27 Carcinoma of the Vagina Higinia R. Cardenes

(1)

27 Carcinoma of the Vagina

Higinia R. Cardenes

H. R. Cardenes, MD, PhD

Clinical Associate Professor, Department of Radiation Oncol- ogy, RT 041, Indiana University School of Medicine, 535 Barn- hill Drive, Indianapolis, IN 46202, USA

Primary vaginal cancer is a rare entity and accounts for only 1–2% of all female genital neoplasias (Herbst et al. 1970), ranking next to last in fre- quency among gynecological malignancies. Most carcinomas found in the vagina represent direct extension or metastasis from other primary gynecological (cervix or vulva) and non-gynecological sites, most commonly breast and gastrointestinal tract. There are a number of controversies in terms of epidemiology, staging and diagnostic evaluation as well as management of vaginal cancer. Given the lack of prospective, randomized studies in patients with vaginal cancer, given its rarity, it is difficult to establish strong, evidence-based recommenda- tions. Therefore, the decisions regarding therapeutic options should be based on the best retrospective data and individual assessment using general principles derived from clinical experience in the management of cancer at other sites. Most of the available data refer to the treatment of primary invasive squamous cell carcinoma (SCC) of the vagina, since this represents the most common histology.

27.1 Anatomy

The vagina is a dilatable tubular structure averaging 7.5 cm in length, lined by non-keratinizing, strati- fied, squamous epithelium extending from the ves- tibule to the cervix of the uterus. It lies dorsal to the base of the bladder and urethra, and ventral to the rectum. The upper fourth of the posterior wall is separated from the rectum by a reflection of peritoneum, the pouch of Douglas. The vaginal wall is composed of three layers: the mucosa, muscularis and adventitia. Beneath the mucosa lies a submu- cosal layer of elastin and a double muscularis layer, highly vascularized with a rich innervation and lymphatic drainage. The muscularis layer is composed of smooth muscle fibers, arranged circularly on the inner portion and longitudinally in the outer por-

CONTENTS

27.1 Anatomy 657 27.2 Pathology 658 27.3 Natural History 658 27.4 Clinical Presentation 659 27.4.1 Diagnostic Work-up 659 27.4.2 Staging 659

27.5 Prognostic Factors Influencing Choice of Treatment 660

27.6 General Management: Treatment Options 662 27.6.1 RT Techniques 663

27.6.1.1 External Beam Radiotherapy 663

27.6.1.2 Low-Dose-Rate Intracavitary Brachytherapy 665 27.6.1.3 High-Dose-Rate Intracavitary Brachytherapy 666 27.6.1.4 Interstitial Brachytherapy 667

27.6.2 SCC: Treatment Options and Outcome by FIGO Stages 669

27.6.2.1 FIGO Stage 0: VAIN – CIS 669 27.6.2.2 Invasive SCC 670

27.6.3 Chemotherapy and Radiation 670 27.6.4 Patterns of Failure in SCC 674 27.6.4.1 Potential RT-Related Factors Influencing Outcome 674

27.7 Clear Cell Carcinoma of the Vagina 675 27.8 Salvage Therapy 675

27.8.1 Surgical Considerations 676 27.8.2 RT Considerations 676

27.9 Treatment Complications and their Management 678

27.10 Palliative RT 680 27.11 Chemotherapy in

Advanced-Recurrent Vaginal Cancer 682 References 682

(2)

tion. The adventitia is a thin, outer connective tissue layer that merges with that of adjacent organs.

The proximal vagina is supplied by the vaginal artery branch from the uterine or cervical branch of the uterine artery. It runs along the lateral wall of the vagina and anastomoses with the inferior vesical and middle rectal arteries from the surrounding vis- cera (Sedlis and Robboy 1987). The accompanying venous plexus, running parallel to the arteries, ulti- mately drains to the internal iliac vein. The lumbar plexus and pudendal nerve, with branches from the sacral roots 2 to 4, provide innervation to the vaginal vault.

The lymphatic drainage of the vagina is complex, consisting of an extensive inter-communicating network. Fine lymphatic vessels coursing through the submucosa and muscularis coalesce into small trunks running laterally along the walls of the vagina. The upper anterior vagina drains along cervical channels to the interiliac chain; the posterior vagina drains into the inferior gluteal, presacral, and anorectal nodes. The distal vagina lymphatics drain into the inguinal and femoral nodes and from there to the pelvic nodes. Lymphatic flow from lesions in the mid-vagina may drain either way (Plentl and Friedman 1971). However, because of the presence of inter-communicating lymphatics along the ter- minal branches of the vaginal artery and near the vaginal wall, the external iliac nodes are at high risk, even in lesions of the lower third of the vagina. Such a complex lymphatic drainage pattern has significant implications for therapeutic planning. There- fore, bilateral pelvic nodes should be considered at risk in any invasive vaginal carcinoma, and bilateral groin nodes considered at risk in those lesions involving the distal third of the vagina.

27.2 Pathology

SCC represents about 80–90% of primary vaginal cancers (Zaino et al. 2002). These tumors occur in older women and are most often located in the upper, posterior wall of the vagina. Histologically, keratinizing, non-keratinizing, basaloid, warty and verrucous variants have been described. Tumors may also be graded, moderately or poorly differentiated, most cases being moderately differentiated and non-keratinizing. Vaginal intraepithelial neoplasia (VAIN) is a precursor of SCC and is graded from I to III, depending on the degree of nuclear atypia

and crowding, and the proportion of the epithelium involved. The true incidence of VAIN and its rate of progression to invasive carcinoma are unknown, ranging in several series from 9% to 28% (Aho et al.

1991; Brinton et al. 1990; Hoffman et al. 1991).

Clear-cell adenocarcinoma (CCA) is associated with intrauterine diethylstilbestrol (DES) exposure (Antonioli and Burke 1975; Kauffman et al. 1982; Maassen et al. 1993; Robboy et al. 1977, 1982, 1984). These tumors have a predilection for the upper third of the vagina and the exocervix. Most are exophytic and superficially invasive (Herbst et al. 1974). Other adenocarcinomas that could involve the vagina include: mucinous type (Ebrahim et al. 2001; Hiroi et al. 2001; Munkarah et al. 1994), endometrioid adenocarcinoma – usually arising with endometriosis (Haskel et al. 1989), meso- nephric (Hinchey et al. 1983) and papillary serous adenocarcinoma (Riva et al. 1997).

Creasman et al. (1998) published in 1998 the National Cancer Data Base (NCDB) report on 4885 patients with primary diagnosis of vaginal cancer registered from 1985 to 1994. Approximately 92%

of the patients were diagnosed with in-situ or invasive SCC or adenocarcinomas; 4% with melanomas;

3% with sarcomas; and 1% with other or unspeci- fied types of cancer. In the NCDB report, invasive carcinomas accounted for 72% of the carcinoma cases, or 66% of all vaginal cancers. In-situ carcinomas accounted for 28%; SCC represented 79% of invasive vaginal carcinomas; and adenocarcinomas represented 14% (Creasman et al. 1998).

27.3

Natural History

The majority (57–83%) of vaginal primaries occur in the upper third or at the apex of the vault, most commonly in the posterior wall; the lower third may be involved in as many as 31% of patients (Gallup et al. 1987; Rubin et al. 1985; Stock et al. 1995). Lesions confined to the middle third are uncommon. Vaginal tumors may spread along the vaginal walls to involve the cervix or the vulva. A lesion on the anterior wall may infiltrate the vesico–vaginal septum and/or the urethra; those on the posterior wall may eventually involve the recto–vaginal septum and, subsequently, infiltrate the rectal mucosa. Vaginal cancer may invade the parametrium and paracolpal tissues, extending into the obturator fossa, cardinal ligaments, lateral pelvic walls and the uterosacral ligaments.

