Radiotherapy is one of the principal modalities of cancer treatment, but the delivery of a curative dose of ionizing radiation (IR) to the tumour is frequently limited by the need to protect the normal tissues within the irradiated area from radiation damage. This problem could be circumvented if tumour cells could be selectively sensitized to killing by IR. One way to achieve this goal would be to transduce the tumour cells with expression vectors carrying toxin genes under the control of promoters that are inactive unless induced by IR. For this approach to be successful, two parameters must be met: (i) the expression vector has to be delivered to the tumour or its immediate vicinity (e.g. its vasculature) and (ii) the promoter driving the expression of the toxin gene has to have negligible basal activity, yet has to be activated by clinically-achievable doses of IR. Several vectors that fulfil these criteria are currently reaching clinical trials. In this review, we examine the response of mammalian cells to IR, and the current status of radiation-induced suicide gene therapy that is dependent on this response.