Associated Lichen Sclerosis Increases the Risk of Lymph Node Metastases of Vulvar Cancer

The most important prognostic factor in vulvar cancer is inguinal lymph node status at the time of diagnosis, even in locally advanced vulvar tumors. The aim of our study was to identify the risk factors of lymph node involvement in these women, especially the impact of lichen sclerosis (LS). We conducted a retrospective population-based cross-sectional study in two French referral gynecologic oncology institutions. We included all women diagnosed with a primary invasive vulvar cancer. Epithelial alteration adjacent to the invasive carcinoma was found in 96.8% (n = 395). The most frequently associated was LS in 27.7% (n = 113). In univariate analysis, LS (p = 0.009); usual type VIN (p = 0.04); tumor size >2 cm and/or local extension to vagina, urethra or anus (p < 0.01), positive margins (p < 0.01), thickness (p < 0.01) and lymphovascular space invasion (LVSI) (p < 0.01) were significantly associated with lymph node involvement. In multivariate analysis, only LS (OR 2.3, 95% CI [1.2–4.3]) and LVSI (OR 5.6, 95% CI [1.7–18.6]) remained significantly associated with positive lymph node. LS was significantly associated with older patients (p = 0.005), anterior localization (p = 0.017) and local extension (tumor size > 2 cm: p = 0.001). LS surrounding vulvar cancer is an independent factor of lymph node involvement, with local extension and LVSI.


Introduction
Vulvar cancer is a rare disease representing only 0.3% of all cancers in women and from 3% to 5% of gynecologic cancers [1,2]. It is most frequently seen in women between 65 and 75 years [3,4]. Ninety percent of vulvar cancers are squamous cell carcinoma (SCC), whereas the other histologic subtypes such as melanoma, Paget's disease and adenocarcinoma of the Bartholin gland are less frequent [5].
The most important prognostic factor is inguinal lymph node status at the time of diagnosis [6][7][8], even in locally advanced vulvar tumors: The 5-year and 10-year survival of patients without metastatic nodes are 62% and 47% respectively, versus 39% and 27% for patients with metastatic nodes [7]. Besides the classic predictive factors of inguinal lymph node involvement, such as depth of invasion and tumor size, the influence of the type of surrounding epithelial disorders may be important. Epithelial disorders are found in more than 80% of patients. They are subdivided into lesions due to carcinogenic genotypes of human papillomavirus (HPV) including unifocal and multifocal usual vulvar intraepithelial neoplasia (uVIN) and non HPV lesions in which lichen sclerosis (LS) is the most frequent observed [9][10][11][12]. It is known that the type of epithelial disorder has prognostic significance in patients with vulvar cancer [9][10][11][12][13], but there are few data to date about clinical and prognostic features of patients with LS.
The objective of this study was to identify the risk factors of lymph node involvement in women with vulvar cancer and especially the impact of LS.

Materials and Methods
The study protocol was approved by the Ethics Committee of Paris X, France. The present report complies with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement [14].
Our departments are referral centers for vulvar disease. We treat about 25 vulvar cancers per year. For this cross-sectional analysis, we examined the records of consecutive patients with invasive vulvar carcinoma managed in the tertiary University Hospital of Creteil from 1992 to 2012 and the University Tenon Hospital (AP-HP) from 2005 to 2010. We excluded patients with recurrent vulvar carcinoma or in situ carcinoma. The following data were recorded: age, body mass index (BMI), American Society of Anesthesiologists (ASA) status, site of lesion, histologic type and characteristics, surgical treatment, complications, adjuvant treatment, relapse and survival data. Stages were assigned based on clinical data and lymph node pathology when available according to the 2009 FIGO (International Federation of Gynecology and Obstetrics) staging [6]. Patients who had been staged before 2009 were retrospectively assigned for the data collection. Surgical treatment of the vulva was radical and conservative (hemivulvectomy or local excision) when the surgical excision could encompass the lesion with at least a 1-cm margin of clinically normal skin. Patients with exclusive lateral vulvar lesions had a unilateral lymphadenectomy. Groin node dissection was avoided in patients older than 70 years without clinically metastatic nodes, because of their medical status or in cases of superficially invasive carcinoma (<1 mm). Type of dissection was reviewed according to the classification described by Rouzier et al. [15]. Since 2002, sentinel node identification was progressively performed in our center, consistent with the guidelines from the literature and the Collège National des Gynécologues et Obstétriciens Français (CNGOF) [16][17][18][19]. Patients with invasive vulvar cancer more than 1 mm, of clinical stages T1 and T2 underwent sentinel lymph node detection, bilateral in case of midline or close-to-midline lesions (<1 cm from the midline). The presence of clinically suspicious or palpable lymph nodes was a contraindication to performing the sentinel node procedure [16][17][18][19]. The technic used for sentinel lymph node mapping was the one described by Christine Louis Sylvestre et al. [19] using a double detection method.
Adjuvant therapy decision was based on final pathological analysis, and both centers followed latest ESGO guidelines at the time the patient was treated.
The associated epithelial disorders were retrospectively separated according to the International Society for the Study of Vulvar Disease (ISSVD) [10,13,20,21]. Because most of the data were not interpretable with the last ISSVD classification, we used the previous ISSVD classification [22].

