MP49-17\u2003INTERNATIONAL MULTI-CENTER ANALYSIS OF FACTORS PREDICTING THE PRESENCE OF NON-SENTINEL NODE INGUINAL METASTASES IN PENILE CANCER PATIENTS UNDERGOING COMPLETION INGUINAL LYMPH NODE DISSECTION FOLLOWING POSITIVE SENTINEL NODE BIOPSY

Abstract

INTRODUCTION AND OBJECTIVES: Currently, penile cancer patients with impalpable inguinal lymph nodes routinely undergo dynamic sentinel node biopsy (DSNB) to diagnose metastatic spread to the inguinal basin. Those with positive DSNB are then offered completion inguinal lymph node dissection (ILND). Given the high morbidity associated with ILND we set out to identify factors within the positive sentinel node that may help to predict those at highest risk of having metastases at ILND. METHODS: Patients who had undergone completion ILND following positive DSNB were identified retrospectively from the penile cancer databases of two high volume European centers. Four potential predictive covariates were analysed; size of the largest positive sentinel node (SL), size of the largest area of metastasis within a positive lymph node (SM), percentage of lymph node replaced by cancer (PM), and extracapsular spread (ECS). ROC curve analysis of each covariate was used to assess strength as a test, and identify a cut-off value for regression analysis. Odds ratios were derived using logistic binary regression. Statistical analysis was performed using IBM SPSS 25. RESULTS: Between November 2004 and November 2017 (Centre 1) and February 1994 and June 2016 (Centre 2) 279 patients underwent 304 completion ILND following positive sentinel node biopsy. Complete data was available for 287. The mean number of sentinel nodes sampled from each groin was 2.21 of which a mean of 1.24 were positive, which was similar for those with positive and negative ILND. Of the variables tested SL did not show any correlation to ILND positivity, but SM, PM and ECS did. When SM was stratified into three groups by increasing size a clear trend was demonstrated with increasing size showing an increasing percentage of patients having positive non-sentinel nodes. ROC curve analysis derived cut-offs of 6.25mm for SM, and 32% for PM. Odds ratio of having a positive ILND with SM>6.25 was 2.09 (p=0.012), with PM>32% it was 2.81 (p=0.001), and with ECS positive it was 1.8 (p=0.081). CONCLUSIONS: Completion groin dissection following positive DSNB is a procedure with a high level of associated morbidity. By using the predictive factors we have identified within a positive sentinel node we can identify those patients who are more likely to have non-sentinel node metastases. This will allow a better informed discussion with our patients regarding the need for further surgery. Source of Funding: None