(3)

The issue of regional nodal metastasis, both the incidence of occult nodal disease and the anatomic pathways of lymphatic spread, are somewhat controversial. The incidence of positive pelvic nodes at diagnosis varies with the stage and location of the lesion. There does seem to be a significant risk of nodal metastasis for patients with disease beyond stage I. Although data on staging lymphadenectomy are sparse, two studies reported a significant incidence of nodal disease in early stage vaginal carcinoma. In Al-Kurdi’s series (Al-Kurdi and Monaghan 1981), the incidence of pelvic nodal metastasis was 14% and 32% for stages I and II, respectively, whereas in Davis’s series (Davis et al. 1991) the incidence was 6% and 26% for stages I and II, respectively. The incidence is expected to be higher for stage III, although no substantial data are available. Involvement of inguinal nodes is most common when the lesion is located in the lower third of the vagina.

Distant metastasis may occur, primarily in patients with advanced disease at presentation, or in those who experienced tumor recurrence after primary therapy. In Perez series (Perez et al. 1988), the incidence of distant metastasis was 16% in stage I, 31% in stage IIA, 46% in stage IIB, 62% in stage III and 50% in stage IV.

27.4

Clinical Presentation

VAIN is most often asymptomatic (Lenehan et al.

1986) and is usually detected by means of cytology.

In patients with invasive disease, irregular vaginal bleeding and/or discharge (often post-coital) is the most common presenting symptom, followed by vaginal discharge and dysuria. Pelvic pain is a relatively late symptom, generally related to tumor extent beyond the vagina (Gallup et al. 1987; Rubin et al. 1985; Tjalma et al. 2001).

27.4.1

Diagnostic Work-up

In general, in patients with suspected vaginal malignancy, thorough physical exam with detailed speculum inspection, digital palpation, colposcopic and cytological evaluation and biopsy constitute the most effective procedure for diagnosing primary, metastatic or recurrent carcinoma of the vagina. In

symptomatic patients, biopsy of any abnormal exophytic or endophytic lesion noted at the time of the exam is indicated. Examination under anesthesia is recommended for the thoroughness of evaluation of all of the vaginal walls and local extent of the disease, primarily if the patient is in great discomfort because of advanced disease, in order to obtain a biopsy. Biopsies of the cervix, if present, are recommended to rule out primary cervical tumor. It is important that the speculum is slowly withdrawn from the vaginal fornix so that the total vaginal mucosa may be visualized.

The patient with a history of pre-invasive or invasive carcinoma of the cervix found to have abnormal cytology following prior hysterectomy or radiotherapy (RT) should be offered colposcopy with application of acetic acid to the entire vault, followed by biopsies as indicated by areas of white epithelium, mosaicism, punctation or atypical vascularity. It can be very helpful for the menopausal patient or the patient previously irradiated to use a short course of topical estrogen (Premarin) into the vault once or twice a week for 1 month prior to the colposcopy, in order to foster epithelial maturation. Another method of identifying the area(s) most in need of biopsy would be, after application of the acetic acid, to place half-strength Schiller’s iodine to determine whether the Schiller-positive (non-staining) areas correspond with the involved areas identified after acetic acid application.

27.4.2 Staging

At present, primary malignancies of the vagina are all staged clinically. In addition to a complete history and physical examination, routine laboratory evaluations including complete blood cell count (CBC) with differential and platelets and assessment of renal and hepatic function should be undertaken.

In order to determine the extent of disease, the following tests are allowed according to the Interna- tional Federation of Gynecology and Obstetrics (FIGO) criteria: chest X-ray, a thorough bimanual and recto–vaginal exam, cystoscopy, proctoscopy and intravenous pyelogram. If the patient is in significant discomfort, the exam should be con- ducted under anesthesia, preferably by a radiation oncologist and gynecological oncologist. However, it can be difficult even for the experienced exam- iner to differentiate between disease confined to the mucosa (stage I) and disease spread to the submu-

(4)

cosa (stage II) (Ball and Berman 1982; Rubin et al. 1985). Cystoscopy and proctoscopy are generally performed on patients with symptoms or clinical findings indicative of bladder or rectal infiltration, respectively.

Pelvic computed tomography (CT) scan is generally performed to evaluate inguino-femoral and/or pelvic lymph nodes, as well as extent of local disease. In patients with vaginal melanoma or sarcoma, chest, abdomen, and pelvis CT scans are often part of the work-up. Magnetic resonance imaging (MRI) has emerged as a potentially important imaging modality in the evaluation of vaginal cancers, pre- dominantly the T1-weighted with contrast and T2- weighted images. An additional role of MRI is dif- ferentiation of tumor from fibrotic tissue in patients with suspected recurrent vaginal carcinoma (Chang et al. 1988).

The two commonly used staging systems for carcinoma of the vagina are the FIGO (Table 27.1a) (Pecorelli et al. 2000) and the American Joint Commission on Cancer (TNM) (Table 27.1b) (AJCC 2002) classifications. According to the FIGO guidelines, cases should be classified as vaginal carcinomas only when “the primary site of the growth is in the vagina”. A tumor of the vagina that involves the cervix or vulva should be classified as a primary cervical or vulvar cancer, respectively. Additionally, in the setting of a prior gynecological malignancy, a neoplasm would be classified as primary carcinoma of the vagina if the current vaginal tumor occurred 5 years or more after the initial cancer diagnosis and if there is no other clinical evidence of recurrence of the initial gynecological lesion (Zaino et al. 2002). It may be difficult or impossible histologically to dis- tinguish a primary vaginal SCC from recurrent cervical or vulvar disease. In this setting, it is unclear whether the vaginal lesion represents a new carcinoma of the vagina, recurrent cervical cancer, or a human papilloma-virus (HPV)-related field effect in these patients.

Perez et al. (1973) proposed in 1973 that FIGO stage-II vaginal cancer should be subdivided into stage IIa (tumor infiltrating the subvaginal tissues but not extending into the parametrium) and stage IIb (tumor infiltrating the parametrium but not extending to pelvic side walls). However, most authors do not use this classification, and there is limited published data to support the prognostic significance of this subclassification (Perez et al. 1988;

Prempree and Amomman 1985). In addition, FIGO does not assign a specific stage for those patients with inguino-femoral lymphadenopathy. Some

authors assign these patients to stage III, whereas others consider them stage IVB. In the AJCC staging system, patients with T1–3 and positive nodes (pelvic or inguinal) are assigned to stage III (AJCC 2002).

27.5

Prognostic Factors Influencing Choice of Treatment

As with most primaries, stage of disease is the dom- inant prognostic factor in terms of ultimate outcome (Chu and Beechinor 1984; Chyle et al. 1996;

Dancuart et al. 1988; Delclos 1984; Dixit et al.

1993; Eddy et al. 1991; Kirkbride et al. 1995; Lee et al. 1994; Leung and Sexton 1993; MacNaught et al. 1986; Perez et al. 1988, 1999; Peters et al.

1985). In Perez series, including 165 patients with primary vaginal carcinomas treated with definitive RT, the 10-year actuarial disease-free survival (DFS) was 94% for stage 0, 75% for stage I, 55% for stage IIA, 43% for stage IIB, 32% for stage III and 0% for those with stage IV (Perez et al. 1988).

The impact of lesion location has been controversial. Several authors (Ali et al. 1996; Chyle et al.

