also be useful in treatment planning but are not classically used to determine disease stage. CT scanning can delineate the size and extent of many tumors (Fig. 32-2). However, if the extent of cancer expansion is unclear, MR imaging is the most useful radiologic tool available to visualize the vagina due to its superior soft tissue resolution. FDG-PET can also be selected to evaluate lymph node involvement and distant metastases. In one study, FDG-PET was more sensitive than CT for detection of abnormal lymph nodes (Lamoreaux, 2005).

Prognosis

The prognosis of vaginal squamous cell carcinoma has improved since the 1950s, when the 5-year survival rate was only 18 percent. Advances in radiation technology and earlier diagnosis are largely responsible for the improved 5-year survival rate, which now ranges from 45 to 68 percent for all stages (Ghia, 2011; Hellman, 2006).

The prognosis of vaginal squamous cell carcinoma depends primarily on FIGO stage (Fig. 32-3 and Table 32-2) (Frank, 2005; Peters, 1985). Other factors associated with poor prognosis include larger tumor size, adenocarcinoma cell type, and older age (Hellman, 2006; Tjalma, 2001; Tran, 2007). The 5-year disease-specific survival rate is 85 to 92 percent for women with stage I disease, 68 to 78 percent for those with stage II, and 13 to 58 percent for those with stage III or IV (Fig. 32-4) (Frank, 2005; Tran, 2007).

Treatment

Stage I

Because of vaginal cancer’s rarity, data are limited to guide evidence-based treatment. Therefore, therapy is individualized and based on factors such as tumor type, stage, location, and size.

For stage I disease, both surgery and radiotherapy are options. However, surgery is preferred for most if negative surgical margins can be achieved. Surgery includes radical vaginectomy, radical hysterectomy (for women with an intact uterus), and pelvic lymphadenectomy for most tumors located in the upper third of the vaginal vault. A review of the National Cancer Data Base showed that women with stage I disease treated with surgery alone had a significantly improved 5-year survival rate compared with those treated with radiation (90 percent versus 63 percent) (Creasman, 1998). However, other reports have found no significant difference in disease-free survival rates in women with stage I disease treated with surgery compared with radiation alone (Stock, 1995). Radiotherapy may be delivered by external beam, with or without brachytherapy. Further, brachytherapy alone has been used successfully to treat selected small stage I lesions (Nori, 1983; Perez, 1999; Reddy, 1991).

Stage II

Depending on circumstances and the treating clinician, primary surgery or radiation may be selected for stage II cancers. Stock and colleagues (1995) found a significant survival advantage at 5 years in those with stage II disease treated with surgery compared with those treated with radiation (62 percent versus 53 percent). Review of the National Cancer Data Base showed that the 5-year survival rate for women with stage II disease treated with surgery alone was 70 percent; with radiotherapy alone, 57 percent; and with a combination of surgery and radiotherapy, 58 percent (Creasman, 1998). However, other researchers have found no survival advantage from surgery compared with radiotherapy in stage II disease (Davis, 1991; Rubin, 1985).

Radiation is generally recommended if negative margins cannot be achieved surgically due to tumor location or size or to patient comorbidities that preclude surgery. If primary radiation is administered for stage II disease, a combination of external beam radiation and brachytherapy is the most common regimen. External beam radiation generally is given first, and depending on tumor response, brachytherapy is tailored to remaining disease. As discussed subsequently, adjuvant chemotherapy is often administered during radiation therapy.

Stage III and IVA

For advanced disease, external beam radiation alone or in combination with brachytherapy is usually administered (Frank, 2005). Concurrent chemotherapy with cisplatin as a radiation sensitizer is generally also recommended.

Stage IVB

Metastatic vaginal cancer is not curable, and treatment may include systemic chemotherapy or supportive hospice care. The most common sites of distant spread include liver, lung, and bone. The choice of chemotherapeutic agents is commonly extrapolated from cervical cancer data. For example, bevacizumab (Avastin) was approved by the Food and Drug Administration (FDA) in 2014 as a treatment for metastatic cervical cancer, in combination with paclitaxel and either cisplatin or topotecan. The addition of bevacizumab to these chemotherapy combinations improved overall survival length by approximately 4 months in women with metastatic cervical cancer (Tewari, 2014).

