The effectiveness of cancer radiotherapy is compromised by the small proportion (approximately 5%) of patients who sustain severe normal tissue damage after standard radiotherapy treatments. Predictive tests are required to identify these highly radiosensitive cases. Patients with the rare, recessively inherited, cancer-prone syndrome ataxia-telangiectasia (A-T) sustain extremely severe normal tissue necrosis after radiotherapy and their cultured cells are also highly radiosensitive. Clinically normal carriers (heterozygotes) of the A-T gene have an increased risk of breast cancer, account for approximately 4% of all breast cancer cases and show a modest increase in cellular radiosensitivity in vitro. It has been suggested that a substantial proportion of highly radiosensitive (HR) breast cancer patients may be A-T heterozygotes, and that screening for mutations in the A-T gene could be used as a predictive test. We have tested this hypothesis in a group of cancer patients who showed adverse reactions to radiotherapy. Sixteen HR breast cancer patients showing mainly acute reactions (and seven HR patients with other cancers) were tested for ATM mutations using the restriction endonuclease fingerprinting assay. No mutations typical of those found in obligate A-T heterozygotes were detected. If the estimate that 4% of breast cancer cases are A-T gene carriers is correct, then ATM mutations do not confer clinical radiosensitivity. These early results suggest that screening for ATM mutations in cancer patients may not be of value in predicting adverse reactions.