Overdiagnosis is the inevitable flip side of early detection resulting in unnecessary labelling of well women with a diagnosis of cancer and possible unnecessary treatment. Overdiagnosis occurs because breast cancers have different rates of growth and slow-growing cancers are preferentially detected by screening. Some of these slow-growing screen-detected cancers may never have been clinically apparent during an individual's lifetime. Evaluating the benefits and risks of screening are complex, but this has been performed for the UK population by an independent review led by Professor Marmot. It might be possible to limit overdiagnosis by identifying women with "low-risk disease" earlier, either at the point of screening when additional investigations could be delayed (possibly for ever) so that they are not subjected to additional diagnostic tests, or at the point of diagnosis. Both these options would require major re-education of clinicians and the public who would need to accept that screening is "deliberately ignoring a cancer". There is a long surgical history of reducing the burden of treatment, which continues today with trials of management of the axilla and reducing or even omitting radiotherapy for low-risk disease. The Low Risk Ductal Carcinoma In Situ trial (LORIS) has started to identify a group of breast cancer patients who could avoid surgery and be offered active monitoring. We need to consider planning a similar trial for low-risk invasive breast cancer.