Statistical Analysis
Characteristics of the included women are described as a number (%) for qualitative variables and median (interquartile range, IQR) for quantitative variables and compared according to presence of LS. To identify factors associated with a poorer prognosis, we compared initial characteristics of patients with and without lymph node involvement using the chi-squared test or Fisher's exact test, as appropriate, for categorical variables and the nonparametric Kruskal-Wallis test for continuous variables. Odds ratios (ORs) and 95% confidence intervals (CIs) were estimated using univariate asymptotic or exact logistic regression analyses, as appropriate, for variables that yielded p values < 0.05. These variables were selected for multivariate analyses. As we were interested in LS, we did not include the variable "usual type VIN" in the multivariate analysis. Confounders and interactions were tested in bivariate models.
All tests were two-sided, and p values ≤ 0.05 were considered significant. Analyses were performed using STATA software version 12.0 (StataCorp, College Station, TX, USA) and the R software (http://cran.r-project.org).
Radiotherapy was delivered to the inguinal area in 17.4% patients (n = 71).

Factors Associated with Lymph Node Involvement
In univariate analysis, the following variables were significantly associated with lymph node involvement: the presence of LS (  Odds ratios and 95% confidence intervals were estimated using univariate logistic regression. ** Multivariate analysis was adjusted for the 7 variables listed in the table. ‡ Odds ratio and confidence interval is expressed for 5 mm increase.

Characteristics of the Population with Lichen Sclerosis
Because the presence of LS had a strong and independent predictive value for positive lymph node status, we analyzed patients' characteristics associated with this epithelial disorder (Table 3). These patients differed significantly from the global population. LS was significantly associated with older patients (p = 0.005), anterior localization (p = 0.017), tumor size > 2 cm (p = 0.0005), positive margins (p = 0.03) and thickness (p < 0.0001). We found an OR for the risk of lymph node involvement of 1.89 [1.14-3.13]. To the contrary, the absence of LS was significantly associated with FIGO 1 (p = 0.009) and tumor size < 2 cm (p = 0.004).

Main Findings
We found that the type of epithelial disorder associated with invasive vulvar carcinoma was an important prognostic factor of lymph node involvement. While LS was significantly associated with positive lymph node (38.9% versus 25.1% in the absence of LS, p = 0.009), uVIN emerged as a protective factor (23.8% versus 35.1% in the absence of uVIN, p = 0.04). In multivariate analysis, LS remained an independent risk factor, as well as LVSI.

Strengths and Limitations
A strength of our study is that the results are based on a large cohort (more than 400 patients) since vulvar cancer is a rare disease. We were able to establish a significant correlation between simple clinical or histologic characteristics and lymph node involvement, which is a major prognostic factor. However, it was a retrospective, two-center study spanning a period of 20 years. During this period, the FIGO staging system was revised to reflect the importance of lymph node involvement as a prognostic factor of vulvar cancer [6][7][8]. Four major changes were thus introduced in 2009 to further characterize stage III, which previously comprised a heterogeneous group of patients with negative or positive nodes. To take this into account, patients who had been staged before 2009 were retrospectively assigned for data collection, rendering analysis more difficult. Another change involved the classification of VIN, which was modified by the ISSVD in 2004. Prior to this date, dVIN was not clearly identified and we chose to use the previous classification [22] as most of the data are older than 2004. This is a limit of the present study and would need a review of the histologic slides. But LS is a clinical entity, and the significative association of LS with lymph node involvement, independently of the presence of atypia, softens this limit. Eventually, since this information was not recorded in our databases, we cannot confirm or infirm that the adjuvant therapy was performed according to the same protocol (especially radiotherapy), which might have introduced a bias.

