Cause-specific mortality in long-term survivors of breast cancer who participated in trials of radiotherapy.

PURPOSE: To examine long-term cause-specific mortality in patients irradiated for breast cancer as part of a randomized clinical trial. PATIENTS AND METHODS: We studied all available information from randomized trials initiated before 1975 in which radiotherapy was the randomized option and surgery was the same for both treatment arms. Eight such trials were identified. RESULTS: The increased all-cause mortality rate in 10-year survivors previously reported is no longer significant, although a numerical difference in favor of non-irradiated patients remains. This result was strongly influenced by the earliest trials, and more recent trials have found a nonsignificant net benefit in overall mortality associated with radiation therapy. An excess of cardiac deaths was apparent in both early and more recent trials (P < .001), but this was offset by a reduced number of deaths due to breast cancer, especially in more recent trials. CONCLUSION: The reduction of breast cancer deaths suggests that radiation therapy may have a value beyond the clearly established improvements obtainable for local control. Use of techniques that minimize cardiac dose is important in reducing the risks of adjuvant radiotherapy, especially in good-prognosis patients.