Role of tumour-free margin distance for loco-regional control in vulvar cancer-a subset analysis of the Arbeitsgemeinschaft Gynäkologische Onkologie CaRE-1 multicenter study.

ABSTRACT

AIM OF THE STUDY A tumour-free pathological resection margin of ≥8 mm is considered state-of-the-art. Available evidence is based on heterogeneous cohorts. This study was designed to clarify the relevance of the resection margin for loco-regional control in vulvar cancer. METHODS: AGO-CaRE-1 is a large retrospective study. Patients (n = 1618) with vulvar cancer ≥ FIGO stage IB treated at 29 German gynecologic-cancer-centres 1998-2008 were included. This subgroup analysis focuses on solely surgically treated node-negative patients with complete tumour resection (n = 289). RESULTS: Of the 289 analysed patients, 141 (48.8%) had pT1b, 140 (48.4%) pT2 and 8 (2.8%) pT3 tumours. One hundred twenty-five (43.3%) underwent complete vulvectomy, 127 (43.9%) partial vulvectomy and 37 (12.8%) radical local excision. The median minimal resection margin was 5 mm (1 mm-33 mm); all patients received groin staging, in 86.5% with full dissection. Median follow-up was 35.1 months. 46 (15.9%) patients developed recurrence, thereof 34 (11.8%) at the vulva, after a median of 18.3 months. Vulvar recurrence rates were 12.6% in patients with a margin <8 mm and 10.2% in patients with a margin ≥8 mm. When analysed as a continuous variable, the margin distance had no statistically significant impact on local recurrence (HR per mm increase 0.930, 95% CI 0.849-1.020; p = 0.125). Multivariate analyses did also not reveal a significant association between the margin and local recurrence neither when analysed as continuous variable nor categorically based on the 8 mm cutoff. Results were consistent when looking at disease-free-survival and time-to-recurrence at any site (HR per mm increase 0.949, 95% CI 0.864-1.041; p = 0.267). CONCLUSIONS: The need for a minimal margin of 8 mm could not be confirmed in the large and homogeneous node-negative cohort of the AGO-CaRE database.