The general utility of the maximum tolerated dose (MTD) paradigm, a strategy aimed at optimizing the chance of total tumor cell eradication, is here questioned. Evidence to date suggests that for many tumors the potential for eradication is in fact remote, with patients consistently demonstrating tumor cell presence subsequent to MTD treatments having eradicative intent. The failure to eradicate is attributed largely to the heterogeneous nature of the tumor. Heterogeneous cell populations demonstrate short-term refractoriness to up-front dose delivery, but "resensitize" as part of dose recovery, showing increased overall susceptibility to a given series of doses when delivered more evenly spaced. It is demonstrated: (1) that the minimization of total tumor burden, rather than complete eradication, may often be the more practical objective; and (2) that regularly spaced, "metronomic" dosing is the best way to achieve it. As a corollary, it is found that the more efficient ability of the tumor endothelial cells to resensitize following dosing predicts a targeting bias towards the endothelial compartment of a tumor when metronomic dosing is employed. This lends theoretical support to recent empirical studies showing that regularly spaced dosing schedules with no extended rest periods act more antiangiogenically, thereby delaying or avoiding the onset of acquired resistance.