Previous clinical and experimental studies suggested that invasion of the brain by metastatic melanoma may follow the external surfaces of vascular channels, that is, angiotropic extravascular migratory metastasis. Such angiotropic invasion seemss analogous to that of neoplastic glial invasion of the nervous system. We, therefore, have retrospectively investigated 20 primary melanoma cases and their respective metastatic brain lesions. The following parameters were analyzed in each primary melanoma: presence of angiotropism, Breslow thickness, Clark level, mitotic rate, sentinel lymph node (SLN) status, and time interval between the primary lesion and the metastasis. The metastatic brain lesions were examined for the presence of angiotropism. Of the 20 cases, 14 showed angiotropism in the primary lesion. The angiotropic group had a significantly deeper Breslow thickness (median 4.4 mm vs. 1.4 mm, P < 0.01) and was more mitotically active (median 11 vs. 4.7 mitoses/mm2, P = 0.04). Interestingly, the angiotropic group had an average time lapse of 33 months from the primary lesion to the brain metastasis, whereas the nonangiotropic group had a 57-month time interval. Although the Kaplan–Meier analysis failed to show a survival difference in this small cohort (P = 0.235), there was a trend toward significance. Seven of 20 brain metastases showed angiotropism; however, no significant correlation between angiotropism in the primary melanomas and the corresponding metastatic lesions could be demonstrated. Indeed, angiotropism in the brain metastases was difficult to assess because the available material were generally small partial biopsy samplings and many showed conspicuous necrosis. Ten melanoma patients underwent SLN biopsy. The 3 of 6 positive cases in the angiotropic group had an average time lapse of 32 months from the primary lesion to the brain metastasis, whereas the 4 positive SLN biopsies in the nonangiotropic group had an average of 63 months. This preliminary study of angiotropism in primary melanomas and their corresponding brain metastasis shows a striking trend suggesting that angiotropism in primary melanomas may predict the rapid development of brain metastases. This study also has demonstrated the difficulty in studying angiotropism in melanoma brain metastases because of small sample sizes and abundance of necrotic tissue. The authors are in the process of collecting larger and more representative numbers of melanoma brain metastases for further investigations.