Chemoradiation

The numbers of women with vaginal cancer have been too small to make a prospective, randomized trial feasible. However, concurrent chemotherapy with cisplatin can be considered to benefit those with locally advanced vaginal cancer because of its proven efficacy in cervical cancer treatment. The characteristics of this agent are described in Chapter 27 and Figure 30-13. In a small series of vaginal cancer cases, adjuvant chemotherapy provided a 10- to 33-percent decline in the total radiation dose delivered (Dalrymple, 2004). Although not intending to show an improved survival rate with chemoradiation, the authors found that local control of tumor growth and survival rates were comparable with those who had received higher radiation doses alone. Smaller total radiation doses may lead to lower rates of vaginal stenosis and fistula formation.

In a Surveillance Epidemiology and End Result (SEER) database analysis of 326 patients with vaginal cancer treated with external beam radiation and/or brachytherapy between 1991 and 2005, a notable increase in sensitizing chemotherapy use was observed after the 1999 National Cancer Institute announcement confirmed the efficacy of chemoradiation in cervical cancer. Despite this increased use, the authors did not observe any survival advantage among those vaginal cancer patients who received chemoradiation compared with radiation alone (Ghia, 2011).

A recent phase II trial, which included 22 cervical cancer patients and three with stage II–IV vaginal cancer, demonstrated a 96-percent response rate for women given weekly cisplatin, radiation, and triapine, a ribonucleotide reductase inhibitor (Kunos, 2013). Larger studies are planned in cervical cancer patients.

Chemotherapy

In general, chemotherapy alone is ineffective to treat vaginal cancer, although data are limited. The Gynecologic Oncology Group (GOG) performed a Phase II trial evaluating 50 mg/m² doses of cisplatin given every 3 weeks for advanced or recurrent vaginal cancer in 26 patients. Only one woman with squamous cell carcinoma achieved a complete response. Five of 16 patients with squamous cell carcinoma had stable disease, and 10 had cancer progression. Based on this trial, single-agent cisplatin is considered to have insignificant activity at that dose and schedule (Thigpen, 1986). To date, this has been the only prospective GOG trial evaluating chemotherapy alone for vaginal cancer.

Radiation Therapy

Radiation to the primary tumor usually involves pelvic external beam with or without brachytherapy, and often concurrent platinum-based sensitizing chemotherapy depending on the stage and other factors described above. Additionally, groin radiation is effective in patients with palpable nodal metastases. Moreover, elective irradiation may be delivered to clinically negative inguinal lymph nodes if the distal third of the vagina is involved. In a retrospective review, Perez and colleagues (1999) found that of 100 women who did not receive groin radiation, if disease was confined to the upper two thirds of the vagina, then none developed groin metastases. However, 10 percent of patients with lower-third primary tumors and 5 percent with tumors involving the entire length of the vagina developed inguinal node metastases.

Surveillance

Treatment failures usually develop within 2 years of primary therapy completion. Thus, patients are typically examined every 3 months for the first 2 years and then every 6 months until 5 years of surveillance is completed (Pingley, 2000; Rubin, 1985). After 5 years following treatment, women can be seen annually. A Pap smear and pelvic examination with careful attention to the inguinal and scalene nodes are performed. Surveillance with CT or MR imaging is at the clinician’s discretion.

Recurrent Disease

Disease recurrence should be confirmed by biopsy if further treatment is planned. Therapeutic options in women with central pelvic recurrence who have had prior pelvic radiation are limited. Pelvic exenteration can be considered if a patient is psychologically and medically fit to undergo radical surgery associated with high morbidity (Chap. 46). It is attempted only in those whose disease is limited to the central pelvis. Therefore, clinicians are alert to the triad of sciatic pain, leg edema, and hydronephrosis, which suggests pelvic sidewall disease. These women are not surgical candidates but can be managed with chemoradiation or with chemotherapy alone for women previously irradiated.

Survival after relapse is poor. In a review of 301 patients, 5-year survival was 20 percent for local recurrence and 4 percent for metastatic disease recurrence (Chyle, 1996).

Primary adenocarcinoma of the vagina is rare, making up only 13 percent of all vaginal cancers (Platz, 1995). Histologic types include clear cell, endometrioid, mucinous, and serous carcinoma, and these may arise in endometriosis foci, in areas of vaginal adenosis, in periurethral glands, or in wolffian duct remnants. Adenosis in the vagina is defined by subepithelial glandular structures lined by mucinous columnar cells that resemble endocervical cells (Sandberg, 1965). These represent residual glands of müllerian origin. Clinically, adenosis appears as red granular spots or patches and does not stain following Lugol solution application. Vaginal adenosis is a condition common in females exposed in utero to diethylstilbestrol (DES) (Chap. 18).

