Recent research has revealed the existence of specific mutations in cancer. These mutations are being investigated as targets to find subjects at high risk for cancer, to detect early cancer, to detect the early recurrence of established cancer, and to find micrometastasis. These mutations are reviewed for the major anatomic sites. Some of the clinical issues related to the application of these mutations and the limitations of using molecular targets are also considered. Current methods for determining the risk of cancer are reviewed. Risk assessment is essential for defining cohorts for chemoprevention and other interventions. The concept of using surrogate anatomic and functional sites for estimating risk is introduced. Finally, the increasing complexity of molecular genetic analysis and the biologic heterogeneity of cancer are discussed in relation to early detection.