Neville F. Hacker
Primary carcinomas of the vagina represent 2% to 3% of malignant neoplasms of the female genital tract. In the United States, it is estimated that there will be 2,160 new cases diagnosed in 2009, and 770 deaths from the disease (1). More than 50% of patients are diagnosed in the seventh, eighth, and ninth decades, and squamous cell histology accounts for about 80% of cases (2).
Until the late 1930s, vaginal cancer was in general considered to be incurable. Most patients presented with disease that had spread beyond the vagina, and radiation therapy techniques were poorly developed. With modern techniques for radiation therapy, cure rates of even advanced cases should now be comparable with those for cervical cancer(3,4,5). According to the Annual Report, the overall 5-year survival rate has increased from 34.1% in 1959 to 1963 to 53.6% in 1999 to 2001 (2).
Fu (6) reported that 84% of carcinomas involving the vagina were secondary, usually from the cervix (32%); endometrium (18%); colon and rectum (9%); ovary (6%); or vulva (6%). Of 164 squamous cell carcinomas, 44 (27%) were primary and 120 (73%) were secondary. Among the latter, 95 (79%) originated from the cervix; 17 (14%) from the vulva; and 8 (7%) from the cervix and the vulva (6). This apparent discrepancy is partly related to the International Federation of Gynecology and Obstetrics (FIGO) classification and staging of malignant tumors of the female pelvis. The staging requires that a tumor that has extended to the portio and reached the area of the external os should be regarded as a carcinoma of the cervix, whereas a tumor that involves the vulva and vagina should be classified as a carcinoma of the vulva. Endometrial carcinomas and choriocarcinomas commonly metastasize to the vagina, whereas tumors from the bladder or rectum may invade the vagina directly.
Primary Vaginal Tumors
The histologic types of primary vaginal tumors are shown in Table 14.1 (7,8,9,10,11,12,13,14,15,16,17,18,19). Squamous cell carcinomas are the most common, although adenocarcinomas, melanomas, and sarcomas are also seen. Sarcomas occasionally follow radiation therapy for cervical cancer.
Squamous Cell Carcinoma
Squamous cell carcinoma is the most common vaginal cancer. The mean age of the patients is approximately 67 years, although the disease occasionally is seen in the third and fourth decades of life (4,7,10,13). About 80% of patients are older than 50 years (2).
Table 14.1 Primary Vaginal Cancer: Reported Incidence of Histologic Types
Women who have been treated for a prior anogenital cancer, particularly of the cervix, have a high relative risk of developing vaginal cancer, although the absolute risk is low (20).
In a population-based study of 156 women with in situ or invasive vaginal cancer, Daling et al. determined that they had many of the same risk factors as patients with cervical cancer, including a strong relationship with human papilloma virus (HPV) infection (20). The presence of antibodies to HPV 16 was strongly related to this risk. A study of 341 cases from the Radiumhemmet reported that the disease seemed to be etiologically related to cervical cancer, and thus HPV infection, in young patients, but in older patients, there was no such association (21).
As many as 30% of patients with primary vaginal carcinoma have a history of in situ or invasive cervical cancer treated at least 5 years earlier (11,12,13). In a report from the University of South Carolina, a past history of invasive cervical cancer was present in 20% of the cases and of cervical intraepithelial neoplasia (CIN) in 7% (15). The median interval between the diagnosis of cervical cancer and the diagnosis of vaginal cancer was 14 years, with a range of 5 years, 8 months to 28 years. Sixteen percent of the patients had a history of prior pelvic irradiation.
There are three possible mechanisms for the occurrence of vaginal cancer after cervical neoplasia:
In the first instance, extension of intraepithelial neoplasia from the cervix to the upper vagina was not appreciated and an adequate vaginal cuff was not taken because vaginal colposcopy was not performed before surgical management of the cervical tumor. Surgical margins of the upper vaginal resection usually show in situ neoplasia, and these persistent foci eventually progress to invasive disease. In the second instance, vaginal colposcopy is negative, and the surgical margins of resection are free of disease.
There is controversy regarding the distinction between a new primary vaginal cancer and a recurrent cervical cancer. Many authorities use a 5-year cut-off because 95% of cervical cancers will recur within this period (22,23) but others prefer a 10-year interval (19).
Prior pelvic radiation therapy has been considered a possible cause of vaginal carcinoma (8). In the series of 314 patients with squamous cell carcinoma of the vagina reported from Sweden, previous pelvic radiation was reported by 44 patients (14%) on an average of 22 years earlier (range 5-55 years)(19). Judicious use of pelvic radiation may be particularly important in young patients, who may live long enough to develop a second neoplasm in the irradiated vagina (24).
Figure 14.1 Squamous cell carcinoma of the vagina in a patient with a procidentia. The cancer was apparently related to long-term pessary use.
