Data derived from histopathological analysis, natural history, radiological characteristics, molecular markers, and clinical outcome indicate that ductal carcinoma in situ (DCIS) is a heterogeneous disease, meaning that no one treatment strategy is best, but rather that treatment should be personalised and entail a systematic and rigorous multidisciplinary approach. Many women with DCIS will develop actual invasive carcinoma over time, whereas others-especially those with low-grade cancers-will not. At the moment, identification of patients at low risk of recurrence risk is very difficult (that is, such people for whom further treatment beyond lumpectomy is not needed). In this context, molecular profiling analysis is a promising method to guide management decisions. In this Review, various treatment strategies for DCIS will be reviewed, highlighting the limitations of randomised trials. Furthermore, discussions about the role of sentinel-node biopsy and tamoxifen in disease management; locoregional recurrence; and special clinical scenarios such as recurrent disease, DCIS after thoracic radiotherapy, ductal carcinoma with concurrent lobular carcinoma in situ, and DCIS in elderly people and in men will be presented.