1996; Kucera and Vavra 1991; Tarraza et al. 1991;

Urbanski et al. 1996) have shown better survival and decreased recurrence rates for patients with cancers involving the proximal half of the vagina when compared with those in the distal half, or those involving the entire length of the vagina. Chyle et al. (1996) observed a 17% pelvic relapse in patients with upper vagina lesions, 36% with mid- or lower vagina tumors and 42% with whole vagina involvement. In addition, lesions of the posterior wall have worse prognosis than those involving other vaginal walls (10-year recurrence rates of 32% and 19%, respectively), which probably reflects the greater difficulty of performing adequate brachytherapy procedures in this location.

The prognostic importance of lesion size has been controversial, with an adverse impact noted by Tjalma et al. (2001) and Chyle et al. (1996) con- trary to the findings of Perez et al. (1983). In the Chyle et al. (1996) series, lesions measuring less than 5 cm in maximum diameter had a 20% 10-year local recurrence rate, compared with 40% for those lesions larger than 5 cm. Similarly, in the Princess Margaret Hospital (PMH) experience, tumors larger than 4 cm in diameter fared significantly worse than smaller lesions (Kirkbride et al. 1995). In the Perez

(5)

series (Perez et al. 1999), stage was an important predictor of pelvic tumor control and 5-year DFS, but the size of the tumor in stage-I patients was not a significant prognostic factor. However, in stage- IIA disease, lower pelvic tumor control and survival were noted with tumors larger than 4 cm. In stages IIB–III, tumor size was not a significant prognostic factor, probably related to the difficulty in assessing size, and the fact that higher doses of RT were delivered for larger tumors. Stock et al. (1995) reported that disease volume, a likely surrogate for stage or lesion size, adversely impacted survival, as well as local control.

Age was a significant prognostic factor in Urban- ski’s series (Urbanski et al. 1996), with 5-year survival of 63.2% for patients below the age of 60 years, compared with 25% for those over 60 years of age (P<0.001). Similar findings were reported by Eddy et al. (1991) and in the NCDB of the American Col- lege of Surgeons (Creasman et al. 1998) with better survival in younger patients (90% versus 30%, respectively). However, most of these series do not correct for death secondary to intercurrent disease in the elderly population. No statistical significance of age to survival was found in the series of Dixit et al. (1993) and Perez et al. (1999).

Table 27.1a. FIGO staging system for carcinoma of the vagina Stage Description

Stage 0 Carcinoma in situ, intraepithelial neoplasia grade III Stage I Limited to the vaginal wall

Stage II Involvement of the subvaginal tissue but without extension to the pelvic side wall Stage III Extension to the pelvic side wall

Stage IV Extension beyond the true pelvis or involvement of the bladder or rectal mucosa.

Bullous edema as such does not permit a case to be allotted to stage IV Stage IVa Spread to adjacent organs and/or direct extension beyond the true pelvis Stage IVb Spread to distant organs

From Pecorelli et al. (2000)

Table 27.1b. American Joint Commission on Cancer (AJCC) staging of vaginal cancer Primary tumor(T)

Tx Primary tumor cannot be assessed T0 No evidence of primary tumor Tis/0 Carcinoma in situ

T1/I Tumor confined to the vagina

T2/II Tumor invades paravaginal tissues but not to the pelvic wall T3/III Tumor extends to the pelvic wall

T4/IVA Tumor invades mucosa of the bladder or rectum and/or extends beyond the pelvis (bullous edema is not sufficient to classify a tumor as T4)

Regional lymph nodes (N)

Nx Regional lymph nodes cannot be assessed N0 No regional lymph nodes

N1/IVB Pelvic or inguinal lymph node metastasis Distant metastasis (M)

Mx Distant metastasis cannot be assessed M0 No distant metastasis

M1/IVB Distant metastasis AJCC stage groupings Stage 0 Tis N0 M0 Stage I T1 N0 M0 Stage II T2 N0 M0

Stage III T1–3 N1 M0, T3 N0 M0 Stage IVa T4, any N, M0

Stage IVb Any T, any N, M1

From American Joint Committee on Cancer (2002)

(6)

With regard to the histological type and grade, several series (Kirkbride et al. 1995; Kucera and Vavra 1991; Urbanski et al. 1996) have shown the histological grade to be an independent, significant predictor of survival. However, the histology of the tumor (SCC versus other) has not been found to be a prognostic factor for NED survival among the patients with invasive tumors. Chyle et al. (1996) noted a higher incidence of local recurrences in adenocarcinomas when compared with SCC (52% and 20%, respectively, at 10 years), higher distant metastasis rate (48% and 10%, respectively) and lower 10- year survival (20% versus 50%, respectively). An increased propensity for distant metastases to the lung and supraclavicular nodes has been reported in patients with CCA (Robboy et al. 1974). Patients with vaginal melanoma (Reid et al. 1989) and malig- nant mesenchymal tumors (Tavassoli and Norris 1979) have a worse prognosis than patients with SCC due to a higher propensity for development of local recurrence and distant metastases.

Lymph node metastasis at diagnosis portends a poor prognosis. However, this has not been exten- sively evaluated in vaginal cancer. The only report of outcome based on lymph node status is by Pingley et al. (2000). They reported a 56% DFS for patients without lymph node involvement and 33% for those with metastatic lymphadenopathy at presentation.

27.6

General Management: Treatment Options

Due to its rarity, data concerning the natural history, prognostic factors and treatment of vaginal carcinoma derive from small, retrospective studies. Most of the currently available literature in terms of radiotherapeutic and surgical techniques refer to primary SCC of the vagina. In general, SCC of the vagina has been treated by means of RT. However, several surgical series have reported acceptable to excellent outcomes in well-selected patients, with survival rates after radical surgery for stage-I disease ranging from 75% to 100%

(Ball and Berman 1982; Creasman et al. 1998;

Davis et al. 1991; Rubin et al. 1985; Tjalma et al.

2001), although few studies directly compared the two treatment modalities. Cases in which surgery may be the preferred treatment include selected stage I–II patients with lesions at the apex and upper-third of the posterior or lateral vagina that could be approached with radical hysterectomy,

upper vaginectomy, and pelvic lymphadenectomy providing adequate margins (Ball and Berman 1982; Davis et al. 1991; Gallup et al. 1987; Rubin et al. 1985; Stock et al. 1995) and very superficial lesions that may be removed with wide local excision. If the margins are found to be close or positive after resection, adjuvant RT is recommended. How- ever, for lesions at other sites and in those cases requiring more extensive resection, definitive RT is the treatment of choice in order to maximize cure and improve quality of life; generally, those patients with isolated central failures are offered exenteration (Rubin et al. 1985).

Furthermore, most patients are elderly, and a radical surgical approach is often not feasible. Local excision and partial and complete vaginectomy have given way to a more individualized approach that takes into consideration the patient’s age, the extent of the lesion, and whether it is localized or multi- centric. In younger patients with early stage disease, treatment can also depend on the desire to preserve a functional vagina as well as ovarian function. Rad- ical surgery in the past precluded vaginal function, but this situation has been improved significantly by the use of split-thickness grafts, intestinal segments and myocutaneous flap reconstruction (Magrina and Basterson 1981). Creasman noted superior survival in those undergoing surgery (Creasman et al. 1998). However, he and Tjalma (Tjalma et al.

2001) recognized that there may be bias in surgical series. Younger, healthier patients with better per- formance status are more likely to be offered radical surgery, whereas older patients with multiple co- morbid medical conditions are offered RT.