The true malignant potential of vaginal intraepithelial neoplasia is unclear because once diagnosed, the condition is usually treated. Benedet and Saunders (25) reviewed 136 cases of carcinoma in situ of the vagina seen over a 30-year period. Four cases (3%) progressed to invasive vaginal cancer in spite of various methods of treatment. Rome and England reported 9 cases (6.8%) of early invasive vaginal cancer detected during the initial management of 132 cases of vaginal intraepithelial neoplasia (VAIN) (26).
Chronic local irritation from long-term use of a pessary may also be of significance (6) (Fig. 14.1), although pessaries are used less commonly in modern gynecology.
For screening to be cost effective, the incidence of the disease must be sufficient to justify the cost of screening. In the United States, the age-adjusted incidence of vaginal cancer is 0.6 per 100,000 women, making routine screening of all patients inappropriate (27). However, women with a history of cervical intraepithelial or invasive neoplasia are at increased risk and should be monitored carefully with Pap smears.
As many as 59% of patients with vaginal cancer have had a prior hysterectomy (9). However, when age and prior cervical disease are controlled for, there is no increased risk of vaginal cancer in women who have had a hysterectomy for benign disease (28).
Symptoms and Signs
Most patients with vaginal cancer present with painless vaginal bleeding and discharge. The bleeding is usually postmenopausal but may be postcoital. In the large series reported from the Radiumhemmet, 14% of patients were asymptomatic, and the diagnosis was made either by routine examination (7%) or by abnormal cytology (7%) (19).
Because the bladder neck is close to the vagina, bladder pain and frequency of micturition occur earlier than with cervical cancer. Posterior tumors may produce tenesmus. Approximately 5% of patients present with pelvic pain because of extension of disease beyond the vagina.
Most lesions are situated in the upper one-third of the vagina, usually at the apex or on the posterior wall (19). Macroscopically, the lesions are usually exophytic (fungating, polypoid), but they may be endophytic. Surface ulceration usually occurs late in the course of the disease.
The diagnosis of carcinoma of the vagina is often missed on first examination, particularly if the lesion is small and situated in the lower two-thirds of the vagina, where it may be covered by the blades of the speculum. Definitive diagnosis is usually made by biopsy of a gross lesion, which can often be performed in the office without anesthesia. Particularly in elderly patients or in those with some degree of vaginal stenosis, examination under anesthesia may be desirable to allow adequate biopsy and clinical staging. The latter may require cystoscopy or proctoscopy, depending on the location of the tumor.
In patients with an abnormal Pap smear and no gross abnormality, careful vaginal colposcopy and the liberal use of Lugol's iodine to stain the vagina are necessary. This is performed in the office initially, but may need to be repeated with the patient under regional or general anesthesia to allow excision of colposcopically abnormal lesions. For definitive diagnosis of early vaginal carcinoma, it may be necessary to resect the entire vaginal vault and submit it for careful histologic evaluation because the lesion may be partially buried by closure of the vaginal vault at the time of hysterectomy. This is usually done with a cold knife, but Fanning et al. have reported the successful use of the loop electrosurgical procedure for partial upper vaginectomy in 15 patients (29). Inadvertent cystotomy may occur occasionally, and this requires immediate repair.
Hoffman et al. (30) at the University of South Florida reported on 32 patients who underwent upper vaginectomy for VAIN 3. Occult invasive carcinoma was found in nine patients (28%). In five cases, the depth of invasion was less than 2 mm, but in four cases, invasion ranged from 3.5 mm to full-thickness involvement.
The FIGO staging (2009) for vaginal carcinoma is shown in Table 14.2. Updated FIGO staging tables can be seen on pages 665 to 668. The staging is clinical and is based on the findings at general physical and pelvic examination, cystoscopy, proctoscopy, chest x-ray, and possible skeletal radiographs if the latter are indicated because of bone pain.
Because it is difficult to determine accurately any spread into subvaginal tissues, particularly from anterior or posterior lesions, differences in observations are common. This is reflected in the wide range of stage distributions reported and the wide range of survivals within a given stage. The distribution by FIGO stage from 13 series is shown in Table 14.3. Fewer than onethird of patients present with disease confined to the vagina, although Hellman et al. reported that significantly more patients were diagnosed at an early stage in the last 20 years of their study, compared to the first 20 years (19).
Surgical staging for vaginal cancer has been used less commonly than for cervical cancer, but in selected premenopausal patients, a pretreatment laparotomy may allow better definition of the extent of disease, excision of any grossly enlarged lymph nodes, and placement of an ovary up into the paracolic gutter beyond the radiation field.
Table 14.2 Carcinoma of the Vagina: FIGO Nomenclature
Table 14.3 Primary Vaginal Carcinoma: Distribution by Stage of Disease
Peters et al. (34) suggested criteria for microinvasive carcinoma of the vagina: focal invasion associated with VAIN 3, no lymph-vascular invasion, free margins on partial or total vaginectomy, and a maximum depth of invasion of less than 2.5 mm, measured from the overlying surface. However, Eddy et al. (35) reported six patients who met these criteria and were treated by either partial or total vaginectomy. In one of the six, a bladder recurrence developed at 35 months.
