Approximately 1 in every 600 women attending breast-screening programmes in the United Kingdom is diagnosed with breast carcinoma in situ (BCIS). However, there is little information on the occurrence of subsequent cancers (other than second breast cancers) in these women. We investigated the occurrence of invasive cancers in 12 836 women diagnosed with BCIS in southeast England between 1971 and 2003, using data from the Thames Cancer Registry. A greater than expected number of subsequent cancers was found for two sites: breast (standardised incidence ratio (SIR) 1.96; 95% confidence interval (CI) 1.79 -2.14) and corpus uteri (SIR 1.42; 95% CI 1.11 -1.78). For subsequent ipsilateral breast cancer in those treated with breast conservation, the excess was independent of the time since diagnosis of BCIS, whereas for subsequent contralateral breast cancer, there was a steady decline in excess over time. For subsequent uterine cancer, the excess became statistically significant only at 45 years after BCIS diagnosis, consistent with a treatment effect. This was further supported by Cox regression anaysis: the risk of subsequent uterine cancer was significantly increased in women receiving hormonal therapy compared with those not receiving it, with a hazard ratio of 2.97 (95% CI 1.84 -4.80). (2008) Currently, approximately 1 in every 600 women attending breastscreening programmes in the United Kingdom is diagnosed with breast carcinoma in situ (BCIS). This amounted to almost 3000 cases in 2004 (NHS Cancer Screening Programmes, 2006. The most common form of treatment is lumpectomy, with or without radiotherapy, although women with a more diffuse pattern of BCIS may undergo mastectomy. A number of previous studies have focused on the risk of second breast cancer following a diagnosis of BCIS (Habel et al, 1997;Wärnberg et al, 2000;Claus et al, 2003;Levi et al, 2005;Rawal et al, 2005), with estimates of relative risk generally of the order of two-to fivefold. Few studies have looked at the occurrence of subsequent invasive cancers at other sites (Ward et al, 1992;Franceschi et al, 1998;Soerjomataram et al, 2006), and these were based on relatively small numbers of cases.
British Journal of CancerIn the current population-based study, we have investigated the occurrence of subsequent invasive cancers in a very large series of unselected women with BCIS in southeast England. We have also examined the relationship between treatment for BCIS and the development of subsequent cancers.
MATERIALS AND METHODSAll recorded cases of BCIS (ICD10 code D05) diagnosed in women between 1971 and 2003 were extracted from the Thames Cancer Registry (TCR) database. The TCR is a population-based registry covering London and a large part of southeast England. Patients with any other cancers diagnosed prior to, or at the same time as, BCIS were excluded in order to avoid confounding the sequelae of BCIS with those of other cancers, particularly breast cancer. The occurrence of subsequent cancers was analysed by calculating the standardised inc...