The growth fraction of tumors from patients with non-Hodgkin's lymphomas (NHL) has been shown to correlate with survival in retrospective studies. The growth fraction can be evaluated using immunohistochemical techniques employing the Ki-67 monoclonal antibody (MoAb) that marks a nuclear protein present in cycling cells. The purpose of this study was to evaluate the clinical utility of the Ki-67 MoAb for predicting survival. Using a prospective trial design in a multi-institutional cooperative trials group, the proliferative index, clinical outcome, and statistical correlations were independently assessed for previously untreated patients with advanced stages of intermediate- and high-grade histologies of NHL treated on Southwest Oncology Group study (SWOG 8516, Intergroup 0067). The proportion of Ki-67-positive cells was determined on snap-frozen thin tissue sections. A proliferative index of 80% or greater, as determined from prior retrospective studies, identified a group of patients (18%) who had a poor outcome. Overall survival was significantly reduced in these patients with a high Ki-67-associated proliferative index compared with those with a low proliferative index (P = .001). One-year survival estimates were 82% (low proliferative index) versus 18% (high proliferative index). A multivariate regression analysis incorporating commonly used clinical prognostic features confirmed the independent effect of proliferation on survival (relative risk estimate 5.9; 95% confidence interval, 2.2, 16.1). The Ki-67 MoAb identifies a group of patients with rapidly fatal NHL for whom currently available chemotherapy is inadequate.