The major health risk associated with low photon energy imaging is thought to be the induction of cancer as a consequence of the radiation exposure and this is the focus of this paper. Low photon energy imaging typically involves exposure to a low dose (<50 mGy) of low linear energy transfer (LET) radiation delivered at high dose-rate. Since epidemiologic data cannot provide an accurate assessment of risk at the doses used in imaging, risk estimates are currently made by fitting a linear response to intermediate and high dose data for cancer induction in radiation-exposed human populations. This method assumes a linear no-threshold (LNT) response and implies that no dose of radiation is safe. This assumption is not borne out by many laboratory studies of cancer-related endpoints that would suggest that the risk at low doses is much less than would be estimated from linear extrapolation from intermediate to high doses. It is also well recognised that the dose-response from many epidemiologic studies could equally well be fit by threshold models. Through the study of radiation-induced neoplastic transformation in vitro J-shaped dose-response curves for a variety of low LET radiations, including those used in low photon energy imaging, have been demonstrated. The relative risks calculated from this data compare remarkably well with those for breast cancer and leukemia incidence in radiation-exposed populations. From this it is concluded that the LNT hypothesis is likely to overestimate the risk of cancer induction by low photon energy imaging, at least for certain tumors.