Although the Breslow measurement of tumor thickness of melanoma is the most significant predictor of survival, the biologic behavior of thick lesions remains unpredictable. MIB-1, a monoclonal antibody to a Ki-67 epitope, recognizes all proliferating cells. Unlike Ki-67 antibody, which requires frozen tissue, MIB-1 can be used on formalin-fixed tissue. Proliferation, measured by MIB-1 expression and mitotic index, was assessed as a prognostic factor in a group of patients with clinical stage I thick cutaneous melanoma (tumor thickness 4 mm or greater), for which predicted survival is low. From a melanoma data base, 97 patients with this type of melanoma were identified. Of these, 64 had lesional tissue available for study. The median follow-up time was 3.8 years (range 0.42-13.6 years). The percentage of MIB-1 reactivity was scored as low at less than 10% (n = 33), intermediate at 10% to 20% (n = 17), and high at greater than 20% (n = 14). Melanomas with high MIB-1 reactivity were associated with significantly poorer cause-specific survival compared with tumors with intermediate (p < 0.0001) or low MIB-1 reactivity (p = 0.0025). Multivariate analysis demonstrated that MIB-1 reactivity was a significant independent prognostic factor related to cause-specific survival (p = 0.0002) and was more sensitive than tumor thickness or mitotic index in this select group of high-risk patients. Identification of individuals with stage I thick cutaneous melanoma who are at risk of recurrent disease may improve patient management as new therapeutic modalities become available.