Using immunohistochemistry with anti-D2-40 for the detection of lymphovascular invasion (LVI-IHC) in 74 cases of invasive melanoma, we found LVI in 23% (16/74) of the tumors. Data on sentinel lymph node (SLN) biopsy were available for 36 patients. Sixty-seven percent (6/9) of patients with LVI-IHC and 19% (5/27) without LVI-IHC had positive SLN. Follow-up data were available for 60 patients. Data on recurrence/metastasis were available for 60 patients. Twenty-five percent (15/60) had LVI with immunohistochemistry. Fifty-three percent (8/15) of these patients had "distant" metastasis or regional recurrence compared with 11% (5/45) in those without LVI-IHC. Overall and disease-specific survival was shorter for patients with LVI. In both the univariate and multivariate Cox proportional hazards regression models, LVI-IHC in addition to ulceration was statistically significant with respect to overall survival. Specifically, in the reduced multivariate model, compared with patients with no LVI, patients with intratumoral LVI had a hazard ratio (HR) of 5.4 (95% CI 1.6-18.4), while patients with peritumoral LVI had a HR of 3.8 (95% CI 0.7-20.9). In addition, patients with ulceration had an increased hazard of 4.4 (95% CI 1.2-16.8). For the first time, we herein show a positive correlation with LVI in melanoma detected with immunohistochemistry and distant metastasis, overall survival and disease-free survival.