Overdiagnosis of screen‐detected breast cancer

Screen‐detected breast cancer overdiagnosis occurs, but each woman has been diagnosed with cancer that cannot be ignored

There are an increasing number of publications estimating the extent of cancer overdiagnosis, which for breast cancer is in the context of population cancer screening programs.1 Researchers investigating overdiagnosis point to a range of related harms, but it is important to view these in the context of screening benefits, such as reductions in risk of breast cancer death.2 Care needs to be taken not to conflate formal screening programs with informal or opportunistic approaches to early detection, such as prostate‐specific antigen (PSA) testing in prostate cancer. This article focuses on the risk of overdiagnosis in the context of population‐based breast screening programs, given that overdiagnosis is often at the heart of calls to cease mammographic breast cancer screening.3,4 Despite the emphasis often given to breast cancer screening in discussions of overdiagnosis, the concept should not be regarded as only applying to breast cancer screening, or to cancer screening more generally, but as an outcome that could apply, to varying degrees, to a wider range of screening and diagnostic practices.

Defining overdiagnosis

Overdiagnosis of a cancer is not a false positive or misdiagnosis; it is a diagnosis with histological verification of a cancer that would otherwise not have gone on to cause morbidity or death — although it cannot be determined at the time of diagnosis whether the cancer would have progressed to cause morbidity or death.5 An overdiagnosed cancer is in part a consequence of our capacity to diagnose cancers at increasingly earlier stages. It depends on competing causes of death; that is, a cancer will not cause morbidity or death in people who die beforehand from other causes, such as respiratory or cardiac diseases or trauma. The reality that a proportion of cancers will therefore be overdiagnosed is inherent in all screening programs, although the issue is not limited to screening. Major international reviews have concluded, after a careful evaluation of the balance between benefits and harms, that there is a net benefit from inviting women to receive breast screening (ie, benefits outweigh harms).2

The problem of overdiagnosis

Concerns about overdiagnosis stem from the potential harms that may be experienced by a person receiving the overdiagnosis. Harms can range from the psychological stress of receiving a diagnosis through to the potential for complications and adverse effects of diagnostic procedures or treatments. However, the challenge is that for any individual, it is not possible at diagnosis to determine whether their cancer is overdiagnosed or not. The cancers that are overdiagnosed are indistinguishable from other cancers histologically. As this is a post mortem classification, cancers can only be classified as overdiagnosed when another cause of death supervenes.

Estimating overdiagnosis and mortality benefits

Estimates of overdiagnosis within breast cancer screening programs vary widely, and this is in part due to methodological as well as programmatic differences. Recent Australian modelling suggested that the rate of overdiagnosis across five cancers (breast, prostate, renal, thyroid and melanoma) was 18% in women and 24% in men.1 However, of these cancers, only breast cancer is part of population screening in Australia. Other studies of breast cancer overdiagnosis specifically point to much lower levels of overdiagnosis. Based on British and European reviews, Cancer Australia has estimated that for every 1000 Australian women screened for breast cancer every 2 years from age 50 to 74 years, around eight breast cancers (range, 2–21) may be found and treated which would not otherwise have been found in a woman’s lifetime.6 In addition, an equivalent number of breast cancer deaths would be avoided in these women.6 The European Screening Network (EUROSCREEN) Working Group calculated a summary estimate of overdiagnosis as about 6.5% of the expected number of diagnosed breast cancers (range, 1–10%) in screened women, based on data from studies in Europe.2 Overall, data from around the world indicate that breast screening confers an estimated reduction in breast cancer mortality of 23% in women invited for screening and 40% or more among those women who are screened.2

Cancer screening programs need to balance benefits and harms

Decisions about whether to implement screening at a national level in any country should follow a comprehensive assessment of likely benefits at a population level (mortality reduction, delivery of more conservative therapy to people diagnosed with cancer via screening), harms (unnecessary treatments, psychological impacts) and costs (health service, individual out‐of‐pocket expenses, societal costs). In Australia, this principle is encapsulated in the Australian Population Based Screening Framework.7

It is well understood that some cancers are slower growing while others are more aggressive, but there are significant limits to our capacity to determine at diagnosis these characteristics at both an individual tumour and patient level. The concern, therefore, is that the potential to discourage women from breast screening through concerns of overdiagnosis would result in harms associated with later diagnosis, including deaths from breast cancer. Based on current international evidence, if left untreated, more than 90% of cancers found through routine screening would progress and become symptomatic and be potentially lethal, depending on modelling assumptions.2 Even drawing from the higher estimates of overdiagnosis, data would support that more than 70% of screen‐detected invasive cancers would progress to become symptomatic without treatment.8 These estimates are indicative of the scale of risk of developing a symptomatic breast cancer which could be detected earlier through screen detection. Screen‐detected cancers are found at an earlier stage and tend to be smaller; treatment guidelines advise less extensive surgery and reduced need for adjuvant treatments, as well as being associated with improved survival.9

Reducing the harms of overdiagnosed cancers

The most readily recognised harm of overdiagnosis is overtreatment. However, overtreatment is distinct from overdiagnosis and the effects can be mitigated by promotion of evidence‐based clinical management guidelines. Significant advances have been made in tailoring treatment for breast cancer. These include advances in surgery and radiotherapy and using tumour characteristics such as oestrogen receptor, progesterone receptor, and HER2 status to tailor systemic therapies. Research is also being undertaken to investigate using active monitoring rather than surgery for ductal carcinoma in situ, as there is debate regarding the potential for this type of carcinoma, if left untreated, to progress to invasive cancer. Research is underway to determine if genomic or other molecular signals in tumours will provide clearer indications of which ductal carcinoma in situ and invasive cancers need treatment, including chemotherapy, and at what level of aggressiveness, irrespective of whether found through screening or other means.10 Advances from this research will better support women to make informed decisions about treatment.

Population screening programs are offered within a policy framework that carefully considers the target population that stands to benefit from screening, including age criteria, and ongoing monitoring and reporting of sensitivity, specificity and interval cancers.7 There are also ongoing research efforts to improve the effectiveness of breast screening, including evaluation of new approaches to tailor screening to the individual woman’s risk profile to maximise benefit and minimise harms. Informed consent about breast screening needs to balance the potential harms with the demonstrated benefits of the current national screening program.