Patterns of Spread
Vaginal cancer spreads by the following routes:
There is little information available on the incidence of lymph node metastases in vaginal cancer because most patients are treated with radiation therapy. Rubin et al. (13) reported that 16 of 38 patients (42.1%) with all stages of disease had lymphangiographic abnormalities, but many of these abnormalities were not confirmed histologically. Al-Kurdi and Monaghan (36) performed lymph node dissections on 35 patients and reported positive pelvic nodes in 10 patients (28.6%). Positive inguinal nodes were present in 6 of 19 patients (31.6%), with disease involving the lower vagina. Stock et al. (32) reported positive pelvic nodes in 10 of 29 patients (34.5%) with all stages of disease who underwent bilateral pelvic lymphadenectomy as part of their therapy or staging. Positive paraaortic nodes were present in one of eight patients (12.5%) undergoing paraaortic dissection.
Van Dam et al. described sentinel node detection in three patients with primary vaginal cancer (37). Preoperatively, 60-mBq technetium-labeled nanocolloid was injected around the tumor, and sentinel nodes were detected using a laparoscopic or hand-held probe. Sentinel nodes could be found in two of the three patients. The first patient described had a 1-cm tumor in the distal anterior vagina. Lymphoscintigraphy revealed sentinel nodes in the groin and obturator fossa, but both were negative histologically. The second patient had squamous cell carcinoma at the vaginal vault, which followed an abdominal hysterectomy two years earlier for a stage Ia1 squamous cervical carcinoma. Regardless of whether this is considered a vaginal or a recurrent cervical cancer, lymphoscintigraphy demonstrated two sentinel nodes just below the common iliac bifurcation, and both were histologically positive.
Apart from the standard staging investigations, a computed tomographic (CT) or magnetic resonance imaging (MRI) scan of the pelvis and abdomen is useful for evaluation of the status of the primary tumor, liver, pelvic and paraaortic lymph nodes, and ureters.
The group in St. Louis compared CT and F-18 fluorodeoxy glucose (FDG) positron emission tomography (PET) for the detection of the primary tumor and lymph node metastases in patients with carcinoma of the vagina (38). Of the 21 patients with an intact primary tumor, CT visualized the tumor in 9 patients (43%), whereas FDG-PET identified abnormal uptake in all 21 cases (100%). Similarly, CT demonstrated abnormally enlarged lymph nodes in 17% of cases, compared to abnormal uptake in 35% of patients with FDG-PET.
Experience with the management of primary vaginal cancer is limited because of the rarity of the disease. Most gynecologic oncology centers in the United States see only two to five new cases per year, and even in some European centers, where referral of oncology cases tends to be more centralized, only approximately one new case per month can be expected (31). Therapy must be individualized and varies depending on the stage of disease and the site of vaginal involvement, further limiting individual experience.
Anatomic factors and psychological considerations place significant constraints on treatment planning. The proximity of the vagina to the rectum, bladder, and urethra limits the dose of radiation that can be delivered and restricts the surgical margins that can be attained unless an exenterative procedure is performed. For most patients, maintenance of a functional vagina is an important factor in the planning of therapy.
Surgery has a limited role in the management of patients with vaginal cancer because of the radicality required to achieve clear surgical margins, but in selected cases, satisfactory results can be achieved (9,17,19,32,36,39,40,41). Surgery may be useful in the following circumstances:
A Chinese report described four patients who had laparoscopic radical hysterectomy, pelvic lymphadenectomy, and total vaginectomy for stage I vaginal carcinoma (39). Vaginal reconstruction was performed using the sigmoid colon. Mean operating time was 305 minutes (range 260-350 minutes), mean blood loss 325 mL (range 250-400 mL), and median number of resected lymph nodes 16 (range 13-20). The mean postoperative stay was 7 days (range 6-8) and all patients were clinically free of disease with a mean follow-up of 46 months (range 40-54 months).
A surgical approach to conserve reproductive and sexual function was described from Italy (40). Three nulliparous women under 40 years of age with stage I squamous cell carcinoma confined to the upper third of the vagina underwent radical tumorectomy and pelvic lymphadenectomy. A fourth patient underwent partial hemivaginectomy plus ipsilateral paracolpectomy and pelvic lymphadenectomy. One patient with microscopic involvement of the paracolpium received adjuvant radiation after laparoscopic ovarian transposition. With a follow-up of 9 to 51 months, all patients were regularly menstruating, sexually active, and clinically free of disease.
Although this approach has obvious appeal, the number of cases is very limited, and such an approach should probably be limited to small lesions, particularly if located in the upper vagina. The authors recommended more intensive follow-up, particularly aimed at early detection of any locoregional relapse and/or second tumor of the lower genital tract (40).
Radiation therapy is the treatment of choice for all patients except those listed previously, and comprises an integration of teletherapy and intracavitary/interstitial therapy(3,4,5,7,18,19,31,33). Selected stage I and II lesions can be treated adequately with brachytherapy alone (3,4,33,42). For larger lesions, treatment is usually started with approximately 5,000 cGy external irradiation to shrink the primary tumor and treat the pelvic lymph nodes. Intracavitary or interstitial treatment follows.
