EGFR expression is associated with groin node metastases in vulvar cancer, but does not improve their prediction.

OBJECTIVES High morbidity of elective inguinofemoral lymphadenectomy in early stage vulvar cancer patients urges the need for defining a group of low-risk patients in whom inguinofemoral lymphadenectomy can be safely omitted. Aim of the study was to evaluate whether in addition to 'classic' clinicopathological factors determination of EGFR expression in vulvar cancer can be helpful in defining such a 'low-risk' group. METHODS Formalin-fixed paraffin-embedded tumor tissue samples of 197 surgically treated T1/2 patients were collected in a Tissue Micro Array (TMA). On this TMA, immunohistochemistry for EGFR was performed. Logistic regression analyses were performed including histopathological characteristics with the presence of nodal metastases as outcome. A predictive model was constructed, and absolute risks were calculated. RESULTS EGFR expression was present in 68% of the vulvar tumors and related to the presence of lymph node metastases (OR 2.12, 95% CI 1.09-4.10). Our predictive model with only clinicopathological factors was able to define a group of patients with a likelihood of absence of lymph node metastases of 13% (95% CI 5-36), which could be decreased to 6% (95% CI 0-29) after inclusion of EGFR expression (p=0.07). CONCLUSIONS EGFR expression is present in the majority of vulvar tumors and is associated with groin node metastases in vulvar cancer. Current classic clinicopathological predictive factors for inguinofemoral lymph node metastases with or without EGFR analysis are not strong enough for identification of "sufficiently low" risk T1/2 vulvar cancer patients. Our predictive model approach however is excellent for evaluation of new cell biological parameters, associated with clinical outcome.