The microdosimetric interaction model simulates individual interactions of alpha particles with nuclei of sensitive target cells in human bronchial airway generations 2, 4, 6 and 10. For a normalized source density of Po 218 and Po 214 alpha particles, number of hits, microdosimetric spectra and related parameters for cell nuclei located at various depths in bronchial epithelium are calculated. The lineal energy spectrum is a spherical nuclear target is then converted into probabilities for cell killing, mutation and transformation by multiplying the single event chord length distribution with event specific effect probabilities per unit track length as a function of LET. These effect probabilities are finally weighted by the depth-density distributions of basal and secretory cells. The predicted transformation probability is compared with other physical indicators of lung cancer risk, such as dose, dose equivalent, biologically weighted dose, mean lineal energy, and number of alpha particle hits.