Radiotherapy is an important weapon in the treatment of cancer, but adverse reactions developing in the co-irradiated normal tissue can be a threat for patients. Early reactions might disturb the usual application schedule and limit the radiation dose. Late appearing and degenerative reactions might reduce or destroy normal tissue function. Genetic markers conferring the ability to identify hyper-sensitive patients in advance would considerably improve therapy. Association studies on genetic variation and occurrence of side effects should help to identify such markers. This survey includes published studies and novel data from our own laboratory. It illustrates the presence of candidate polymorphisms in genes involved in the cellular response to irradiation which could be used as predictive markers for radiosensitivity in breast or prostate cancer patients. For other tumor types such as head and neck cancers or brain tumors, the available data are much more limited. In any case, further validation of these markers is needed in large patient cohorts with systematically recorded data on side effects and patient characteristics. Genetic variation contributing to radiosensitivity should be screened on a broader basis using newly developed, more comprehensive approaches such as genome-wide association studies.