In most patients, the primary treatment modality is RT, as noted by the Society of Gynecologist Oncologists in practice guidelines published in 1998 (Creasman et al. 1998). RT provides excellent tumor control in early and superficial lesions, with satisfactory functional results. This makes it imper- ative that RT techniques yielding optimal tumor control and functional results are utilized. Despite the acceptance of RT as the treatment of choice for this disease, in particular for patients with lesions involving the mid-third of the vagina or stage II and greater, the optimal therapy for each stage is not well-defined in the literature. Intracavitary and interstitial irradiation is used in small superficial stage-I disease. A combination of external beam RT (EBRT), intracavitary brachytherapy (ICB) and/or interstitial brachytherapy (ITB) with or without chemotherapy is used in more extensive stage-I and stages II–IV disease.

(7)

27.6.1

RT Techniques

Radiation prescription and technique varies according to stage, disease bulk and anatomic location. In recent years, several series have reported improved outcomes with techniques emphasizing the higher radiation doses that can be achieved when all or part of the therapy is accomplished with interstitial technique (Leung and Sexton 1993; Perez et al. 1999;

Pingley et al. 2000; Puthawala et al. 1983).

27.6.1.1

External Beam Radiotherapy

EBRT is advisable in patients with deeply infiltrating or poorly differentiated stage-I lesions, and in all

patients with stage II–IVA disease. The treatment is generally delivered using opposed anterior and posterior fields (AP/PA). The pelvis receives between 20 Gy and 45 Gy, depending on the stage of the disease. High-energy photons (>10 MV) are usually preferred. Treatment portals cover at least the true pelvis with a 1.5- to 2-cm margin beyond the pelvic rim. Superiorly, the field extends to either L4–L5 or L5–S1 to cover the pelvic lymph nodes up to the common iliacs, and extends distally to the introitus to include the entire vagina. Lateral fields, if used, should extend anteriorly to adequately include the external iliac nodes, anterior to the pubic symphy- sis, and at least to the junction of S2–S3 posteriorly (Fig. 27.1).

In patients with tumors involving the middle and lower vagina with clinically negative groins, the bilateral inguino-femoral lymph node regions

Fig. 27.1a–c. Invasive Squamous cell carcinoma of the proximal vagina with Squamous-cell carcinoma in-situ in the mid-third of the vaginal squamous cell carcinoma in situ, without involvement of inguino-femoral nodes. Digital reconstructed radiographs (DRRs): a AP/PA whole pelvis ﬁ elds.

b Right/left lateral whole pelvis ﬁ elds, intended to treat the entire length of the vagina. c Axial, sagittal and coronal isodose distributions

a b

c

(8)

should be treated electively to 45–50 Gy. Planning CT is recommended to adequately determine the depth of the inguino-femoral nodes. A number of techniques have been used to treat the areas at risk without over-treating the femoral necks. Some of the most commonly used techniques include the use of unequal loading (2:1, AP/PA), a combination of low- and high-energy photons (4–6 MV, AP, and 15–18 MV, PA), or equally weighted beams with a transmission block in the central AP field, utilizing small AP photon or electron beams to deliver a daily boost to the inguino-femoral nodes. A technique has been developed and implemented at Indiana

University which uses a narrow PA field to treat the pelvis, and a wider AP field encompassing the pelvis and inguino-femoral nodes, with daily AP photon boost to the inguinal nodes being delivered using the asymmetric collimator jaws (Dittmer and Randall 2001). Advantages of this technique include simplicity of set-up and treatment (single isocenter, no need for transmission block), dose homogeneity, reduced dose to the femoral necks, low potential risk of nodal underdose, and elimination of dosimetric difficulties inherent in electron boosts (Figs. 27.2).

In patients with clinically palpable inguinal nodes, additional doses of 15–20 Gy (calculated at a depth

Fig. 27.2a,b. Vaginal cancer with distal third vaginal involvement, squamous cell carcinoma. Technique for pelvic and inguino-femoral nodal irradiation. a Digi- tal reconstructed radiographs (DRRs). AP ﬁ eld intended to treat pelvis and groins;

PA ﬁ eld intended to treat the pelvis only, in order to decrease the dose to the femoral heads; small AP photon ﬁ eld for additional daily boost to the inguino-femoral nodes.

b Axial, sagittal and coronal isodose dis- tributions

a

b

(9)

determined by CT scan) are necessary with reduced portals. This is generally achieved using low-energy photons or electron beam (12–18 MeV).

For patients with positive pelvic nodes, or those patients with advanced disease not amenable to interstitial implant, additional boost to the areas of gross disease, as defined by CT scan, should be given using conformal therapy to deliver a total dose between 65 Gy and 70 Gy with high-energy photons (Fig. 27.3).

27.6.1.2

Low-Dose-Rate Intracavitary Brachytherapy

VAIN and small T1 lesions with less than 0.5 cm depth can be adequately treated with ICB alone. Low-dose- rate ICB (LDR-ICB) is performed using vaginal cylinders such as Burnett, Bloedorn, Delclos (Delclos et al. 1980) or MIRALVA (Perez et al. 1990) loaded with cesium-137 (¹³⁷Cs) radioactive sources. Delclos afterloading vaginal cylinders have a central hollow metallic cylinder in which the sources are placed, and plastic rings of varying diameter (2.5–4 cm), 2.5 cm in length, are inserted over the cylinder. Domed cylinders are used to irradiate the vaginal cuff homo- geneously, when indicated. Delclos (Delclos et al.

1980) recommended a short ¹³⁷Cs source to be used at the top to obtain a uniform dose around the dome, because a lower dose occurs at the end of the linear cesium source. Some cylinders have a lead shielding to protect selected portions of the vagina, the bladder and/or the rectum. The largest possible diameter that can be comfortably accommodated by the patient should be used to improve the ratio of mucosa to tumor dose and eliminate vaginal rogations (Fig. 27.4). The cylinders can be mounted in the vaginal component of an intrauterine tandem or along the stem of a dome cylinder. Before the cylinders are mounted over the vaginal component of an intrauterine tandem, this is inserted into the uterus (when present) and the cylinders are fitted along the protruding tandem. To minimize the rotation of the tandem, a flat, round flange with keel is placed below the last cylinder and is kept in position with some packing if required. In general, the vulva is sutured-closed with proline or silk for the duration of the implant in order to secure the position of the applicators (Fig. 27.4c, d).

In patients with upper vagina lesions with less than 0.5 cm depth of invasion, vaginal colpostats alone (after hysterectomy) or in combination with intrauterine tandem, loaded with ¹³⁷Cs sources similar to that used in treatment of cervical cancer, can be used to treat the proximal vagina to a minimum

dose of 65–70 Gy, estimated to 0.5 cm depth, including the contribution of EBRT if given. When indicated, the remainder of the vagina can be treated by performing a subsequent implant using vaginal cylinders generally 60Gy total dose to the vaginal surface will be delivered including the contribution of the EBRT and the intracavitary implant. It is important to avoid the placement of a protruding source over the vulva, with the subsequent increased risk of complications. The use of LDR remote control afterloading technology allows the reduction of radiation exposure to hospital personnel and optimization of the isodose distribution. When appropriate dose specification points are chosen, a very uniform dose distribution over the entire length of the vagina can be obtained.

Perez et al. (1990; Slessinger et al. 1992) designed and constructed a vaginal applicator, MIRALVA, which incorporates two ovoid sources and a central tandem that can be used to treat the entire vagina (alone, or in combination with the uterine cervix). The applicator has vaginal apex caps and additional cylinder sleeves to allow for increased dimensions (Fig. 27.5a). A tandem in the uterus can be used if clinically indicated. The dosimetry and dose specifications for this applicator have been published (Slessinger et al. 1992), showing that the applicator delivers 1.1–1.2 Gy/h to the vaginal apex and 0.95–1 Gy/h to the distal vaginal surface when loaded with 20 mgRaEq ¹³⁷Cs tubes in each ovoid

Fig. 27.3. Locoregionally advanced vaginal cancer. External irradiation – beam arrangement including initial whole pelvis and inguino-femoral ﬁ elds and three-dimensional conformal radiotherapy boost. In magenta, the isodose distribution cor- responds to the 65 Gy isodose line

(10)

a b

c d

Fig. 27.4a–d. AP (a) and lateral (b) views of vaginal cylinders only. AP (c) and lateral (d) views of intrauterine tandem and vaginal cylinders

and 10–10–20 mgRaEq tubes in the vaginal cylinder (Fig. 27.5b). When the tandem and vaginal cylinder are used, the strength of the sources in the ovoids should always be 15 mgRaEq. The vaginal cylinder or uterine tandem never carries an active source at the level of the ovoids to prevent excessive doses to the bladder or rectum.

