CHAPTER 41 Malignant disease of the vulva and vagina
Introduction
Vulval cancer is an uncommon condition that largely affects elderly women (Figure 41.1). The Office of National Statistics recorded 842 cases in 2005 (Office of National Statistics 2008). This incidence of 3.3 per 100,000 ranks vulvar cancer as the 20th most common cancer in women. The most recent mortality figures recorded 270 deaths for all age groups, giving a death rate of 1.05 per 100,000 women (0.53/105 persons), ranking it the 19th most common cause of cancer death in women.
Figure 41.1 Age distribution of vulval cancer.
Source: Office of National Statistics 2008 Cancer Statistics Registration, Registration of Cancer Diagnosed in 2005, England. Series MB1 No. 36. Her Majesty’s Stationery Office, London.
Aetiology
There seem to be two distinct types of vulval carcinoma: the first type occurs predominantly in older women, is typically a well-differentiated keratinizing tumour that is not associated with usual-type vulvar intraepithelial neoplasia (VIN) or human papilloma virus (HPV) infection, but which often shows adjacent epithelial disorders such as lichen sclerosus and/or differentiated VIN; and a second type which mainly occurs in younger women, is associated with usual-type VIN, shows evidence of oncogenic HPV infection, and may be associated with synchronous or metachronous squamous preneoplastic or neoplastic lesions of the cervix, vagina and anal canal (Crum et al 1997, van der Avoort et al 2006). Smoking may be an important cofactor involved in the aetiology of HPV-related vulvar tumours (Hussain et al 2008).
The following are recognized risk factors for all types of vulval cancer:
Histology
The majority of vulvar cancers (90%) are squamous in origin (Figure 41.2). Verrucous and basal cell cancers are squamous cell variants.
Non-squamous cancers include Bartholin’s gland cancer and other adenocarcinomas, malignant melanoma, Paget’s disease and sarcomas. The vulva may also be a site for lymphoma and metastases from elsewhere (Finan 2003). Rarely, primary breast cancers have been reported (Piura et al 2002).
Lymphatic Drainage
Lymph drains from the vulva to the superficial inguinal glands and then to the deep femoral glands in the groin. Drainage continues to the external iliac glands. Drainage to both groins occurs from midline structures — the perineum and the clitoris — but some contralateral spread may take place from other parts of the vulva (Iversen and Aas 1983). Direct spread to the pelvic nodes along the internal pudendal vessels occurs very rarely, and no direct pathway from the clitoris to the pelvic nodes has been demonstrated consistently. An important aspect of the lymphatic drainage is the concept of sentinel nodes in each groin. This is the first node that draining lymph encounters as it drains bilaterally from the vulvar basin (Cabanas 1977). This anatomical concept has been exploited recently to develop selective lymphadenectomy in this disease.
Natural History of Squamous Cancers
Malignant transformation may occur in more than one site, and this appears to apply in both HPV-negative and -positive scenarios, although more commonly in the latter. The earliest recognizable phase of vulval cancer is termed ‘superficially invasive disease’ (stage Ia). The tumour is less than 2 cm in lateral dimension, and there is less than 1 mm invasion when measured from the base of an adjacent dermal papilla. Accurate identification and classification of this stage requires expert pathological interpretation, and is exceptionally important as the current consensus would suggest a virtually negligible risk of lymph node metastases. As the cancer gradually increases in size and progressively invades the deeper layers of the dermis, it spreads locally. The tumour will eventually involve the local lymphatics, hence the propensity for groin lymph node involvement (Table 41.1).
Table 41.1 Groin recurrences in women who did not undergo any form of groin node surgery. None had clinically suspicious lesions in the groin, and all were regarded as ‘low-risk cases’. In many, groin node surgery was omitted because of associated medical problems
Presentation
The medial aspects of the labia majora are the most common sites for disease to develop (70%). The labia minora, clitoris, periurethral areas, posterior fourchette and perineum account for the remainder (30%). Small lesions may be asymptomatic and go unnoticed by the patient. However, both Hacker et al (1981) and Monaghan (1990) have commented that, in a significant minority of patients, there appears to be considerable delay in presentation. The intimate nature of the disease, fear and/or ignorance and the advanced age of many patients might partly explain this observation. The fact that the disease is uncommon may also be contributory, as primary carers may not recognize the significance of symptoms or fail to recognize the clinical signs. Whatever the cause, a significant number of cases continue to present in advanced stages.
The reasons for presenting have been analysed by Podratz et al (1983a). Pruritus is the most common presenting symptom (71%), and an ulcer or mass will have been noted in nearly 80% of cases. Surprisingly, pain is only a feature in 23% of cases. Bleeding (26%), discharge (13%) and urinary tract dysfunction (14%) are other common reasons for presentation.
