With the aim to relate the effects observed in a clinical boron neutron capture therapy protocol to the corresponding outcomes in a standard photon radiation therapy, “RBE-weighted” doses are customarily calculated by adding the contributions of the different radiations, each one weighted by a fixed (dose and dose rate independent) relative biological effectiveness factor. In this study, the use of fixed factors is shown to have a formal inconsistency, which in practice leads to unrealistically high tumor doses. We then introduce a more realistic approach that essentially exploits all the experimental information available from survival experiments. The proposed formalism also includes first-order repair of sublethal lesions by means of the generalized Lea-Catcheside factor in the modified linear-quadratic model, and considers synergistic interactions between different radiations. This formalism is of sufficient simplicity therefore to be directly included in all BNCT treatment planning systems. In light of this formalism, the photon-isoeffective doses for two BNCT clinical targets were computed and compared with the standard dose calculation procedure. For the case of brain tumors and clinically relevant absorbed doses, the proposed approach derives isoeffective doses that are much lower than the fixed RBE method, regardless of considering synergism. Thus, for a tumor that receives a mean total absorbed dose of 15 Gy (value achievable with 50 ppm of boron concentration and typical beams used in the clinic), the photon-isoeffective doses are 28 Gy (IsoE) and 30 Gy (IsoE) (without and with synergism, respectively), in contrast to 51 Gy (RBE) for the fixed RBE method. When the clinical outcome of the Argentine cutaneous melanoma treatments is assessed with regard to the doses derived from the standard procedure, it follows that the fixed RBE approach is not suitable to understand the observed clinical results in terms of the photon radiotherapy data. Moreover, even though the assumed ¹⁰B concentration in tumors is lowered to reduce the obtained doses with the standard procedure, the fixed RBE approach is still unsuitable to explain the observed outcomes (the model is always rejected with P values of virtually zero). Additionally, the numbers of controlled tumors predicted by the proposed approach are statistically consistent with observed outcomes. As a by-product of this work, a dose-response clinical reference for single-fraction melanoma treatments is developed.