27.6.1.3

High-Dose-Rate Intracavitary Brachytherapy

High-dose-rate ICB (HDR-ICB) is typically performed with a 10 Ci single iridium-192 (¹⁹²Ir) source (Micro-Selectron HDR, Nucletron). The applicators are similar to those described for LDR, consisting of vaginal cylinders of 2.5–4 cm in diameter.

Little information regarding HDR-ICB in the treatment of primary carcinoma of the vagina is available (Nanavati et al. 1993; Stock et al. 1992).

Few patients have been treated, follow-up is short, publication bias is likely and there is no agreement on treatment regimen. With HDR, there is a need to adjust the total dose, and additional fractionation is necessary, compared with LDR brachytherapy, because of the biologically equivalent dose considerations. Generally, the number of insertions ranges from 1 to 6 (median 3), with the dose per fraction ranging from 300 cGy to 800 cGy (median 700 cGy).

Nanavati et al. (1993) reported 13 patients with primary vaginal cancer (5 St I, 4 St IIA and 4 St IIB) treated with external beam RT (45 Gy) and HDR- ICB (20–28 Gy in three to four fractions, calculated at 0.5 cm from the surface of the applicator). All 13

(11)

Fig. 27.5. a MIRALVA structure. b Coronal, sagittal and axial isodose distributions for the applicator loaded with Cesium- 137 radioactive sources dose distribution, A’ B’ without and C’

D’ with intrauterine tandem. Reprinted with permission from Perez et al. (1990) and Slessinger et al. (1992)

a

b

patients had a complete response, and local control was achieved in 92% of the patients with a median follow-up of only 2.6 years (range 0.7–5.2 years).

Originally, the planned dose was 2100 cGy in three fractions, 700 cGy/fraction, but because of reports of decreased complications with increased fractionation, the planned dose was changed to 2000 cGy in four equal fractions. They did not observe any acute or chronic intestinal or bladder grade 3 or 4 toxicity. However, moderate to severe vaginal stenosis occurred in 46% of the patients. The authors recognize that “late-occurring toxicity could be missed at a medium follow-up of 2.6 years.” In a recent report, Mock et al. (2003) showed similar outcome and toxicity rate, stage by stage, in 86 patients with primary vaginal cancer treated with a variety of external beam and HDR brachytherapy.

Many aspects remain unknown or not well- understood in the use of HDR-ICB. These include the radiobiological equivalency of HDR to LDR, fractionation schedule, total dose, specification of dose prescription, and how to combine HDR with EBRT and/or LDR brachytherapy. In addition, optimization approaches and methods of dose calculation, such as the inclusion of anisotropic corrections, are not well-described in the sparse literature available to date (Gore et al. 1995; Li et al. 1998). Li et al. (1998)

have shown that when optimized dose distribution at a distance from the cylinder surface is calculated using an accurate dose calculation model, the vaginal mucosa dose becomes significantly higher than calculated, and therefore should be carefully moni- tored. These factors could result in an increased incidence of severe complications, such as vaginal necrosis, and recto–vaginal or vesico-vaginal fistulas (Rutkowski et al. 2002; Tyree et al. 2002).

In the opinion of the authors, until further data are available with longer follow-up, as well as a better understanding of the physical and radiobiological principles involved in the HDR-ICB, this should not be routinely used in the management of primary vaginal carcinoma. We strongly encourage the continued use of LDR-ICB, given its excellent results and widely documented long-term outcome and complications (Chyle et al. 1996; Kirkbride et al. 1995; Perez et al. 1988, 1999).

27.6.1.4

Interstitial Brachytherapy

ITB is an important component in the treatment of more advanced primary vaginal carcinomas, typically in combination with EBRT and/or ICB. In

(12)

the first place, a careful definition of the “target volume”, this is gross tumor volume (based on clinical, radiological and operative findings) and a margin of adjoining normal tissue, is required.

Other considerations include whether a permanent (¹⁹⁸Gold or ¹²⁵Iodine) or temporary implant (¹⁹²Ir) is optimal, the geometry of the implant (e.g., single or double plane or volume implant), source distribution, dose rate and total dose, based on tumor size, location, local extent and proximity of normal structures (Hilaris et al. 1987). The principal advantages of temporary implants are readily controlled distribution of the radioactive sources and easier modifi- cation of the dose distribution. The main advantages of a permanent seed implant include relative safety/

simplicity, easy applicability, cost-effectiveness and ability, in most cases, to be performed under local anesthesia. As a general rule, temporary implants are more commonly used in the curative treatment of larger gynecological malignancies, whereas permanent implants are usually performed for smaller volume disease.

The number and strength of the radioactive sources and their intended distribution within the target volume are determined pre-operatively, making use of available guidelines such as nomo- grams, tables and computer-assisted optimization techniques. Following this, it is necessary to spec- ify an approximate dose rate to the target volume, which requires careful localization of the sources and computer calculation of the three-dimensional radiation dose distribution. Finally, a dose prescription, based on the treatment volume, tumor sensitivity, dose rate, prior treatments and tolerance of normal surrounding tissues, is required (Hilaris et al. 1987).

When performing an interstitial procedure, free- hand implants or template systems designed to assist in pre-planning and to guide and secure the position of the needles in the target volume can be employed.

Popular commercially available templates include the Syed-Neblett device (SNIT) (Alpha Omega Ser- vices, Bellflower, CA) (Syed et al. 1986), the modified Syed-Neblett (Disaia et al. 1990) and the “MUPIT”

(Martinez Universal Perineal Interstitial Template) (Martinez et al. 1984). All rely on pelvic examination to help guide the location and depth of needle placement.

The Syed-Neblett device consists of two identical superimposed Lucite plates, each about 1 cm thick, held together by six screws. Both plates are drilled in an identical pattern of predetermined needle positions that can be afterloaded with iridium-192

(¹⁹²Ir) sources (Syed et al. 1986). The modified Syed- Neblett (Disaia et al. 1990) applicator consists of a perineal template, vaginal obturator, and 17-gauge hollow guides of various lengths. The vaginal obturator is 2 cm in diameter, and 12 cm or 15 cm in length. The vaginal obturator has seven grooves on its surface for the placement of guide needles and is centrally drilled so it can allow the placement of a tandem to be loaded with ¹³⁷Cs sources. This makes possible to combine an interstitial and intracavitary application simultaneously (Fig. 27.6).

A similar afterloading applicator, referred to as

“MUPIT” (Fig. 27.7) was developed by Martinez et al. (1984). The device basically consists of two acrylic cylinders, an acrylic template with predrilled holes that serve as guides for trocars and a cover plate.

Some of the guide holes on the template are angled outward to permit a wide lateral coverage without danger of striking the ischium. The cylinders have an axial hole large enough to pass a central tandem or suction tube for drainage of secretions. Thus, the device allows for the interstitial placement of ¹⁹²Ir ribbons as well as the intracavitary placement of either ¹³⁷Cs tubes or ¹⁹²Ir ribbons. In use, the cylinders are placed in the vagina and rectum and then fastened to the template, so that a fixed geometric relationship among the tumor volume, normal structures and source placement is preserved throughout the course of implantation.