There is improved local control with total tumor doses of at least 7,000 cGy (3,43). If the uterus is intact and the lesion involves the upper vagina, an intrauterine tandem and ovoids can be used. If the uterus has been previously removed, a Bloedorn type of applicator or vaginal cylinder may be used. If the lesion is more deeply invasive (thicker than 0.5 cm), interstitial irradiation, alone or in conjunction with the intracavitary therapy, improves the dose distribution (18,33). Extended-field radiation has rarely been used for patients with vaginal cancer, but if positive paraaortic nodes are documented after either surgical staging, CT scanning and fine-needle aspiration cytologic evaluation, or PET scanning, this treatment should be given. If the lower one-third of the vagina is involved, the groin nodes should be treated or dissected.
There is limited reported experience with chemoradiation for vaginal cancer (4,44), although many centers now routinely use this combined therapy, as is being done for cervical carcinoma. In our center, cisplatin 35 mg/m2 given on the first day of each week of external beam therapy is used. The small number of cases makes it virtually impossible to ever conduct a randomized, prospective study to compare standard radiation with chemoradiation for patients with vaginal cancer.
Complications of Therapy
Major complications of therapy are usually reported in 10% to 15% of patients treated for primary vaginal cancer, whether the treatment is by surgery or radiation. Workers at the M. D. Anderson Hospital reported serious complications in 39 of 311 patients (13%), but estimated an actuarial incidence of 19% at 20 years (3). The close proximity of the rectum, bladder, and urethra predisposes these structures to injury, and radiation cystitis, radiation proctitis, rectovaginal or vesicovaginal fistulas, and rectal strictures or ulceration may occur. Radiation necrosis of the vagina occasionally occurs, and radiation-induced fibrosis and subsequent vaginal stenosis are a constant concern.
Stryker reported vaginal morbidity in 9 of 15 patients (60%) undergoing external beam therapy plus intracavitary brachytherapy, and 3 of 10 patients (30%) undergoing external beam therapy plus interstitial brachytherapy. He suggested that when combining external beam therapy with brachytherapy, interstitial techniques were preferable (45).
Patients who are sexually active must be encouraged to continue regular intercourse, but those who are not sexually active or for whom intercourse is temporarily too painful should be encouraged to use topical estrogen and a vaginal dilator, at least every second night. There is inadequate documentation of the adequacy of vaginal function after radiation therapy for vaginal cancer.
The reported overall 5-year survival rate for vaginal cancer is approximately 52%, which is more than 15% poorer than that for carcinoma of the cervix or vulva, and reflects the difficulties involved in treating this disease and the late stage at presentation (Table 14.4). Even for patients with stage I disease, the 5-year survival rate in combined series is only approximately 74%. In the 26th volume of the Annual Report on the Results of Treatment in Gynecological Cancer, the 5-year survival for 224 patients with vaginal cancer was as follows: stage I, 77.6%; stage II, 52.2%; stage III, 42.5%; stage IVA, 20.5%; stage IVB, 12.9% (2).
Table 14.4 Primary Vaginal Carcinoma: 5-Year Survival Rates
Better results have been reported from individual centers. In a report of 193 cases of primary vaginal squamous cell carcinoma from the M. D. Anderson Hospital in Houston, Frank et al. reported 5-year disease-specific survival (DSS) rates of 85% for 50 patients with FIGO stage I disease, 78% for 97 patients with stage II, and 58% for 46 patients with stages III-IV(5). Fiveyear DSS rates were 82% and 60% for patients with tumors ≥4 cm and >4 cm respectively (p = 0.0001). Kirkbridge et al. (4) from the Princess Margaret Hospital in Toronto reported on 138 patients with invasive vaginal carcinoma. The 5-year cause-specific survival rates by stage were 77% for stages I/II and 56% for stages III/IV. In a multivariate analysis, only tumor size and stage of disease were significant variables.
Most recurrences are in the pelvis (48), so improved radiation therapy, which may include chemoradiation and/or increasing experience with interstitial techniques, may improve the results. Chyle et al. (2) reported that salvage after first relapse was uncommon, with a 5-year survival rate of only 12%.
Because of the rarity of the disease, patients with vaginal cancer should be referred centrally to a limited number of tertiary referral units so that increasing experience can be gained in their management.
Approximately 10% of primary vaginal carcinomas are adenocarcinomas, and they affect a younger population of women, regardless of whether exposure to diethylstilbestrol (DES) in utero has occurred (49). Adenocarcinomas may arise in areas of vaginal adenosis, particularly in patients exposed to DES in utero, but they probably also arise in Wolffian rest elements, periurethral glands, and foci of endometriosis. They have been reported to have a worse prognosis than squamous carcinomas (3,41) although this has not been confirmed by other groups (2,19).
