Histopathological risk factors for ipsilateral breast events after breast conserving treatment for ductal carcinoma in situ of the breast--results from the Swedish randomised trial.

Aim: The primary aims were to study risk factors for an ipsilateral breast event (IBE) after sector resection for ductal carcinoma in situ of the breast (DCIS) in a trial comparing adjuvant radiotherapy to no therapy and to assess predictive factors for response to radiotherapy. Secondary aims were to analyse reproducibility of the histopathological evaluation and to estimate correctness of diagnosis in the trial.

Setting: A randomised trial in Sweden (the SweDCIS trial), including 1046 women with a median of 5.2 years of follow-up in a population, offered routine mammographic screening.

Methods: A case-cohort design with a total of 161 cases of IBE (42 of those being members of the subcohort) and 284 sampled for the sub-cohort. Ninety five percent of the participants' slides could be retrieved and were re-evaluated by three experienced pathologists.

Results: Low nuclear grade (NG 1-2) and absence of necrosis halves the risk of IBE in both irradiated and non-irradiated patients. Lesion size, margins of excision and age at diagnosis did not modify these associations. The presence of necrosis modified the effect of radiotherapy: relative risk was 0.40 with necrosis present and 0.07 with necrosis absent (p-value for interaction 0.068). In all subsets of prognostic factors, radiotherapy conferred a substantial benefit. The risk factors for in situ and invasive IBE were similar. The agreement between pathologists was moderate (kappa=0.486). Correctness of diagnosis in the subcohort of SweDCIS was 84.8%.

Conclusion: Although nuclear grade and necrosis carry prognostic information, we could not define a group with very low risk after sector resection alone. Radiotherapy has a protective effect in all substrata of risk factors studied. The interaction between the presence of necrosis and radiotherapy is a clinically and biologically relevant research area.