The major advantage of these systems is greater control of the placement of the sources relative to tumor volume and critical structures due to the fixed geometry provided by the template and cylinders. In addition, improved dose-rate distributions are obtained by means of computer-assisted optimization of the source placement and strength during the planning and loading phases. Due to the inac-

Fig. 27.6. Syed-Neblett and modiﬁ ed templates

(13)

curacies of pelvic examination and close proximity of the rectum and bladder to the target volume, there exists a serious risk of either underdosing the target volume or causing bladder and rectal morbidity. In order to improve the accuracy of target localization and needle placement, several investigators have explored performing ITI under trans- rectal ultrasound (TRUS) (Stock et al. 1997), CT, MRI-planned implants with endorectal coil (Corn et al. 1995), laparotomy and laparoscopic guidance (Childers and Surwit 1993; Corn et al. 1995).

While laparotomy facilitates the displacement of bowel during the procedure using slings or tissue expanders and/or lysis of adherent bowel, there is some degree of associated morbidity, such as ileus, bleeding and increased operative time. Laparoscopy is a shorter and less invasive procedure. Real time TRUS-guided Syed-Neblett template implantation technique was reported by Stock et al. (1997). TRUS allows the ultrasound (US) probe to be in closer proximity to the pelvic structures (cervix, parametria, vagina) than trans-abdominal US and can more accurately guide needle placement into tumor and avoid needle insertion into critical surrounding normal tissues. Transverse US imaging is used to assure that the needles cover the target area and do not enter the bladder, rectum or small bowel. The longitudinal mode of the US probe is equally important in the implant procedure due to its ability to guide the optimum depth of needle insertion. With this technique, invasive laparotomy and/or laparoscopy can often be avoided, providing an interactive, non-invasive technique allowing for highly accurate needle placement (Stock et al. 1997).

27.6.2

SCC: Treatment Options and Outcome by FIGO Stages

27.6.2.1

FIGO Stage 0: VAIN – CIS

VAIN has been approached both surgically and medically by multiple investigators. Treatment options include local excision, partial or complete vaginectomy, laser vaporization, topical 5%

fluorouracil (5-FU) administration, or ICB alone.

Overall, the reported control rates are very similar among the different approaches, ranging from 48% to 100% for laser (Diakomanolis et al. 2002;

Hoffman et al. 1991; Jobson and Homesley 1983;

Julian et al. 1992; Stafl et al. 1977; Townsend et

al. 1982), 52% to 100% for colpectomy (Creasman et al. 1998; Fanning et al. 1999; Hoffman et al.

1992; Robinson et al. 2000), 75% to 100% for topical 5-FU (Krebs 1989; Petrilli et al. 1980; Piver et al. 1979; Woodruff et al. 1975) and 83% to 100%

for RT (Chyle et al. 1996; Kirkbride et al. 1995;

Perez et al. 1977). The degree of VAIN and the age and general health of the patient are important treatment considerations. RT has a long history of documented efficacy and has a significantly better therapeutic ratio than other modalities (Chyle et al. 1996; Kirkbride et al. 1995; Perez et al. 1977;

Prempree and Amomman 1985). Using conventional LDR-ICB techniques, the entire vaginal mucosa should receive between 60 Gy and 70 Gy in one or two implants (Perez et al. 1977). Perez et al.

(1977) reported only one distal local failure in the 20 patients treated for CIS. This recurrence developed distally to the vaginal vault in a patient inadequately treated with vaginal ovoids only.

There have been some reports in the literature regarding the use of HDR-ICB for patients with VAIN-3. Ogino et al. (1998) reported 6 patients treated with HDR to a mean dose of 23.3 Gy (range 15–30 Gy), none of whom developed recurrent disease. Limited rectal bleeding and moderate to severe vaginal mucosa reactions were noted in patients treated to the entire length of the vagina. MacLeod et al. (1997) used HDR-ICB to treat 14 patients with VAIN 3 with a dose of 34–45 Gy in 4.5-Gy to 8.5- Gy fractions, with a local control of 78.5%. With a median duration of follow-up of 46 months, two patients developed grade-3 vaginal toxicity. At the present time, no definite conclusions can be drawn from the limited data published in the literature regarding the use of HDR-ICB. Based on the excellent local control and functional results obtained

Fig. 27.7. A. MUPIT template structure. Reprinted with per- mission from Martinez et al. (1984)

(14)

with LDR-ICB, this remains, in the authors’ opinion, the treatment of choice when definitive RT is used.

27.6.2.2 Invasive SCC

Most authors emphasize that brachytherapy alone is adequate for superficial FIGO stage-I patients, for whom 95–100% local control has been achieved with intracavitary and interstitial techniques (Chu and Beechinor 1984; Kucera and Vavra 1991; Leung and Sexton 1993; Perez et al. 1988; Peters et al.

1985; Reddy et al. 1991; Stock et al. 1995; Urbanski et al. 1996). Superficial lesions can be adequately treated with ICB alone using afterloading vaginal cylinders. Mucosal doses of 80–120 Gy are typically delivered, depending on the diameter of the cylinders, when prescribing 65–70 Gy at 0.5 cm depth beyond the vaginal surface (Perez et al. 1977). For lesions thicker than 0.5 cm at the time of implantation, it is advisable to combine ICB and ITB in order to deliver tumor dose in the range of 65–70 Gy, calculated to the base of the lesion, limiting the proximal and distal vaginal mucosal doses to 140 Gy and 100 Gy, respectively (Figs. 27.8 and 27.9).

There are no well-established criteria regarding the use of EBRT in patients with stage-I disease.

Perez et al. (1988, 1999) did not find a significant correlation between the technique of irradiation used and the probability of local or pelvic recurrence, probably since the treatment technique varied based on tumor-related factors. There is general consensus that EBRT (20–50 Gy) is advisable for larger, more infiltrating or poorly differentiated tumors that may have a higher risk of lymph node metastasis.

Chyle et al. (1996) recommended EBRT in addition to brachytherapy for stage-I disease to cover at least the paravaginal nodes and, in larger lesions, to cover the external and internal iliac nodes. The 5-year survival for patients with stage-I disease treated with RT alone ranges from 70% to 95%.

Patients with FIGO stage-II disease are uniformly treated with EBRT, followed either by ICB and/or ITB (Fig. 27.8a–c). Perez et al. (1999) showed that in stage IIA, the local tumor control was 70% (37/53) in patients receiving brachytherapy combined with EBRT, compared with 40% (4/10) in patients treated with either brachytherapy or EBRT alone. In stage IIB, the local-regional control was also superior with combined EBRT and brachytherapy (61% versus 50%, respectively). Generally, 40–50 Gy is delivered to the whole pelvis, followed by an additional boost

of 30–35 Gy given with brachytherapy. Patients with lesions limited to the upper third of the vagina can be treated with an intrauterine tandem and vaginal ovoids or cylinders. In patients with parametrial infiltration, a “boost” with EBRT and/or an interstitial implant is advisable to deliver a minimum tumor dose of 70–75 Gy and 55–60 Gy to the pelvic side wall (Fig. 27.8a–c). The 5-year survival for patients with stage-II disease treated with RT alone ranges between 35% and 70% for stage IIA and 35%

and 60% for stage IIB. The results of several series published in the literature using RT with or without limited surgical resection for the treatment of stage- I and -II vaginal cancer are shown in Table 27.2 (Chyle et al. 1996; Creasman et al. 1998; Davis et al. 1991; Kirkbride et al. 1995; Kucera and Vavra 1991; Perez et al. 1999; Stock et al. 1995; Urbanski et al. 1996).