Vaginal adenosis is the sequestration of müllerian glandular epithelium into the vaginal mucosa during embryogenesis. It is thought to arise as a consequence of disrupted development and has been reported in non-DES-exposed women from infancy to old age (50). Intestinal metaplasia occasionally occurs in tissues of müllerian origin, and primary vaginal adenocarcinoma of intestinal type has been described (51). Secondary tumors from such sites as the colon, endometrium, or ovary should be considered when vaginal adenocarcinoma is diagnosed.
Diethylstilbestrol Exposure In Utero
In 1970, Herbst and Scully (52) initially reported on seven women 15 to 22 years of age with clear cell adenocarcinoma of the vagina, seen over a 4-year period. Subsequently, Herbst et al. (53) reported an association with maternal DES ingestion during pregnancy in six of these seven cases. A Registry for Research on Hormonal Transplacental Carcinogenesis was established by Herbst and Scully in 1971 to investigate the clinical, pathologic, and epidemiologic aspects of clear cell adenocarcinoma of the vagina and cervix occurring in women born after 1940 (i.e., during the years when DES was used to maintain high-risk pregnancies). Such high-risk situations included diabetic and twin pregnancies in women with a history of spontaneous abortion. The use of DES for pregnant patients was discontinued in the United States in 1971.
More than 500 cases of clear cell carcinoma of the vagina or cervix have been reported to the registry, although only approximately two-thirds of the completely investigated cases have a history of prenatal exposure to DES. In all instances, the mother was treated in the first half of the pregnancy (54). An additional 10% of the mothers received some unknown medication, but in 25% of the cases, there was no indication of maternal hormone ingestion.
These cancers become most frequent after the age of 14 years, and the peak age at diagnosis is 19 years. The oldest reported DES-exposed patient with vaginal clear cell carcinoma was 52 years of age. The estimated risk of clear cell adenocarcinoma in an exposed offspring is 1:1,000 or less. Approximately 70% of vaginal adenocarcinomas are stage I at diagnosis.
Although DES exposure in utero rarely leads to vaginal adenocarcinoma, vaginal adenosis occurs in approximately 45% of such patients, and approximately 25% of exposed women have structural changes to the cervix and vagina. Such changes include a transverse vaginal septum, a cervical collar, a cockscomb (a raised ridge, usually on the anterior cervix), or cervical hypoplasia. The occurrence of these abnormalities is related to the dosage of medication given and the time of first exposure, the risk being insignificant if administration is begun after the 22nd week.
Two types of cells have been described in vaginal adenosis and cervical ectropion: the mucinous cell, which resembles the endocervical epithelium, and the tuboendometrial cell.Robboy et al. (55) reported foci of atypical tuboendometrial epithelium in 16 of 20 (80%) cases of clear cell adenocarcinoma of the cervix or vagina. The foci were almost immediately adjacent to the tumor, and they suggested that atypical vaginal adenosis and atypical cervical ectropion of the tuboendometrial type may be precursors of clear cell adenocarcinoma. Sandberg and Christian (56) reported the appearance of cervicovaginal clear cell adenocarcinoma in only one of a genetically identical (monozygotic) pair of twins, simultaneously exposed to DES in utero. Benign teratologic changes were present in both twins. This discordance suggests that factors other than embryonic exposure to DESmay be operative in tumorigenesis.
Areas of vaginal adenosis and cervical ectropion are progressively covered with metaplastic squamous epithelium as the individual matures, and areas of adenosis may disappear completely and be replaced by normally glycogenated squamous epithelium. Structural abnormalities (e.g., cervical hoods) also tend to disappear progressively (57).
In addition to benign changes in the lower genital tract, a number of other abnormalities in the upper genital tract have been reported in DES-exposed female offspring. Kaufman et al. (58) reported abnormalities of the hysterosalpingogram in 185 of 267 (69%) exposed women. The most common abnormality was a T-shaped uterus, with or without a small cavity; less common abnormalities included constriction rings, uterine filling defects, synechiae, diverticulae, and uterus unicornis or bicornis. These abnormalities translate into an impaired reproductive experience for DES-exposed offspring, with an increased incidence of primary infertility, ectopic pregnancy, spontaneous abortion, and premature delivery (59).
It is recommended that a young woman exposed to DES in utero should be initially seen when she begins to menstruate, or at approximately 14 years of age. The most important aspects of the examination are careful inspection and palpation of the entire vagina and cervix, and cytologic sampling by direct scraping of the vagina and cervix. Colposcopy is not essential if clinical and cytologic evaluations are negative, but staining with half-strength Lugol's iodine delineates areas of adenosis.
In general, DES-related tumors may be treated in a way similar to that for squamous carcinomas, except that in these young patients, every effort should be made to preserve vaginal and ovarian function. For early-stage tumors, particularly those involving the upper vagina, the performance of radical hysterectomy, pelvic lymphadenectomy, vaginectomy, and replacement of the vagina with a split-thickness skin graft has been successful in a high percentage of cases. A combination of wide local excision, retroperitoneal lymphadenectomy, and local irradiation can be effective therapy for stage I tumors (60). Local surgical excision alone for small primary tumors is associated with a higher incidence of local and regional recurrence. Approximately 16% of patients with stage I disease have positive pelvic nodes (61). If radiation alone is used, a pretreatment staging laparotomy to allow pelvic lymphadenectomy and ovarian transposition may facilitate an optimal functional outcome. Freezing of embryos with a view to a subsequent surrogate pregnancy may be considered if ovarian function is in jeopardy and the patient has a suitable partner.
