Tuesday April 3 2012
Breast cancer screening research highlights doubts on risk
“For every life saved by breast cancer screening up to 10 women ‘have unnecessary treatment’,” the Daily Mail reported today.
The news is based on a study that looked at the impact of screening in Norway, and how often it led to “overdiagnosis”. Overdiagnosis is the phenomenon where women are correctly diagnosed but given treatment even though it will not improve their survival. This could be because the tumour identified would not have progressed to a clinical stage (would never have caused symptoms) or because the woman would have died from another cause, such as old age, before the cancer reached a clinical stage. Using nearly 20 years of data, researchers estimated that Norway’s screening programme had an overdiagnosis rate of 15–25%. This equated to 6 to 10 women being overdiagnosed for every 2,500 women invited for screening. In comparison, 20 in 2,500 women screened were diagnosed with breast cancer that was not overdiagnosed. Screening 2,500 women was estimated to prevent one death from breast cancer.
The issue of screening is of key interest to many people, and this study provides further data that could be relevant to the UK's current review of breast screening, which was announced in late 2011. This independent review will look at both screening in the UK and the type information that is given to women when deciding whether screening is right for them.
Where did the story come from?
The study was carried out by researchers from Harvard School of Public Health, Bringham and Women’s Hospital, and Harvard Medical School in the US, the Karolinska Institute in Sweden, and Oslo University Hospital Rikshospitalet in Norway. It was funded by the Norwegian Research Council and Frontier Science.
The study was published in the peer-reviewed medical journal Annals of Internal Medicine.
This story was widely reported, and most news sources provided accurate coverage of the research.
What kind of research was this?
There are several key considerations surrounding the use of routine breast cancer screening. These are primarily the need to ensure it can detect cancers and save lives, and that it does not cause unnecessary invasive treatments. In particular, there is currently a great deal of debate over the extent to which screening leads to overdiagnosis. This is where a woman is diagnosed and treated with no survival benefit, either because the tumour identified would not have progressed to a clinical stage or because the woman would have died from other causes before the cancer reached a clinical stage.
In late 2011, it was announced that the UK’s breast screening policy would be independently reviewed after some people questioned the overall benefit of the programme. The review will be performed by independent researchers but led by Professor Richards of the Department of Health and Harpal Kumar, chief executive at Cancer Research UK. The publication date for the review has not yet been announced.
This research compared the incidence of invasive breast cancer in women with and without screening. It estimated the percentage of overdiagnoses that might be attributed to mammography screening. It did this by comparing the number of cases of breast cancer found in screened women to the number of cases in matched unscreened women in Norway, where mammography screening was rolled out over a decade by geographical region. The researchers took into account the improvements in breast cancer care and changes in new breast cancer rates that have occurred over time.
The researchers stated that a randomised controlled trial with a lifelong follow-up of participants assigned to screening or a control group with no screening would be the ideal way to estimate the rate of overdiagnosis in research terms, but that due to practical reasons such a trial is unlikely ever to be performed. In this study, the researchers tried to take into account the many variables that affect the calculation of overdiagnosis, including changes in breast cancer incidence over time and “lead time”, the variation in length of time when a breast cancer can go undetected, between the groups.
What did the research involve?
The researchers retrieved details on the female population from the Statistics Norway national database, and information on all women with a first diagnosis of invasive breast cancer between January 1 1986 and December 31 2005 from the Cancer Registry of Norway.
The women were divided into four groups:
- current screening group – a group of women living in counties with screening between 1996 and 2005
- historical screening group – women living in these counties between 1986 and 1995, before screening
- current non-screening – women diagnosed in counties without screening between 1996 and 2005
- historical non-screening – women living in these counties between 1986 and 1995
By comparing current groups against historical groups, the researchers accounted for changes in breast cancer incidence over time. The researchers then used two different approaches to account for lead time, where an earlier diagnosis is made when women are younger in the screened group. The researchers also analysed the results by stage of cancer.
What were the basic results?
The researchers included 39,888 patients with invasive breast cancer diagnosed over twenty years, between 1986 and 2005. There were 18,708 cases in women aged between 50 and 69 years old, the age range invited for screening, and 7,793 of these women were diagnosed after the screening programme started, making them part of the “current screening group”.
Combining the two statistical approaches, the researchers calculated that the estimated rate of overdiagnosis due to the mammography screening programme was 15–25%. In other words, this proportion of women were diagnosed with breast cancer and potentially received treatment unnecessarily. They also estimated that over the 10 years of biennial mammography screening in Norway, for every 2,500 women invited for screening:
- 6 to 10 women were overdiagnosed
- 20 women were diagnosed with breast cancer that was not overdiagnosed
- 1 death from breast cancer was prevented
How did the researchers interpret the results?
The researchers concluded that “mammography screening entails a substantial amount of overdiagnosis”.
The benefits and risks of breast cancer screening have come under much scrutiny in recent years, not just from the press and the public, but also within the medical community. Many of the key arguments centre on the rates of legitimate treatment and treatment that would have otherwise proved unnecessary following screening. The issue is highly complex as it relies on finding high-quality data that can reliably indicate what would happen with the presence and absence of screening in a particular population.
This study looked at screened and unscreened segments of the Norwegian population to estimate the percentage of overdiagnoses of invasive breast cancer due to mammography screening. Overdiagnosis occurs where a woman is diagnosed and treated with no survival benefit even though the diagnosis is correct. This can be because the tumour identified would not have progressed to a clinical stage (i.e. it would never have caused symptoms) or because the woman would have died from other causes before the cancer reached a clinical stage.
The authors estimated the rate of overdiagnosis due to mammography screening as 15–25%. They calculated that for every 2,500 women invited for screening, 6 to 10 women are overdiagnosed, 20 women are diagnosed with breast cancer that is not overdiagnosed, and 1 death from breast cancer is prevented.
This study has some limitations. It was based on data taken from registries rather than gathered to specifically assess screening. Also, the researchers could not adjust for all factors that could have influenced the results, some of which might have offered alternative explanations for the results. In addition, another type of breast cancer called ductal carcinoma in situ was not studied.
However, despite its limitations, the study’s findings are similar to those in other recent publications, including some that have prompted a UK review of breast cancer screening and the type of information offered to women. This review is currently being conducted, although its expected completion date has not yet been announced.
Analysis by Bazian