Generally, patients with FIGO stages III–IV disease will receive 45–50 Gy EBRT to the pelvis and, in some cases, additional parametrial dose with midline shielding to deliver up to 60 Gy to the pelvic side walls. Ideally, ITB brachytherapy boost is performed, if technically feasible, to deliver a minimum tumor dose of 75–80 Gy (Fig. 27.8a–c).

If brachytherapy is not feasible, a shrinking-field technique can be used, with fields defined using the three-dimensional treatment planning capabilities to deliver a tumor dose around 65–70 Gy (Fig. 27.3).

The overall cure rate for patients with stage-III disease is 30–50%. Stage IVA includes patients with rectal or bladder mucosa involvement or, in most series, positive inguinal nodes. Although some patients with stage-IVA disease are curable, many patients are treated palliatively with EBRT only.

Pelvic exenteration can also be curative in highly selected stage-IV patients with small volume central disease. Table 27.3 (Chyle et al. 1996; Creasman et al. 1998; Kirkbride et al. 1995; Kucera and Vavra 1991; Perez et al. 1999; Stock et al. 1995; Urbanski et al. 1996) shows the treatment results in patients with advanced disease. However, each of these series reported a greater number of patients with similar stage disease treated with RT, which represents the preferred approach in contemporary practice (Perez et al. 1988; Prempree and Amommam 1985).

27.6.3

Chemotherapy and Radiation

The control rate in the pelvis for stages-III to -IV patients is relatively low, and about 70–80% of the

(15)

a1

b2

c

Fig. 27.8a–c. Interstitial brachytherapy boost in a patient with locally advanced vaginal cancer. Left anterior oblique (a1) and right anterior oblique (a2) radiographs of the implant. Coronal (b1) and sagittal (b2) isodose distributions. c Placement of funnel needles using the modiﬁ ed Syed-Neblett template

b1

a2

(16)

a b

c1 c2

Fig. 27.9a–c. AP (a) and lateral (b) radiographs of an interstitial and intracavitary implant in carcinoma of the vagina using the modiﬁ ed Syed-Neblett template. c Three-dimensional isodose distribution: c1 Ir¹⁹² only; c2 Ir¹⁹² plus Cs¹³⁷

patients have persistent or recurrent disease in the pelvis, despite high doses of external beam RT and brachytherapy. Failure in distant sites does occur in about 25–30% of the patients with locally advanced tumors, much less than pelvic recurrences. There- fore, there is a need for better approaches to the management of advanced disease, such as the use of concomitant chemo-radiotherapy. Agents such as 5-FU, mitomycin and cisplatin have shown promise when combined with RT, with complete response rate as high as 60–85% (Evans et al. 1988; Roberts et al.

1991) but long-term results of such therapy have been variable. In these small studies, many of the patients had advanced (stage III) disease at the initiation of combined modality therapy, perhaps explaining

the lack of long-term disease control. Evans et al.

(1988) found no local recurrences, however, among patients achieving a complete response with RT and 5-FU plus mitomycin-C (12 of 25 patients), with a median follow-up period of 28 months, suggesting that local control may be improved with combined modality therapy since local failure is common with radiation alone in large volume pelvic disease. The survival for the entire population was 56% (66%

for patients with primary vaginal cancer). Only 2 patients had severe complications, although the authors recognize that longer follow-up is probably required to assess the true incidence of late effects. More sobering are the data from Roberts et al. (1991) who reported 67 patients with advanced

(17)

cancers of the vagina, cervix and vulva treated with concurrent 5-FU, cisplatin and RT. Although 85%

experienced a complete response, 61% of them experienced a tumor recurrence, with a median time to recurrence of only 6 months, and an overall survival at 5 years of 22%. Further, of 67 patients, severe late complications developed in 9 (13%), 8 of whom required surgeries. Kersch et al. (1990) reported that 5 of 8 vaginal cancer patients achieved local control with combined modality therapy. Studies of primary chemo-radiation in primary vaginal cancer are small or heterogeneous populations including cervical and vulvar cancers, making it difficult to truly assess the role of combined modality therapy in the management of locally advanced disease.

No randomized trials comparing radiation with or without chemotherapy have been reported.

Dalrymple et al. (2004) published recently a small study including 14 patients, primarily stages I and II SCC of the vagina treated with reduced doses of RT (median 63 Gy) concurrently with different 5-FU-based chemotherapeutic regimens. They report a 93% control rate, probably reflecting a more favorable stage distribution. Interestingly, none of the patients required interstitial implants and no patients developed fistulas. The authors indicated that this approach, similar to that used in the management of anal and vulvar cancer, would allow reducing the RT dose with the subsequent improve- ment in organ function and late toxicity.

Table 27.2 FIGO stages I–II vaginal cancer. Treatment outcome with radiation therapy with/

without surgery

Author No. of patients Outcome (survival)

Chyle et al. (1996) 59 stage I 10 years 76%

104 stage II 10 years 69%

Creasman et al. (1998) 169 stage I 5 years 73%; 79% S+RT (47), 63% RT (122) 175 stage II 5 years: 58%; 58% S+RT (39), 57% RT (136) Davis et al. (1991) 19 stage I 5 years 100% S+RT (5), 65% RT (14)

18 stage II 5 years 69% S+RT (9), 50% RT (9) Kirkbride et al. (1995) 40 stage I 5 years 72%

Kuceraand Vavra (1991) 16 stage I 5 years 81%

23 stage II 5 years 43.5%

Perezet al. (1999) 59 stage I 10 years 80%

63 stage IIA 10 years 55%

34 stage IIB 10 years 35%

Stocket al. (1995) 8 stage I 5 years 100% S+RT, 80% RT 35 stage II 5 years 69% S+RT, 31% RT Urbanski et al. (1996) 33 stage I 5 years 73%

Table 27.3 FIGO stages III–IV vaginal cancer. Treatment outcome with radiation therapy with/

without surgery

Author No. of patients Outcome (survival)

Chyle et al. (1996) 55 stage III 10 years 47%

16 stage IV 10 years 27%

Creasman et al. (1998) 180 stage III–IV 5 years 36%; 60%-S+RT (36), 35%-RT (144) Kirkbride et al. (1995) 42 stage III–IV 5 years 53%

Kucera and Vavra (1991) 46 stage III 5 years 35%

19 stage IVA 5 years 32%

Perez et al. (1999) 20 stage III 10 years 38%

15 stage IV 0%

Stock et al. (1995) 9 stage III 5 years 0%

8 stage IV 0%

Urbanski et al. (1996) 40 stage III 5 years 22.5%

15 stage IVA 0%

(18)

Further investigation is needed to determine the therapeutic efficacy of the concurrent chemo- radiotherapy and the optimal chemotherapy regimen. Recently published data on locally advanced cervical cancer have demonstrated an advantage in loco-regional control, overall survival and DFS for patients receiving cisplatin-based chemotherapy concurrently with RT (Keys et al. 1999; Morris et al. 1999; Rose et al. 1999; Whitney et al. 1999).

The only drug common to all the studies was cisplatin, suggesting it may be the only agent needed to improve radiation sensitivity. Based on these data, as well as data on loco-regionally advanced vulvar cancer (Moore et al. 1988), consideration should be given to a similar approach in patients with advanced vaginal cancer. Randomized trials comparing RT alone to chemo-radiation therapy, however, are unlikely due to small patient numbers.