The overall 5-year survival rate for registry patients with clear cell carcinoma of the vagina, regardless of the mode of therapy, is 78%. The survival rate correlates well with stage of disease: 87% for patients with stage I disease, 76% for patients with stage II, and 30% for those with stage III (61).
Small Cell Carcinoma
Primary small cell carcinoma of the vagina is extremely rare, but as many as 5% of such tumors arise in extrapulmonary sites. In the female genital tract, such tumors arise most commonly in the cervix, followed by the ovary, endometrium, vagina, and vulva (62). As with other primary sites, small cell carcinoma of the vagina has a proclivity for distant failure and a poor prognosis. Management should be with concurrent chemoradiation.
Verrucous carcinomas of the vagina are rare, but their clinical and pathologic features are similar to those of their vulvar counterparts (63). They are large, warty tumors that are locally aggressive but have a minimal tendency to metastasize. Wide surgical excision of the tumor is the treatment of choice. Crowther et al. (63), in a literature review, reported a successful outcome in four of five patients with small lesions treated by wide excision. Similarly, for larger lesions, exenteration or vaginectomy was successful in seven of seven patients, but there were three postoperative deaths. Regional lymphadenectomy is not required, provided there is no suspicious lymphadenopathy. Radiation therapy has been implicated in the rapid transformation of such lesions to a more malignant tumor, and Crowther et al. (63) reported recurrence in all four patients treated with primary radiation therapy.
Malignant melanomas of the vagina are rare, with less than 250 reported cases to 2002 (64). The 26th Annual Report documents 13 cases (4%) among 324 cases of vaginal cancer (2). They presumably arise from melanocytes that are present in the vagina in 3% of normal women (65). The average age of the patients is 58 years, but vaginal melanomas have been reported from the third to the ninth decades of life (66). Almost all cases occur in white women (67).
Clinically, most patients present with vaginal bleeding, a vaginal mass, or vaginal discharge (66,67). The lesions most commonly arise in the distal part of the vagina, particularly on the anterior wall (67,68). They may be nonpigmented and are frequently ulcerated, making them easily confused with squamous carcinomas. Most are deeply invasive. Expressing the lesion in terms of Chung's level of invasion (as defined for vulvar melanomas ), Chung et al. (68) reported that 13 of 15 (87%) vaginal melanomas were at level IV. A more recent study from Houston reported that 20 of 26 cases (77%) were level IV lesions (70). Approximately 60% of Chung's cases exhibited spread of melanocytic cells into the adjacent epithelium, and in approximately 30% of the cases, the lateral spread was extensive (68).
Radical surgery has traditionally been the mainstay of treatment, and this has often involved anterior, posterior, or total pelvic exenteration, depending on the location of the lesion. Small upper vaginal lesions have been treated with radical hysterectomy, subtotal vaginectomy, and pelvic lymphadenectomy, whereas small distal vaginal lesions have been treated by partial vaginectomy, total or partial vulvectomy, and bilateral inguinofemoral lymphadenectomy.
More recently, conservative operations (e.g., wide local excision) have been used, followed frequently by pelvic radiotherapy (64,67,71,72,73), and there appears to be no significant benefit in terms of survival or disease-free interval for radical versus conservative surgery. The most important issue is to remove all gross disease whenever possible (73). Postoperative radiation therapy is effective for prevention of local recurrence, and high-dose fractions (>400 cGy) may be preferable to conventional or low-dose fractions (74). Chemotherapy (e.g., with methyl-CCNU [semustine] or dacarbazine) is disappointing, which is unfortunate because most patients die of distant metastases (70).
The overall prognosis for patients with vaginal melanoma is poor because most patients have deeply penetrating lesions at the time of diagnosis. Buchanan et al. (71) reviewed the literature and reported that only 18 of 197 patients (9.1%) survived for 5 years or longer. Six of the 18 patients were treated with radical operative procedures, 4 with radiation, 6 with wide local excision, and 1 with radiation plus wide excision. In one patient, the mode of therapy was unknown. Both reviews by Reid et al. (67) and Buchanan et al. (71) noted that size of the lesion was the best prognostic indicator. Among thirteen 5-year survivors who had their tumor size noted, 11 (84.6%) had lesions less than 3 cm in maximal diameter. Among ten 5-year survivors who had depth of invasion noted, only 2 (20%) had invasion of greater than 2 mm (71).
Adjuvant therapy with interferon alfa-2b has been shown to improve relapse-free and overall survival in patients with high-risk cutaneous melanomas (75), but there are as yet no data on this treatment for vaginal melanomas.
