30-44% of patients with clinical groin node melanoma have involved pelvic nodes. Clinical guidelines selectively target pelvic lymph node dissection (PLND) to those meeting radiological and clinico-pathological criteria, but we lack satisfactory diagnostic tools to preoperatively identify patients with pelvic node disease. We evaluate routine PLND for all patients undergoing superficial node dissection (SLND), performed as a combined single-stage ilioinguinal lymph node dissection (ILND). Retrospective analysis of 67 ILNDs in consecutive patients presenting with palpable, cytologically melanocytic groin nodes. We examine predictors of pelvic node status and determine efficacy of 2010 UK guidelines in patient selection for PLND. 28 patients (42%) had histologically positive pelvic nodes; half had just one involved node (53.6%). 43% of pelvic metastases were radiologically occult. Significant predictors of pelvic melanoma were stage N3 groin nodes (p = 0.049), one third of groin nodes involved (p = 0.0009), positive Cloquet's node (p = 0.005), previous in transit disease (p = 0.001), and staging CT (p = 0.007). UK guidelines, primarily reliant upon staging CT, were effective selection criteria (p = 0.04), identifying 57% of pelvic metastases. CT and in-transit disease status in combination was the strongest predictor of pelvic disease (p = 0.006, RR 4.5, PPV 0.75, NPV 0.83). A combined CT and in-transit disease status provides a potentially clinically useful preoperative selection tool for ILND. With a high prevalence of occult pelvic node involvement, potential to avoid the morbidity of untreated pelvic nodes, and 5 year survival figures of 24-35% following surgery, we advocate ILND in all patients with clinically evident melanoma in a single groin node.