27.6.4

Patterns of Failure in SCC

Of those patients who experienced a tumor recurrence, at least 85% will have loco-regional failure, and the vast majority of these recurrences will be confined to the pelvis and vagina (Chyle et al. 1996;

Davis et al. 1991; Kirkbride et al. 1995; Kucera and Vavra 1991; Perez et al. 1999; Stock et al. 1995;

Tabata et al. 2002; Urbanski et al. 1996). The rate of loco-regional recurrence in stage I is approximately 10–20% versus 30–40% in stage II (Davis et al. 1991;

Perez et al. 1999; Tabata et al. 2002). The pelvic control rate for patients with stage III and stage IV is relatively low, and about 50–70% of the patients have recurrences or persistence, despite well-designed RT (Perez et al. 1999; Tabata et al. 2002). The median time to recurrence is 6–12 months. Tumor recurrence is associated with a dismal prognosis, with only a few long-term survivors after salvage therapy. Failure in distant sites alone or associated with loco-regional failure does occur in about 25–40%

of patients with locally advanced tumors (Chyle et al. 1996; Davis et al. 1991; Kirkbride et al. 1995;

Kucera and Vavra 1991; Perez et al. 1999; Tabata et al. 2002; Urbanski et al. 1996).

27.6.4.1

Potential RT-Related Factors Influencing Outcome

It is important to recognize that analysis of RT doses and techniques and their impact on local/pelvic

tumor control is fraught with difficulty, since the available data is retrospective, and not the result of prospective randomized or dose-escalation studies.

Given the fact that more than 75% of the recurrences are local, the necessity for optimizing local therapy is clear. Clinical experience dictates that higher doses of RT, greater than 70–75 Gy, when feasible, are generally prescribed for more advanced stages of the disease (Andersen 1989; Chyle et al. 1996; Perez et al. 1999; Spirtos et al. 1989). Perez et al. (1988, 1999) reported increased tumor control in patients with stages IIA–IVA with EBRT and brachytherapy, compared with patients receiving brachytherapy alone. In patients with stage-I disease, no correlation was found between the technique of RT used and the incidence of local or pelvic recurrences. In addition, they suggested that doses in the range of 70–75 Gy to the primary tumor volume and 55–65 Gy to the medial parametria for patients with more advanced disease are necessary to optimize tumor and pelvic control. Furthermore, of 100 patients with primary tumors involving the upper and middle third of the vagina who received no elective irradiation to the groin, none developed metastatic inguino-femoral lymph nodes, in contrast to 3 of 29 (10%) with lower third primaries, and 1 of 20 with tumors involving the entire length of the vagina. Of 7 patients with initially palpable inguinal lymph nodes treated with doses in the range of 60 Gy, only 1 developed a nodal recurrence. The authors recommended that elective RT of the inguinal lymph nodes should be carried out only in patients with primary tumors involving the lower third of the vagina.

Stock et al. (1995) found a significant increase in local control and 5-year survival for patients receiving EBRT and brachytherapy, compared with those treated with EBRT alone. The 5-year actuarial local control and survival in the EBRT and brachytherapy group were 44% and 50%, respectively, compared with 12% and 9%, respectively, in the EBRT alone group. However, these two groups were not evenly matched with a large percentage of stage-IV lesions in the EBRT alone group compared with the brachytherapy group.

Lee et al. (1994) identified overall treatment time as the most significant treatment factor predicting pelvic tumor control in 65 patients with carcinomas of the vagina treated with definitive RT. If the entire course of RT, including EBRT and brachytherapy, was completed within 9 weeks, pelvic tumor control was 97%, in contrast with only 57% when treatment time extended beyond 9 weeks (P<0.01). Similarly, Pingley et al. (2000)reported that DFS rate was

(19)

reduced from 60% to 30% if the overall treatment time was prolonged. Conversely, Perez et al. (1999) did not find a significant impact of prolongation of treatment time on pelvic tumor control. Neverthe- less, these authors advocate completion of treatment within 7–9 weeks.

27.7

Clear Cell Carcinoma of the Vagina

Since Herbst’s first report (Herbst and Scully 1970) of 7 adenocarcinomas arising in the vagina of ado- lescent females after in-utero exposure to DES, there have been several reports limited to DES-related vaginal CCA (Herbst and Anderson 1990; Herbst et al. 1971, 1972, 1979). In 1979, Herbst et al. (1979) reported 142 cases of stage-I CCA of the vagina. An 8% risk of recurrence was seen after radical surgery (n=117), and an 87% survival was achieved. There was a 36% risk of recurrence after RT for stage-I lesions; however, the authors acknowledged that, in general, RT was reserved for large stage-I lesions which involved more of the vault and were less amenable to surgical resection. As the majority of CCAs occur in the upper third of the vault, the largest series (Herbst et al. 1979; Senekjian et al. 1987, 1988) addressing the surgical approach to these lesions have advocated radical hysterectomy, pelvic and para-aortic lymphadenectomy, and sufficient colpectomy to achieve negative margins. Senekjian has also reported a series of exenterations done for CCA (Senekjian et al. 1989). However, there have been efforts to also attempt fertility-sparing radical resections (Hudson et al. 1983, 1988) or more limited wide local excisions followed by some form of RT (Senekjian et al. 1987).

Senekjian et al. (1987) reported a series of 219 stage-I CCA cases with 92% overall 5-year and 88%

10-year survival rates, respectively, in 176 patients receiving conventional therapy (identical to 43 who had undergone local therapy). Of the 176 treated conventionally, 128 underwent radical hysterectomy and vaginectomy; 16 had the same operation followed by adjuvant RT; and 32 were treated with RT alone. Because of the risk of node metastases, 14 of 43 patients treated with local therapy underwent extraperitoneal pelvic lymphadenectomy. Of the 43 patients treated with local therapy, 9 had vaginectomy, 17 had local excision alone, 6 had brachytherapy alone and 11 had combined local excision and brachytherapy. The 10-year actuarial recur-

rence rate was an unsatisfactory 45% in those who underwent local excision alone, versus only 16% if they had received conventional therapy and 27%

if they had received local excision followed by RT.

Senekjian et al. (1987) advocated a combination of wide local excision and extraperitoneal node dissec- tion followed by brachytherapy for patients desirous of fertility preservation.

In a subsequent report, Senekjian et al. (1988) reviewed the experience with 76 cases with stage-II CCA from the Registry for Research on Hormonal Transplacental Carcinogenesis. The overall 5- and 10-year survival rates were 83% and 65%, respectively. Of the 76 patients, 22 received surgery exclu- sively (either radical hysterectomy with vaginectomy, 13 patients, or exenterative type procedure, 9 patients), 38 received RT alone, 12 received combination therapy and 4 underwent other approaches.

Patients treated with primary RT achieved an 87%

5-year survival rate versus 80% for those treated with surgery and 85% for those receiving both treatments. The authors concluded that most patients with stage-II vaginal CCA should be treated with combination EBRT and brachytherapy; however, small, easily resectable lesions in the upper fornix might undergo resection, allowing better preservation of coital and ovarian function (Senekjian et al. 1988). In 1989, Senekjian et al. (1989) reported their experience of 20 pelvic exenterations for CCA of the vagina, including 13 for primary lesions and 7 for recurrent disease. They reported a 72% suc- cess rate if the exenterations were done as part of primary therapy. The authors advocated reserving exenterative approaches for those who have failed RT in order to maximize quality of life for the great- est number of patients.

There are few published reports regarding the use of systemic therapy for this tumor. Fowler et al. (1979) reported one complete and one partial response after treatment with melphalan (1mg/kg qd×5 days). Robboy et al. (1974) reported responses in recurrent disease to both 5-FU and vin- blastine.

27.8

Salvage Therapy

In general, the patient with recurrent cancer of the lower female genital tract presents a difficult clinical dilemma. Optimal therapy for patients with recurrent gynecological cancer after potentially curative