RESUMO
Progress in dosimetry and in radiobiology together with the evolution of radiation generators led to adopt very elaborated radiotherapy protocols. However, the therapist should have confidence in these protocols on condition that irradiation parameters are strictly respected. For obvious reasons, this is impossible despite the meticulous rules. In this, paper, we therefore intend to study the role of the angle of incidence on the actual distribution of doses; we also study the inaccuracy of certain parameters recorded during preparation for the therapy or shown during successive irradiation courses. In this paper we retain only errors which can be easily quantified. These are errors concerning the irradiation apparatus (source to skin distance, angle of the rays, field dimension) and those due to factors related to the patient treated: contour measurement, entrance point of rays and positional changes of the patient. The different errors can be associated in various ways for a given therapy protocol. For the same reason, an isodose of a given nominal value will have several curves. All these tracings are lined by two curves which we call "superior and inferior envelops": these border the "incertitude area". It can be said that at the end of the treatment, the given isodose lies in this area; it is however impossible to define its form or exact position. A few examples in the text illustrate these results and show the possible practical angles of incidence. If we assume that errors during successive irradiations are distributed at random, a certain compensation is noted and the "areas of incertitude" decrease. But if the errors are constantly on the same side, as is sometimes the case, the dose distribution can be very different from that predicted in the dosimetric protocol.