More commonly, vaginal adenocarcinoma is metastatic disease, typically from a lesion higher in the genital tract. Disease is frequently metastatic from the endometrium, although it also may originate from the cervix or ovary (Saitoh, 2005). In addition, adenocarcinoma metastases from the breast, pancreas, kidney, and colon also have been identified in the vagina.

Treatment is similar to that for squamous cell carcinoma. Thus, surgery, radiation, or a combination of both can be used. Primary vaginal adenocarcinoma in general is more aggressive than squamous cell carcinoma. In one series of 30 patients, it was associated with local and metastatic relapse rates that were more than double those for squamous cell carcinoma (Chyle, 1996).

Clear Cell Adenocarcinoma

Of primary vaginal adenocarcinomas, the clear cell type is most closely associated with DES exposure. In the United States starting around 1940, DES was used off-label to prevent miscarriage. In 1971, clear cell adenocarcinoma of the vagina was linked to in utero DES exposure, and subsequently, pregnancy became a contraindication to DES use. Despite this, an estimated 1 to 4 million women used this synthetic estrogen and approximately 0.01 percent of females exposed in utero developed vaginal clear cell adenocarcinoma (Melnick, 1987). Most DES-exposed patients with vaginal cancer were born between 1951 and 1953, when the drug was prescribed most frequently. The median age at diagnosis of vaginal clear cell carcinoma in the United States is 19 years.

However, in the Netherlands, a bimodal distribution of vaginal clear cell carcinoma has been observed—the first peak occurring with a mean age of 26 years and the second at 71 years. The younger group had all been exposed in utero to DES, whereas the older group, born before 1947, had not been exposed (Hanselaar, 1997). It remains to be seen if the incidence of vaginal clear cell carcinoma will rise as the DES-exposed population ages.

Treatment is similar to that for squamous cell carcinoma of the vagina. The 5-year survival rate for 219 women with stage I vaginal clear cell adenocarcinoma was 92 percent and was equivalent regardless of therapy mode (Senekjian, 1987). The reported 5-year survival for 76 patients with stage II disease was 83 percent (Senekjian, 1988). A smaller study from MD Anderson shows worse pelvic control (31 versus 81 percent) and worse 5-year overall survival rate (34 versus 58 percent) for women with primary adenocarcinoma of the vagina not associated with DES exposure compared with squamous cell carcinoma of the vagina. Five-year survival rates for non-DES-associated vaginal cancer were 80 percent for stage I disease, 34 percent for stage II, 26 percent for stage III, and no survivors with stage IV disease (Frank, 2007).

Embryonal Rhabdomyosarcoma

This is the most common malignancy of the vagina in infants and children, and most embryonal rhabdomyosarcomas are the sarcoma botryoides subtype. This rare tumor develops almost exclusively in girls younger than 5 years, although vaginal and cervical sarcoma botryoides have been reported in females aged 15 to 20 years (Copeland, 1985). In infants and children, sarcoma botryoides usually is found in the vagina; in reproductive-aged women, within the cervix; and after menopause, within the uterus. Its name, derived from the Greek word botrys, which means “bunch of grapes,” describes its appearance (Fig. 32-5). The gross specimen can exhibit multiple polyp-like structures or can be a solitary growth with a nodular, cystic, or pedunculated appearance (Hilgers, 1970). Although this distinctive appearance may guide diagnosis, the classic histologic finding of this tumor is the rhabdomyoblast. Typical complaints include bleeding or a vaginal mass.

Embryonal rhabdomyosarcomas have a poor prognosis, but sarcoma botryoides is the easiest to treat and has the best chance of cure. It may be that its superficial location allows earlier detection (Copeland, 1985). Childhood sarcoma botryoides treatment has dramatically changed. It has gradually shifted away from radical pelvic exenteration surgery and toward primary chemotherapy plus adjuvant conservative surgery to excise residual tumor (Andrassy, 1995, 1999; Hays, 1981, 1985).

Leiomyosarcoma

This is the most common type of vaginal sarcoma in adults. However, it makes up no more than 1 percent of vaginal malignancies, and only 140 cases have been described in the literature to date (Ahram, 2006; Khosla, 2014; Suh, 2008). The age of affected individuals is broad, but most are older than 40 years (Zaino, 2011). Because of the small number of these tumors, their epidemiology has not been widely studied, and few risk factors have been identified. However, patients previously treated with pelvic radiotherapy for cervical cancer appear to be at risk.

