Understanding the influence of dipolar interactions in magnetic hyperthermia experiments is of crucial importance for fine optimization of nanoparticle (NP) heating power. In this study we use a kinetic Monte Carlo algorithm to calculate hysteresis loops that correctly account for both time and temperature. This algorithm is shown to correctly reproduce the high-frequency hysteresis loop of both superparamagnetic and ferromagnetic NPs without any ad hoc or artificial parameters. The algorithm is easily parallelizable with a good speed-up behavior, which considerably decreases the calculation time on several processors and enables the study of assemblies of several thousands of NPs. The specific absorption rate (SAR) of magnetic NPs dispersed inside spherical lysosomes is studied as a function of several key parameters: volume concentration, applied magnetic field, lysosome size, NP diameter, and anisotropy. The influence of these parameters is illustrated and comprehensively explained. In summary, magnetic interactions increase the coercive field, saturation field, and hysteresis area of major loops. However, for small amplitude magnetic fields such as those used in magnetic hyperthermia, the heating power as a function of concentration can increase, decrease, or display a bell shape, depending on the relationship between the applied magnetic field and the coercive/saturation fields of the NPs. The hysteresis area is found to be well correlated with the parallel or antiparallel nature of the dipolar field acting on each particle. The heating power of a given NP is strongly influenced by a local concentration involving approximately 20 neighbors. Because this local concentration strongly decreases upon approaching the surface, the heating power increases or decreases in the vicinity of the lysosome membrane. The amplitude of variation reaches more than one order of magnitude in certain conditions. This transition occurs on a thickness corresponding to approximately 1.3 times the mean distance between two neighbors. The amplitude and sign of this variation is explained. Finally, implications of these various findings are discussed in the framework of magnetic hyperthermia optimization. It is concluded that feedback on two specific points from biology experiments is required for further advancement of the optimization of magnetic NPs for magnetic hyperthermia. The present simulations will be an advantageous tool to optimize magnetic NPs heating power and interpret experimental results.