The aim of this work was to estimate risk of lung tumour occurrence after inhalation of actinide oxides from published studies and rat studies in progress. For the same delivered dose, the risk increases when homogeneity of irradiation increases, i.e., the number of particles deposited after inhalation increases (small particles and (or) low specific alpha activity). The dose-effect relationships reported appear linear up to a few gray, depending on the aerosol considered, and then the slope decreases. This slope, which corresponds with the risk, can vary over one order of magnitude depending on the aerosol used. An effective threshold at about 1 Gy was not observed for the most homogeneous dose distributions. A dosimetric and biological approach is proposed to provide a more realistic risk estimate.