Advanced treatment optimization is possible using quantitative radiobiological dose response models. Although all present models are necessarily linked to a certain degree of uncertainty, this will only have a small influence on the relative shape of the resultant optimal dose distribution. However the exact dose level should perferably be determined clinically by dose escalation with the optimized dose plan as a control arm. It is shown that a large part of the biological effect of high linear energy transfer radiation is due to the spectrum of low-energy delta-electrons that can generate dense clusters of complex DNA damage. Such electrons are efficiently generated by low-energy photons or densely ionizing ion beams and to a considerably smaller degree by high-energy electrons, photons and protons. A new analytical expression is developed for the effective radiation resistance of heterogeneous tumors, making it possible to approximate the response of such tumors by the effective clonogen number N0,eff and the effective D0 value D0,eff. It is shown that a relatively small number of resistant tumor cells may well be sufficient to dominate the response of hypoxic or otherwise heterogeneous tumors. Finally, several examples are given of intensity-modulated dose distributions generated by multiple radiation modalities, the total effect of which is biologically optimized.