Adjuvant radiotherapy for DCIS.

Sir—The investigators of the European Organization for Research and Treatment of Cancer (EORTC) (Feb 12, p 528) on breast-conserving surgery with or without radiation for treatment of ductal carcinoma in situ (DCIS) point out that if their finding that there is an increased rate of cancers in the contralateral breast in the surgery plus radiation group was truly radiation induced, it might well cancel out (and more) the clinical gains afforded by adjuvant radiotherapy. There is convincing evidence that this apparent increase in contralateral breast cancer is unlikely to be radiation induced. Specifically, we have used data from the surveillance, epidemiology, and end results (SEER) tumour registry for a cohort analysis of 32 000 women with DCIS treated between 1973 and 1993, directly comparing second-cancer risk in the women who did, or did not, receive radiotherapy. This large number allows good statistical power to test the hypothesis that radiotherapy induces an increased rate of second cancer in patients with DCIS. Details of the SEER DCIS subcohorts are shown in the table. For comparison, in the EORTC study, the mean age at DCIS diagnosis was 53 years and the mean follow-up was 4·3 years. The techniques used to estimate the relative risk of second cancer in the radiation versus the no-radiation SEER DCIS subcohorts are described elsewhere. The relative risks for radiation versus no radiation were estimated using Mantel-Haenszel Poisson models, adjusting for age at, and calendar year of, DCIS diagnosis, and follow-up time. For second malignancies in the contralateral breast, the adjusted relative risk for radiation versus no radiation was 0·98 (95% CI 0·80–1·19); when only long-term (>10 years) survivors are taken into account, the adjusted relative risk was 1·07 (0·54–2·12). For all second malignancies, the adjusted relative risk was 1·00 (0·85–1·18). None of the adjusted relative risks were significant. These null results are consistent with the corresponding US study on DCIS treatment, in which no increase in contralateral breast cancers was observed in the radiotherapy group. The mean radiation doses to the contralateral breast are probably comparable in the EORTC and US studies, in that wedge compensators were used in the EORTC protocol, and in the US protocol half-beam blocks and, optionally, wedges were used. These null results are also consistent with estimates of radiationinduced breast cancer derived from studies of atom-bomb survivors. For example, a fractionated radiation dose of 2 Gy (an upper-end estimate of the average contralateral breast dose) to the breast of a white woman aged 55 years gives a predicted lifetime breast-cancer risk of 0·33%—a relative risk of 1·03 compared with the background lifetime risk of about 11%. The increased cancer incidence in the contralateral breast reported in the radiation arm of the EORTC DCIS study is unlikely to be a consequence of radiation exposure to the contralateral breast, and is more likely to be an artifact of confounding variables. Any gains afforded by adjuvant radiotherapy in treating DCIS are unlikely to be ameliorated by an increase in second cancers.