Since the first radio occultation measurements were performed to study planetary atmospheres, the calculations returning refractivity profiles always adopted the so-called "spherical symmetry assumption". This hypothesis imposes a null horizontal gradient of the atmospheric parameters within the region sampled during the occultation: a scenario that can be quite unrealistic for regions in proximity of the terminator line. This study suggests that another way of calculating refractivity profiles is doable, dropping the spherical symmetry assumption and assuming a straight line propagation of the radio signal through the atmospheric medium (which, according to the laws of refraction, should travel in a curved line, corresponding to the fastest trajectory according to Fermat's law). The study proved that the theoretical difference between the two methods is negligible (within 1%) and that numerical methods can be developed to account for the asymmetric regions within the atmosphere.