Interpretation
According to previous reports [10,12,13,20], epithelial disorders associated with vulvar carcinoma are divided into two distinct etiologies: LS and HPV induced disorders, respectively associated with dVIN and uVIN (Table 4). Usual VIN is HPV related in most cases, including warty, basaloid and mixed type. It is typically seen as a unifocal or multifocal lesion with a variety of clinical presentations: erosions, plaques and nodules, pigmented, red or white. Invasive squamous carcinomas of warty or basaloid type are associated with uVIN. To the contrary, differentiated VIN is seen particularly in older women with LS and/or squamous cell hyperplasia in some cases [12]. Neither dVIN, nor associated keratinizing SSC is HPV related [10,11,20]. LS was present in 27.7% of our patients, in accordance with another large study of 1287 patients with vulvar carcinoma reporting 33% of patients with LS [23]. Table 4. Classification of epithelial disorders associated with invasive vulvar carcinoma [8,11,14,15]. The prognostic significance of adjacent lesions in the setting of vulvar cancer has already been studied. Before the new classification of epithelial disorders proposed by the ISSVD in 2004 [10], the presence of VIN in adjacent lesions had been identified as a predictor of reduced disease-free survival, even before using the recent classification of epithelial disorders [24]. However, in 2001, Rouzier et al. showed that the presence of HPV-related VIN was a positive prognostic factor [9]. In their article, patients with uVIN had a 5-year survival rate of 87%, against 42% for patients with non-neoplastic epithelial disorders, dVIN or without associated epithelial alteration (p < 0.01). However, the data regarding lymph node involvement was not reported. Our study shows that 30 patients (23.8%) with uVIN had positive lymph nodes. Usual VIN was more likely to be associated with superficially invasive lesions, which have an excellent prognosis: no spread to the groin has been reported for a depth of infiltration <1 mm [25][26][27]. Most of the authors studied HPV-related uVIN [9,24,28]. The good prognosis of HPV is widely proven and also well known in oropharyngeal cancers [29]. However, only a few studied the prognostic significance of LS in vulvar cancer, which is more frequent and underdiagnosed [12,30,31].

Usual VIN Differentiated VIN
In our study, the site of onset in case of LS was more frequently of anterior localization and of larger and deeper extension. These differences in invasion may be due to a later diagnosis in women with LS. As they are older, they may be more prone to ignoring lesions than women with HPV-related VIN. Moreover, the clinical distinction with cancer can be difficult.
Another hypothesis is that these two forms of dermatosis could have a difference in invasive potential. Ansink et al. showed that patients with an HPV-positive pCR tumor had a better prognosis than those with an HPV-negative pCR tumor (p = 0.03) [28]. Histologically, HPV-independent vulvar carcinomas-associated with dVIN and LS-frequently show mutations of p53 and are histologically keratinizing, whereas HPV-associated vulvar carcinomas are of the basaloid or warty type and arise from uVIN. The viral gene of HPV is involved in a specific process of malignant transformation; p16 immunohistochemistry is diffusely positive in these lesions, and a high Ki-67 proliferation index is observed [13,32]. These two different oncogenic pathways lead to different degrees of malignancy. Indeed, our study suggests two different natural histories and locoregional power of invasion, depending on the type of adjacent epithelial disorder. In the literature, dVIN has a higher risk of progression to invasive SCC than uVIN (33% vs. 5.7%, respectively). In addition, time to progression to SCC is significantly shorter in women with dVIN compared with those with uVIN [13,[33][34][35]. Consequently, it could be interesting to compare the prognosis of patients with and without lymphadenectomy and the effect of radiotherapy depending the type of VIN.

Conclusions
In conclusion, LS surrounding vulvar cancer is an independent predictive factor of positive lymph node status. LS has its own clinical presentation requiring careful monitoring, independently of the presence of atypia. A biopsy should be performed to identify invasive lesion as soon as a suspect area is observed as LS increases the risk of lymph nodes metastasis. Prospective clinical studies could show whether the treatment of this pre-invasive lesion could reduce the risk of vulvar carcinoma.