Selective sentinel lymph node (SLN) dissection is widely used in the management of cutaneous melanoma patients without clinical evidence of nodal metastases. A series of 274 consecutive melanoma patients who underwent melanoma primary excision and SLN mapping at our institutions since 1998, and were thereafter followed up and eventually treated, is reported in this prospective study. The aim was to analyse the parameters associated with a higher risk of occult nodal metastases, to evaluate the clinical outcome of melanoma patients who underwent SLN procedure, and to identify by means of multivariate analysis the prognostic parameters with independent predictive value on disease-free survival (DFS) in node-positive and negative patients. The SLN was tumour-negative in 228 patients (83.2%). A disease progression occurred in 25 (10.9%); among them, 10 patients in whom the initially identified SLN had been negative, developed a clinically and histologically evident positive lymph node in the same basin during follow-up. Five-year DFS and overall survival were 75% and 82%, respectively. In 46 patients (16.8%), the SLN proved to be tumour positive. The percentage of SLN-positive patients varied according to the primary thickness, from 11.8% in patients with Breslow of 2 mm or lower, to 34.7% in patients with Breslow from 2 to 4 mm, up to 55.9% in patients with Breslow greater than 4 mm (P<0.001). Only two patients with Breslow thickness lower than 1 mm had positive SLN biopsy. Five-year DFS and overall survival (OS) were 42 and 69%, respectively, significantly lower than those of negative SLN-patients (P<0.001). Multivariate analyses showed that the parameters with prognostic independent value on DFS were SLN status (micrometastases or macrometastases; P=0.0001), and to a lesser extent, Breslow thickness (P=0.04). In conclusion, our data support the clinical usefulness of SLN dissection as a reliable and accurate staging method in patients with cutaneous melanoma. SLN-positive patient OS (5-year survival 69%) seems to be superior to that historically reported for stage III patients treated with curative nodal dissection only after the clinical evidence of palpable adenopathies (5-year survival 36%). The prognostic relevance of the pattern of SLN invasion (micrometastases/macrometastases) could be the basis for the planning of adjuvant treatment trials on selected groups of patients.