The effectiveness of anticancer therapies relies on the ability of these substances to selectively eliminate the malignant cells with little or no toxicity to normal cells. The isolation in most human tumors of a rare subpopulation of cancer stem cells (CSCs) associated with chemo resistance leads to the "stem cell theory" (SCT). The SCT proposed that eliminating this fraction will eventually cure cancer but experimental data supporting this classical view are controversial and now being gradually replaced by other models. These novel models of cancer biology predict that to cure cancer only drugs or combination of drugs that eliminate all (CSCs and non-CSCs) cancer cells at once ("pankiller drugs") will be effective. The search for "pankiller drugs" will require tests to assess (i) the elimination of all cancer cells in in vitro systems (ii) the ability to eradicate the tumors and prevent tumor relapse in in vivo systems. However, at present, most drugs are being tested in assays that can only provide a picture of the short term activity of anticancer compounds. This in part explains why only a small fraction of the drugs that enter clinical trials are actually approved for clinical use. This article will provide a concise review of the systems, assays and endpoint parameters routinely used to screen for potential anticancer drugs and propose, based in the current knowledge of cancer biology, a more rationale anticancer drug screening program.