The impact of radiation on early breast carcinoma survival. A Bayesian analysis

A Bayesian statistical approach was used to examine the effect of local control on survival of patients with early breast carcinoma.

The effect of radiation on overall survival was examined in the four published randomized trials on conservation surgery with or without radiation: the Uppsala-Orebro, Canadian, NSABP-B06, and Milan III trials. Classical and Bayesian statistical approaches were used to evaluate the effect of sample size and follow-up on the results.

Combined 5-year survival results of patients with negative lymph nodes and surgical margins in the Uppsala-Orebro, Canadian, and NSABP-B06 trials indicated a 79% probability of a positive benefit with the addition of radiation and a 9.6% relative reduction in the annual mortality rate in favor of the irradiated patients. For the same subgroup of patients in the NSABP-B06 trial, the probability of a positive benefit with radiation increased from 65% at 5 years to 87% at 10 years, and is reflected in the change of the relative reduction in annual mortality from 8.2% at 5 years to 17.5% at 10 years. For all patients who accepted randomized treatment in the NSABP-B06 trial, the probability of a positive survival benefit with radiation was 80%, 98%, and 91% at 5, 10, and 12 years, respectively, with corresponding relative reductions in the annual mortality rates of 12.8%, 20.5%, and 15.2%.

The addition of radiation to lumpectomy offers a small but important survival advantage, even in patients with negative surgical margins and lymph nodes. Because locoregional control and breast preservation are significantly improved with radiation, even a small survival advantage substantiates the importance of radiation in the treatment of patients with early breast carcinoma.