Once a recurrence is noted, prognosis is extremely poor, with a mean survival time of 8.5 months.
Vaginal sarcomas, such as fibrosarcomas and leiomyosarcomas, are rare tumors. They are usually bulky lesions and occur most commonly in the upper vagina. Tavassoli and Norris (76) reported 60 smooth muscle tumors of the vagina, only 5 of which recurred. All recurrences were seen in tumors more than 3 cm in diameter with moderate to marked cytologic atypia and more than five mitoses per 10 high-power fields. A review of vaginal leiomyosarcomas in 2000 revealed fewer than 70 cases reported in the English literature (77). The average age at presentation was 47 years, and the overall 5-year survival was 43%.
Surgical excision is the mainstay of treatment. If the lesion is well differentiated and the surgical margins are not involved, as is likely with tumors of low malignant potential, the likelihood of cure is good. For frankly malignant lesions, lymphatic and hematogenous dissemination is common. Adjuvant pelvic radiation is indicated for such tumors (78).
Rhabdomyosarcoma, a malignant tumor of the rhabdomyoblasts, is the most common soft tissue tumor in children. Rhabdomyosarcoma of the female genital tract is one of the most curable forms of the disease, but accounts for less than 4% of all pediatric rhabdomyosarcomas (79).
Sarcoma botryoides (embryonal rhabdomyosarcoma) is a highly malignant, grapelike tumor. The term botryoides comes from the Greek word botrys, which means “grapes,” and, grossly, the tumor usually appears as a polypoid mass extruding from the vagina. Microscopically, the characteristic feature is the presence of cross-striated rhabdomyoblasts (strap cells). More than 75% of vulvovaginal rhabdomyosarcomas are of the botryoid histological type (80). In the female genital tract, sarcoma botryoides is usually found in the vagina during infancy and early childhood, in the cervix during the reproductive years, and in the corpus uteri during the postmenopausal period.
Approaches to management have varied over the years, but radical surgery, including pelvic exenteration for primary management, has progressively been replaced by chemotherapy and more conservative surgery and radiation. The most influential group in this regard has been the Intergroup Rhabdomyosarcoma Study Group (IRSG), which conducted four consecutive trials commencing in 1972 (79,81). The initial study (1972-1978) involved radical surgical resection followed by chemotherapy, with or without radiotherapy. In the second study (1978-1984), multiagent chemotherapy was given first to reduce tumor size, with the objective of allowing more conservative surgery and organ preservation. Radiotherapy was reserved for those patients with residual disease after surgery. In both the third (1984-1991) and fourth (1991-1997) studies, chemotherapy was intensified further, with the aim of reducing further the extent of surgery necessary. After induction chemotherapy, clinical and radiological response was closely monitored and local treatment given only if residual tumor persisted (79).
The IRSG grouping classification is shown in Table 14.5. Chemotherapy was given according to the tumor group. Patients with group I, II, or III tumors received 12 months of treatment with either VAC (vincristine, actinomycin D, and cyclophosphamide), VAI (vincristine, actinomycin D, and ifosfamide) plus VAC, or VIE (vincristine, ifosfamide and etoposide) plus VAC. Patients with group IV tumors were randomized to receive VM (vincristine and melphelan) or ID (ifosfamide and doxyrubicin) given as initial therapy, followed by VAC plus RT.
As with surgery, radiotherapy has been more limited in recent years with a view to decreasing morbidity and preserving reproductive capability. Since 1990, only residual disease has been included in the brachytherapy-treated volume at the Gustave Roussy Institute in Paris (79). Ovarian transposition should be undertaken surgically prior to radiation therapy.Tumors larger than 40 mm after chemotherapy are not candidates for exclusive brachytherapy (79).
Table 14.5 IRSG Grouping Classification
In a review of the records of 82 cases of vaginal rhabdomyosarcoma treated on IRSG protocols I- IV, the estimated 5-year survival was over 85% for patients with locoregional tumors(81).
Other Vaginal Sarcomas
A variety of other malignant mesodermal tumors have been described to arise in the vagina, including malignant fibrous histiocytoma (82); angiosarcoma (83); and hemangiopericytoma (84). Petur and Young suggested that the latter tumors probably represent extrauterine endometrial stromal sarcomas (85).
Endodermal Sinus Tumor (Yolk Sac Tumor)
These rare germ cell tumors are occasionally found in extragonadal sites such as the vagina. Leverger et al. (86) reported 11 such cases from the Institut Gustave-Roussy. The average age of the patients was 10 months, and the presenting symptom was vaginal bleeding. Diagnosis was made by examination and biopsy with the patient under anesthesia. All children had high serum alpha fetoprotein levels. From 1977 to 1983, six of eight children were cured, with an average follow-up of 3 years. Treatment consisted of primary chemotherapy to reduce the tumor volume, followed by either partial colpectomy, radiation therapy, or both.
