Recent progress in elucidating the complex and heterogeneous interactions between malignancy and coagulation or fibrinolysis reactions in humans has clarified the pathogenesis of disseminated intravascular coagulation that occurs with malignancy and has revealed evidence for two distinct pathways of growth regulation based on production by tumor cells of initiators of thrombin formation versus plasminogen activators. We have proposed a preliminary classification of tumors (see Table 2) based on these interactions. Type I tumors are those in which the tumor cells are associated with an intact coagulation pathway that leads to thrombin formation at the tumor periphery but in which the tumor cells lack u-PA. Examples of tumors in this category include SCCL, malignant melanoma, and renal cell carcinoma. Type II tumors are those in which the tumor cells express u-PA but lack an associated coagulation pathway leading to thrombin formation. Examples of type II tumors include prostate cancer, colon cancer, breast cancer, and N-SCLC. Type III tumors are those that express neither of these pathways, or exhibit some other pattern of interaction. Obviously, this formulation must be regarded as hypothetical. However, this concept fits with the limited data available to date from clinical trials. More importantly, this hypothesis can be tested further by means of intervention aimed at interrupting pathways relevant to specific tumor types. Characterization of additional tumor types by the methods described should permit amplification of this classification of tumors and other patterns of interaction may be defined. Exploration of the coagulation-cancer interaction holds considerable promise for gaining new understanding of both the coagulation mechanism and tumor biology. Most intriguing is the prospect that imaginative approaches to cancer treatment may be devised that are not only relatively nontoxic and low cost, but also effective.