[Dose-response relationship in radiotherapy: an evidence?].

The dose-response relationship is a fundamental basis of radiobiology. Despite many clinical datas, difficulties remain to demonstrate a relation between dose and local control: relative role of treatment associated with radiation therapy (surgery, chemotherapy, hormonal therapy), tumor heterogeneity, few prospective randomized studies, uncertainty of local control assessment. Three different situations are discussed: tumors with high local control probabilities for which dose effect is demonstrated by randomized studies (breast cancer) or sound retrospective datas (soft tissues sarcomas), tumors with intermediate local control probabilities for which dose effect seems to be important according to retrospective studies and ongoing or published phase III trials (prostate cancer), tumors with low local control probabilities for which dose effect appears to be modest beyond standard doses, and inferior to the benefit of concurrent chemotherapy (lung and oesophageal cancer). For head and neck tumors, the dose-response relationship has been explored through hyperfractionation and accelerated radiation therapy and a dose effect has been demonstrated but must be compared to the benefit of concurrent chemotherapy. Last but not least, the development of conformal radiotherapy allow the exploration of the dose response relationship for tumors such as hepatocellular carcinomas traditionally excluded from the field of conventional radiation therapy. In conclusion, the dose-response relationship remains a sound basis of radiation therapy for many tumors and is a parameter to take into account for further randomized studies.