In modern conformal radiotherapy (CRT), we attempt to increase its therapeutic ratio, thus improving the survival chances and/or quality of life for patients. It is common to acknowledge that poor local tumor control or increased normal tissue complications may arise from inaccurate targeting of the tumor, failure to conform the high-dose distribution to the target volume, and inaccurately delivered radiation doses. A further cause for concern is the influence that errors or inaccuracies in the dose calculation may have on the management of radiation therapy. Such errors arise from inherent limitations in the calculation algorithm used, which are more significant in some anatomical sites than others. Furthermore, an estimate of the therapeutic ratio is given by the ratio of tumor control probability (TCP) and normal tissue complication probability (NTCP). The effectiveness of these predictive indicators also depends on the accuracy of the calculated dose distributions in the target and surrounding normal structures. In this work, we compared CRT dose distributions of plans for the treatment of prostate, head-and-neck, and lung tumors using the measurement-based Clarkson and model-based Superposition dose calculation algorithms. Dose-volume histograms (DVHs) for the planning target volume (PTV) and sensitive structures, as well as NTCP and TCP, were compared. Dose distributions, observed in the lung and head-and-neck plans, vary significantly with respect to dose conformity as a function of algorithm used. Differences in the calculated maximum dose of up to 14% were observed in the PTV and sensitive structures for the lung and head-and-neck Clarkson-based plans, respectively, compared to the Superposition-based plans. Furthermore, a difference in the biological outcomes of up to 14% in the NTCP and 4% in the TCP was noticed. The CRT plans show the importance of accurate modeling of the effect of tissue inhomogeneities on dose distributions in the target and critical structures for lung and head-and-neck treatments.