Cancer risks from low dose exposure to ionising radiation – Is the linear no-threshold model still relevant?

A review of current knowledge of the biological consequences of diagnostic radiography is well overdue. Despite the monumental investment of time and effort by epidemiologists and biologists over the last 60 years, the ability of low doses of ionising radiation to cause cancer has not been proven. While there is little doubt that serious stochastic and deterministic consequences exist for moderate to large doses, the appropriateness of extrapolating cancer risks to low doses using the linear no-threshold (LNT) model is debatable. Current epidemiological evidence only has sufficient statistical power to detect excess malignancies above around 100 millisieverts (mSv). The lack of detectable excesses below this level could be due to either insufficient statistical power, or genuine lack of carcinogenic potential. The matter has been further complicated by the discovery of various cellular processes including bystander effects, hypersensitivity and adaptive responses, none of which are well understood. A substantial weight of evidence is required to produce a paradigm shift in radiation protection. At present there is insufficient evidence to allow complete rejection of the LNT model, although it must now be acknowledged that the concept has serious limitations.