[Hypofractionation and radiotherapy: "the eternal return"].

Hypofractionation is not a new idea in radiotherapy. The use of a few high-dose fractions has been proposed by some pioneers of our specialty in the early years of the 20th century. Hypofractionation then reappeared several times in the next decades, based on successive radiobiological concepts, a number of them having been shown to be wrong. The nominal single dose (NSD), for example, so fashionable in the 1970's, dramatically underestimated the late toxicity of the high-dose fractions. Consequently, the NSD was directly responsible for a significant increase of the incidence and of the severity of late complications in large cohorts of patients. The linear-quadratic model (LQ) unequivocally improved our understanding of fractionation sensitivity, but one has to keep in mind its limitations, both in the areas of low and high doses per fraction. For more than a decade, prostate cancer has been the subject of fierce discussions about its sensitivity to fractionation. A number of studies have suggested an unusually low (for a malignant tumor) alpha/beta ratio. However, the available data do not allow a precise evaluation of this ratio; "very low" (1.5 Gy), with an advantage of hypofractionation in terms of local control? Or simply "low" (3-4 Gy), only allowing a reduction of the total number of fractions (with a dose adequately reduced)? While waiting for complementary data, it is advised to remain very careful when modifying the classical schemes towards hypofractionation.