Affected women most often complain of an asymptomatic vaginal mass, but other symptoms mirror those of their squamous cell counterpart. Any wall of the vagina may be affected, but most tumors develop posteriorly (Ahram, 2006). Microscopically, tumors resemble uterine leiomyosarcoma (Fig. 34-2). Tumors spread by local invasion and hematogenous dissemination.

Surgical resection with negative margins is the preferred primary treatment. The benefit of adjuvant radiation is unclear due to a lack of controlled trials. However, some clinicians recommend adjuvant radiation for those with high-grade tumor or local recurrence (Curtin, 1995).

Primary malignant melanoma in the vagina is rare, accounting for less than 3 percent of all vaginal cancers. In women, only 1.6 percent of melanomas are genital. The most common site is the vulva (70 percent), followed by the vagina (21 percent) and the cervix (9 percent) (Miner, 2004). Using data from the SEER database, the incidence of vaginal melanoma is between 0.26 and 0.46 per 1 million women per year (Hu, 2010; Weinstock, 1994). Both U.S. and Swedish studies have shown the mean age at diagnosis to be 66 years (Ragnarsson-Olding, 1993; Reid, 1989).

The most frequent presenting symptoms include vaginal bleeding, discharge, and vaginal mass (Fig. 32-6) (Gupta, 2002; Reid, 1989). Most are located in the distal vagina (Frumovitz, 2010; Xia, 2014). Vaginal melanoma is often detected late, and this may be largely responsible for poor treatment outcomes.

With a reported 5-year survival rate ranging from 10 to 20 percent, the prognosis is among the worst of vaginal malignancies (Ragnarsson-Olding, 1993; Weinstock, 1994; Xia, 2014). Although survival rates are significantly better for those with vaginal lesions measuring <3 cm, FIGO staging of vaginal melanomas does not accurately predict survival (Reid, 1989). Thus, staging criteria specific to melanoma are used. Cutaneous melanomas at other body sites are staged by microstaging systems, including the Chung, the Clark, and the Breslow systems, which use criteria such depth of invasion, tumor size, and tumor thickness (Chap. 31). However, Clark levels are not applicable to vaginal melanoma because the typical microscopic skin landmarks used are not present. Therefore, staging is based on tumor thickness, as described by Breslow or Chung.

Treatment is extrapolated from cutaneous melanomas, due to the rarity of vaginal melanoma. Surgery is preferred, when feasible. Although some advocate radical surgery, including exenteration, growing evidence shows that wide local excision has similar survival rates and less morbidity (Buchanan, 1998; Xia, 2014). However, the recommended clinical surgical margin for a melanoma with a Breslow thickness ≤1 mm is 1 cm; a thickness 1 to 2 mm warrants a 1- to 2-cm margin; and a thickness >2 mm requires a 2-cm margin (National Comprehensive Cancer Network, 2014). One study demonstrated a survival benefit to the wide local excision approach (Frumovitz, 2010). However, because of their size or location, many vaginal melanomas are not amenable to this less radical approach.

Melanomas generally are thought to be radioresistant. However, radiation therapy in one series was found to provide local tumor control in women who had surgically unresectable disease (Miner, 2004).

Treatment of advanced and metastatic cutaneous melanoma has progressed, and multiple targeted biologic agents are now available. Mutations in the BRAF and KIT oncogenes have been found in cutaneous and mucosal melanomas, and women with vaginal melanoma have their tumor tested for these mutations (Leitao, 2014). Options for those with BRAF ^V600E mutations include vemurafenib, dabrafenib, and trametinib (Chapman, 2011; Flaherty, 2012a,b; Hauschild, 2012; Sosman, 2012). Imatanib may be used for tumors with the c-KIT mutation (Carvajal, 2011). Ipilimumab is a monoclonal antibody that promotes T-cell activation, which in turn produces antitumor effects. Its use in patients with metastatic melanoma improves overall survival rates (Hodi, 2010). Recently, the monoclonal antibodies pembrolizumab and nivolumab, which target programmed death 1 protein (PD-1), have been compared with and also combined with ipilimumab for advanced melanoma. The anti-PD-1 immunotherapy drugs are now preferred for the treatment of advanced melanoma, as they offer increased progression-free survival rates, objective-response rates, but fewer serious adverse events than ipilimumab monotherapy (Postow, 2015; Robert, 2015).