Examination
The character, size and location(s) of all lesions should be documented. Particular attention should be paid to assess any involvement of the vagina, urethra, bladder base or anus. Palpation is important to determine possible involvement of the surrounding bony structures. Discomfort and tenderness may necessitate an examination (and biopsies) under general anaesthesia. The groin should be examined, although a negative assessment cannot reliably exclude cancer. Hard, enlarged, fixed or ulcerated nodes (Figure 41.3) need to be documented as this will almost certainly influence management.
Diagnosis
Diagnosis is based upon a representative biopsy of the tumour. The biopsy should include an area where there is a transition from normal to malignant tissue (Figure 41.4). Biopsies should be of a sufficient size to allow differentiation between superficially invasive tumours and frankly invasive tumours, and orientated to allow quality pathological interpretation. Radical excision should not be undertaken without prior biopsy confirmation of malignancy unless there are extenuating circumstances.
The histopathological report should include:
Staging
There are two staging systems: the Tumour, Node, Metastasis system and the International Federation of Gynaecology and Obstetrics (FIGO) system. Most gynaecologists use the latter which, since its last update in 2000, is based upon clinical and surgical findings. Table 41.2 details both staging systems.
Table 41.2 FIGO staging of vulval cancer
| FIGO stage | Description |
|---|---|
| I | Tumour confined to the vulva |
| Ia | Lesions ≤2 cm in size, confined to the vulva or perineum and with stromal invasion ≤1 mm. No nodal metastasis |
| Ib | Lesions >2 cm in size or with stromal invasion >1 mm confined to the vulva or perineum. No nodal metastasis |
| II | Tumour of any size with extension to adjacent perineal structures (lower 1/3 urethra; lower 1/3 vagina; anus) with negative nodes |
| III | Tumour of any size with or without extension to adjacent perineal structures (lower 1/3 urethra; lower 1/3 vagina; anus) with positive inguinofemoral nodes |
| IIIa | |
| IIIb | |
| IIIc | With positive nodes with extracapsular spread |
| IV | Tumour invades other regional (upper 2/3 urethra; 2/3 vagina) or distant structures |
| IVa | Tumour invades any of the following |
| IVb | Any distant metastasis including pelvic lymph nodes |
FIGO, International Federation of Gynaecology and Obstetrics.
Prognostic Factors
The size of the primary tumour and the status of the regional lymph nodes are the only two factors that have been consistently associated with outcome (Homesley et al 1991). The 5-year survival rate in cases without lymph node involvement is in excess of 80%, falling to less than 50% if the inguinal nodes are involved and 10–15% if the iliac or other pelvic nodes are involved.
Depth of invasion
Invasion less than 1 mm (superficially invasive or stage Ia) has a negligible risk of lymph node involvement, and groin node dissection may be omitted (Hacker et al 1984b). The risk of nodal metastasis rises to 8% for a depth of 1–2 mm, 11% for 2–3 mm, and over 25% for lesions with a depth greater than 3 mm (Morrow et al 1993) (Figure 41.5).
Lymphovascular space permeation
This factor along with tumour border pattern (infiltrating vs pushing) and perineural invasion are not included in the surgicopathological staging of vulvar cancer, although they are associated with an increased risk of metastasis (Heaps et al 1990, Hopkins et al 1991).
Lymph node status
This is the most important prognostic variable. The number of lymph nodes involved and the nature of involvement influence the prognosis. Microscopic involvement of a single groin node, without extracapsular involvement, represents a low risk of recurrence, and adjuvant therapy is not necessary (Hacker et al 1983). Conversely, multiple nodal diseases and macroscopic and/or capsular involvement are indications for adjuvant radiotherapy in order to reduce the risk of groin recurrence, which is usually fatal (van der Velden et al 1995). It is well recognized that groin node dissection is associated with significant morbidity. Almost 50% of patients undergoing groin node dissection will suffer postoperative complications, most commonly wound infection, wound breakdown and lymphoedema (Gaarenstroom et al 2003). For these reasons, a presurgical investigation that could identify those at risk could have a significant therapeutic benefit. Assessment by clinical palpation of the groins is inadequate; of patients with clinically normal lymph nodes, 16–24% have metastases, while 24–41% of those with clinically involved nodes are negative when examined histologically (Sedlis et al 1987, Homesley et al 1993). Even though the majority of vulval cancers are still assessed clinically, there is increasing interest in utilizing additional imaging, particularly of the regional node groups. Ultrasonography, computed tomography (CT), magnetic resonance imaging (MRI) and positron emission tomography have all been assessed but lack the sensitivity to exclude microscopic disease reliably (Selman et al 2005). Sentinel node sampling (vide infra) has shown promise, and the collective experience of this technique now suggests that it is reliable, at least in early-stage disease.
Sentinel lymph node detection
The concept of the ‘sentinel lymph node’ (SLN) was introduced in 1977 by Cabanas (1977). He defined the SLN as the first node in the lymphatic basin that receives primary lymphatic flow. Subsequent studies have confirmed that a tumour-free SLN implies absence of lymph node metastases in the entire draining lymphatic basin (Balega and van Trappen 2006). The last two decades have seen the technique become established in the management of cutaneous melanoma and breast cancer. The concept (Figure 41.6) has now been introduced for the management of early vulval cancers, and the published performance data to date are promising (Table 41.3).
