In thin melanomas, the involvement of regional nodes is very uncommon. Recent sentinel node (SN) studies have confirmed the absence of positive regional lymph nodes in melanomas < 0.76 mm and a 5% positivity in melanomas between 1.0 and 1.99 mm. The chance of regional lymph node involvement - and therefore whether it is relevant to perform the SN procedure - seems to depend on the Breslow thickness of the primary tumour. However, a Breslow thickness cut-off point has not yet been established. We evaluated a melanoma population that had undergone an SN procedure to determine this point, so that the procedure can be restricted to a smaller group of patients in future. In a total of 348 patients with proven American Joint Committee on Cancer (AJCC) stages I or II cutaneous melanoma with a Breslow thickness > or = 0.5 mm the triple technique was used, consisting of preoperative visualization of the lymph channels from the initial site of the melanoma towards the SN by (dynamic) lymphoscintigraphy, intraoperative visualization of those particular lymph channels and nodes with blue dye, and a gamma probe to measure accumulated radioactivity in radiolabelled lymph nodes. In melanomas thinner than 0.90 mm, no positive SN was found (95% confidence interval 0-5%). This group consisted of 75 patients (22%), with a median follow-up of 31 months. Our data suggest that this procedure need no longer be indicated for almost a quarter of the patient population, because the cut-off point for nodal involvement appears to be a Breslow thickness of 0.90 mm.