One of the visionary ideas of the late Dr. Judah Folkman was to treat cancer based on discovery of ultra-early biomarkers long before tumors become symptomatic or determination of their anatomical location is feasible. This idea was based on the hypothesis that most tumors remain dormant as microscopic lesions for a long period of time, until they gain the ability to recruit vessels in a process called tumor angiogenesis. Dormant tumors that switch to the angiogenic phenotype grow exponentially and ultimately become clinically apparent, which is referred to as the disease state of cancer. A major limitation in investigating the tumor dormancy process is the lack of suitable experimental models and the limited clinical accessibility of dormant tumors. In this report, Almog and colleagues identified the molecular determinants of tumor dormancy by characterizing the consensus transcriptome signature of dormant versus fast-growing experimental human osteosarcoma, liposarcoma, breast carcinoma, and glioblastoma multiforme tumor models. Differential regulation and activation of several signaling pathways not previously associated with the dormancy process are also reported and validated based on protein expression and activation levels. The molecular signature of tumor dormancy was further used to test the feasibility of selected proteins as blood (plasma)-based biomarkers in mice and in human specimens. The enhanced sensitivity of dormant tumors to angiostatin, compared with that of fast-growing angiogenic tumors, was correlated with selective uptake of angiostatin (green) in angiomotin (red)-positive dormant osteosarcoma cells (blue, cell nucleus). This work may provide molecular instructions for discovery of ultra-early tumor biomarkers and possible therapeutic targets for the blockage of early tumor development. For details, see the article by Almog and colleagues on page 836 of this issue.