A large-scale international trial, the Groningen International Study on Sentinel Nodes in Vulvar Cancer (GROINSS-V), was established to test the safety and clinical utility of SLN dissection in early-stage vulvar cancer. This was an observational study. Patients with stage I and II disease with tumour size less than 4 cm were recruited. Only patients found to have disease in either SLN underwent full inguinofemoral lymphadenectomy. Among 259 patients with unifocal vulvar cancer and a negative sentinel node, groin recurrences occurred in six (2.3%) women, and the 3-year survival rate was 97%. This result was comparable with similar historical groups of patients that had full inguinofemoral lymphadenectomy. However, both short- and long-term morbidity were significantly reduced, including shorter hospital stay and less wound breakdown and lymphoedema (van der Zee et al 2008). The authors thus proposed that sentinel node dissection should be the standard treatment for patients with unifocal early-stage vulvar cancer. In the second phase of the GROINSS-V study, investigators are evaluating radiotherapy, rather than complete inguinofemoral lymphadenectomy, in patients with metastatic SLNs. The objective is to further reduce morbidity by avoiding double-modality treatment.
Treatment
Factors influencing the management plan
Site of the tumour
Centrally located tumours lie close to midline structures such as the clitoris, urethra, vagina and anus, and are associated with a higher risk of bilateral inguinal nodal spread compared with more lateral tumours. If the medial margin lies within 1 cm of the midline, they should be defined as central (de Hullu and van der Zee 2006). Bilateral groin node dissection is recommended for central tumours, and ipsilateral groin node dissection is recommended for lateralized tumours (Figure 41.7). Furthermore, if central tumours approximate to the urethra or anal canal such that these structures would be sacrificed to achieve satisfactory clearance, urinary or bowel diversion may be necessary.
Surgery
Surgery is the mainstay of treatment. The traditional ‘butterfly’ incision en-bloc radical (Figure 41.8) vulvectomy introduced by Taussig (1940) and Way (1960) clearly improved survival. The excised tissues included the vulvar lesion, the inguinofemoral lymph nodes and the lymphatics in between. Large areas of normal tissue were frequently included, and primary wound closure was rarely achieved. Protracted postoperative recovery and severe wound complications and disfigurement were associated with this procedure. This postoperative morbidity has been the main driver for progress over the last two decades. Surgical treatment has become more individualized and conservative, aimed at reducing morbidity without compromising survival.
Separate incisions
The use of three separate incisions for groin dissections and vulvectomy instead of the traditional ‘butterfly incision’ has improved wound healing and reduced the frequency of wound breakdown (Hacker et al 1981). The suggestion that this lesser approach might result in an increase in skin bridge recurrence was not born out in a Cochrane review, which showed a skin bridge recurrence rate of less than 1% (Ansink and van der Velden 2000). The hypothesis underpinning this strategy is based upon the belief that in early tumours, cells disseminate by embolization rather than in continuity (Willis 1973); therefore, there should be no residual tumour in the lymphatic channels between the tumour and the groin nodes. Reported cases of skin bridge recurrences were mainly seen in women with large groin node metastasis (Rose 1999). No prospective randomized controlled studies have been performed comparing the traditional radical vulvectomy with en-bloc inguinofemoral lymphadenectomy with the separate incision technique. However, de Hullu et al (2002) did undertake a retrospective comparison and noted a significant increase in groin and skin bridge recurrence with the less radical procedure (6.3% vs 1.3%, P=0.029). This, however, did not result in a difference in overall survival. They concluded that in view of the morbidity of the ‘butterfly incision’ and relatively low skin bridge recurrence rate, separate incisions were the preferable technique for stage I and II disease.
Modified radical vulvectomy or wide local excision
The main factor causing severe morbidity and disfigurement following radical vulvectomy is the extensive dissection and removal of normal tissue from the vulvar region. In the last two decades, less radical surgery, such as modified radical vulvectomy, hemivulvectomy and wide local excision, has been introduced to reduce morbidity without compromising survival. Tumour-free margins are the most important predictive factor for local recurrences. Both Heaps et al (1990) and de Hullu et al (2002) showed that a tumour-free margin of more than 8 mm was associated with significantly lower local recurrence rate. These studies found that suboptimal tumour-free margins (<8 mm) were more likely when the intentional macroscopic margin was 1 cm or less. A lesion might extend further than judged by macroscopic inspection, and tissues usually shrink during preparation for histological examination (Balega et al 2008). A macroscopic surgical margin of 2 cm is thus recommended. This may be difficult to achieve in some cases where the tumour is adjacent to the urethra and anus, but every effort should be made to maximize the margins. The tumour-free margins should be the same regardless of whether a radical vulvectomy or a wide local excision is performed.
