A major goal of personalized medicine in oncology is the identification of drugs with predictable efficacy based on a specific trait of the cancer cell, as has been demonstrated with gleevec (presence of Bcr-Abl protein), herceptin (Her2 overexpression), and iressa (presence of a specific EGFR mutation). This is a challenging task, as it requires identifying a cellular component that is altered in cancer, but not normal cells, and discovering a compound that specifically interacts with it. The enzyme NQO1 is a potential target for personalized medicine, as it is overexpressed in many solid tumors. In normal cells NQO1 is inducibly expressed, and its major role is to detoxify quinones via bioreduction; however, certain quinones become more toxic after reduction by NQO1, and these compounds have potential as selective anticancer agents. Several quinones of this type have been reported, including mitomycin C, RH1, EO9, streptonigrin, β-lapachone, and deoxynyboquinone (DNQ). However, no unified picture has emerged from these studies, and the key question regarding the relationship between NQO1 processing and anticancer activity remains unanswered. Here, we directly compare these quinones as substrates for NQO1 in vitro, and for their ability to kill cancer cells in culture in an NQO1-dependent manner. We show that DNQ is a superior NQO1 substrate, and we use computationally guided design to create DNQ analogues that have a spectrum of activities with NQO1. Assessment of these compounds definitively establishes a strong relationship between in vitro NQO1 processing and induction of cancer cell death and suggests these compounds are outstanding candidates for selective anticancer therapy.