Using an empirical model for the beam-hardening effect, obtained through calibration, in an iterative reconstruction method enables a robust correction of beam-hardening artifacts in low-dose small animal studies independently of the bone distribution.