Carcinoma of the Urethra
Primary carcinoma of the female urethra is a rare malignancy, accounting for less than 0.1% of all female genital malignancies (87). The disease has been reported from the third to the ninth decades of life, with a median age of approximately 65 years. The most common presenting symptoms are urethral bleeding, hematuria, dysuria, urinary obstruction, and a mass at the introitus (88). Uncommon presenting symptoms include urinary incontinence, perineal pain, and dyspareunia.
Most tumors involve the anterior or distal urethra and may be confused with a urethral carbuncle or mucosal prolapse. Histologically, these distal lesions are usually squamous cell carcinomas. Tumors involving the posterior or proximal urethra are usually adenocarcinomas or transitional cell carcinomas. The relative frequency of the various histologic variants is shown in Table 14.6. Urethral carcinomas occasionally arise in a urethral diverticulum (90).
There is no FIGO staging for the disease, and several staging classifications have been suggested (89,91,94,95). The TNM staging system is shown in Table 14.7. Distal tumors spread to the lymph nodes of the groin, whereas proximal tumors spread to pelvic nodes, and treatment planning should take this into consideration. Bladder neck involvement is a common cause of local recurrence, and examination under anesthesia, endoscopic evaluation, and biopsy of the bladder neck should be undertaken as part of the pretreatment workup.
The treatment of urethral cancer must be individualized (87) and multimodal (88,93,96). For small distal lesions, the distal half of the urethra can be excised without loss of urinary continence. Bilateral inguinofemoral lymphadenectomy should be performed for all but the most superficial lesions involving the distal half of the urethra. Interstitial radiation may be satisfactory for more proximal early lesions.
Table 14.6 Histology of Urethral Carcinomas
The majority of patients present with advanced disease, and surgery alone is suboptimal. Radiation therapy is considered to be the treatment of choice, although it may cause complications such as urinary stricture, fistula, or total incontinence. Surgery may be used in conjunction with radiation for advanced lesions (91). Anterior exenteration and high-dose 192iridium (192Ir) intraoperative radiotherapy, followed several weeks later by external beam pelvic radiation, has been reported to give local control in four of six patients (67%), with a median follow-up of 21 months (range 12 to 47 months) (96). Four of the six patients were also treated with neoadjuvant or concomitant platinum-based chemotherapy. Klein et al. from Memorial Sloan-Kettering reported on five women who were treated with preoperative radiation followed by anterior exenteration combined with resection of the inferior pubic rami (97). Two died with distant metastases, and one died of surgical complications at 1 month. For medically unfit patients, external beam therapy followed by high dose-rate brachytherapy delivered with a remote afterloader, using a shielded vaginal applicator and modified urethral catheter, has been successfully used (98).
Experience with chemoradiation is limited to case reports, but the results appear to be favorable (99,100,101). Chemotherapeutic agents used have included cisplatin, 5-fluorouracil, and mitomycin-C. In view of the experience with other primary sites, this would seem to be an acceptable initial approach for locally advanced cases.
A Japanese case report described bladder-sparing urethrectomy with resection of the anterior vaginal wall and continent urinary diversion using the appendix (Mitrofanoff procedure) for a 2 cm × 2 cm distal urethral cancer invading the anterior vaginal wall. The patient was 77 years old and 4 years after surgery, she remained disease-free (102).
Bracken et al. (89) from the M. D. Anderson Hospital reported an overall 5-year survival rate of only 32% for 81 cases of carcinoma of the female urethra. When analyzed with respect to tumor size, 5-year survival rates were 60%, 46% and 13% for lesions <2, 2-4, and >4 cm respectively. Grigsby (92), from the Mallinckrodt Institute of Radiology in St. Louis, reported a 5-year survival rate of 42% for 44 cases. Stage distribution was as follows: T1 in 8, T2 in 5, T3 in 22, and T4 in 9. Treatment was with surgery in 7 cases, radiation therapy in 25 cases, and combined surgery and radiation therapy in 12. The severe complication rate was 29% for treatment with surgery, 24% for radiation therapy, and 8% for combined therapy. The most important clinical factors affecting prognosis were tumor size and histologic type—none of 13 women with adenocarcinomas was alive at 5 years, and only 1 of 10 women with tumors greater than 4 cm diameter was a 5-year survivor. The main cause of treatment failure is local recurrence.
Table 14.7 TNM Staging for Urethral Cancer
This rare tumor accounts for 0.2% of all melanomas. Di Marco et al. reported the Mayo Clinic experience of 11 cases (mean age 68 years) treated from 1950 to 1999 (103). Most patients presented with hematuria or a urethral mass. Four patients were treated by radical surgery, including anterior exenteration in two patients. Two (50%) of the four patients undergoing radical surgery recurred at 4 and 34 months, respectively. The remaining seven patients underwent local excision with partial urethrectomy. This group experienced urethral recurrence in five of the seven patients (71%). No patient received any adjuvant therapy. The authors suggested that radical urethrectomy with bladder preservation and a continent catheterizable stoma may be a more appropriate option. The catheterizable stoma could be constructed using an appendicovesicostomy or an ileovesicostomy for urinary diversion.