Friday June 27 2014
Blood tests are ineffective for most cases of breast cancer
"Blood test could give early warning of breast cancer," The Guardian reports. Researchers have identified a genetic signature that may be useful in predicting whether a woman is likely to develop non-inherited breast cancer.
Reliable blood tests for inherited (genetic) cases of breast cancer already exist. These tests look for mutations in the BRCA gene and may be used in women with a strong family history of breast cancer.
They can predict whether a woman is at risk of developing breast cancer with a high degree of accuracy. Such tests provide women with the opportunity to take preventative treatment, as was the case recently with the actress Angelina Jolie.
But only a very small proportion of all breast cancers are in people with BRCA mutations, described as being fewer than 1 in 10.
New research has focused on looking at DNA methylation of the BRCA1 gene. Methylation is when chemical groups attach to a gene. This does not amount to a mutation as the DNA sequence remains unchanged, but it can nevertheless alter gene activity.
In this study the researchers identified a set of DNA methylations that could provide a unique "signature" and potentially predict breast cancer risk in people without BRCA mutations.
This is interesting and promising research, but a reliable blood test for non-inherited breast cancer does not seem to be on the cards any time soon. The test showed potential, but it currently isn't very accurate – it is only marginally better than a guess at predicting risk.
Where did the story come from?
The study was carried out by researchers from University College London and the University of Manchester in the UK, the First Faculty of Medicine and General University Hospital, Charles University in the Czech Republic, and the Shanghai Institute for Biological Sciences in China.
The research received various sources of funding, including the European Union's Seventh Framework Programme.
It was published in the peer-reviewed medical journal Genome Medicine, an open access publication available free online. This is a provisional publication and there may be some revisions before its final version.
The quality of the UK media's reporting of the study is mixed. Most sources included useful discussion from experts, who on the whole are encouraged by the findings and discuss the need to build on the results.
However, the media does not make it clear that no such screening or diagnostic test is imminent. Much more work is needed before developing such a test can be considered or before it could be brought into widespread use.
Such considerations would need to include the possibility of false negative and false positive test results. A false negative result is where a woman is told she isn't at risk of breast cancer when she is, and a false positive tells her she is at risk of breast cancer when she isn't.
Even if a test correctly identified a woman as being at risk of breast cancer, the psychological consequences of this risk and what to do about it are great.
The Daily Mail also suggests the research has developed a test that can tell whether a woman's lifestyle puts her at risk of breast cancer. This is a misleading and incorrect interpretation of the study.
While it is plausible such a test could become part of a more detailed risk stratification process also taking lifestyle factors into account, no actual research was carried out looking at this.
What kind of research was this?
This was a case control study that aimed to see whether it was possible to develop a blood test that could predict whether a person was at risk of developing non-hereditary breast cancer.
The BRCA1 gene has long been associated with the risk of breast cancer. People who have inherited mutations of this gene are reported to have an 85% risk of developing breast cancer.
However, most people who develop breast cancer have not got a mutation of the BRCA1 gene (or the BRCA2 gene, which is similarly associated with risk). This makes it difficult to predict the risk of breast cancer among the majority of people, who have not inherited mutations of the BRCA genes.
The research centred on trying to identify people who have DNA methylation of the BRCA1 gene. This means they have not inherited a mutation of the BRCA1 gene, but the gene has a methyl chemical group attached to it.
Although the DNA sequence of the BRCA1 gene is "normal", this methyl group addition still alters the activity of this gene.
The case control study design involved analysing and comparing the blood samples of "cases", who carry a BRCA1 gene mutation, and "controls", who have normal BRCA1 and 2 genes.
What did the research involve?
The DNA from the blood samples of 72 people with a BRCA1 gene mutation and 72 people without BRCA gene mutations was extracted in the laboratory (specifically, their white blood cells were examined).
DNA methylation at specific CpG sites was examined (C and G being two of the four base molecules in the DNA sequence) in a model that took age and presence of cancer into account.
The researchers then verified the predictive accuracy of the methylation profiles using blood and tissue samples collected in two additional studies:
- the Medical Research Council National Survey of Health and Development (NSHD) – taking a birth cohort of men and women born in 1946, the researchers analysed the blood and buccal cell (inside of the cheeks) samples of 75 people who developed cancer (19 cases of breast cancer) and 77 who did not develop cancer
- the United Kingdom Collaborative Trial of Ovarian Cancer Screening (UKCTOCS) – the researchers analysed the blood samples from 119 postmenopausal women who went on to develop breast cancer (on average two years later) and 122 who remained cancer free during the average 12-year follow-up
What were the basic results?
The researchers observed a vast number of different methylations at the CpG sites in people who had BRCA1 gene mutation. From this they identified a "DNA methylation signature" that included 1,829 CpGs, which are different in people with and without BRCA1 gene mutations.
When they validated this in the NSHD study, they found the DNA methylation signature was a possible predictor of breast cancer. In statistical terms, the area under the curve for this signature was 0.65 (95% confidence interval [CI] 0.51 to 0.78).
The area under the curve – related to a curved line on a graph – is a measurement of predictive accuracy. When interpreting area under the curve, a value of 1.0 would be perfect accuracy, while a value of 0.5 would be considered random chance (the test would be little better than guessing).
Values of less than 0.5 would be a very poor diagnostic test, worse than guessing. Therefore, 0.65 may be indicated as a bit better than guessing. The DNA methylation signature also similarly predicted the risk of other types of cancer (area under the curve 0.62).
This predictive ability only worked when using the blood samples taken from this cohort – analyses using the buccal tissue sample didn't work.
When the researchers then looked at the UKCTOCS samples, they found the signature predicted the development of oestrogen receptor-positive breast cancer, with an area under the curve of 0.57 (suggesting this was a bit better than guessing).
In this cohort, they carried out further sub-analyses using data on breast cancer risk factors they had information on for this cohort. They found the DNA methylation signature predicted breast cancer (and breast cancer deaths) only in the group of women without a family history of breast cancer, but not in the women with a family history.
However, as the researchers acknowledge, the number of women who developed breast cancer and had a family history of the disease was very small.
How did the researchers interpret the results?
The researchers conclude that the DNA methylation signature derived from BRCA1 carriers "is able to predict breast cancer risk and death years in advance of diagnosis".
Importantly, they say that future studies may need to focus on DNA methylation profiles in body cells (rather than the white blood cells used here) to reach the area under the curve thresholds required of preventative measures or early detection strategies.
This is early-stage research into the development of an indicator that could predict the risk of breast cancer in people who are not carrying the hereditary BRCA gene.
This accounts for the vast majority of people who develop breast cancer. Only a very small proportion of all breast cancers are in people with BRCA mutations.
Although there are many different health and lifestyle factors associated with the risk of breast cancer, predicting breast cancer risk in people without hereditary gene mutations is not currently possible.
This laboratory research focused on examining white blood cells for DNA methylation of the BRCA1 gene, which does not alter its DNA sequence but nevertheless alters its activity. The research found a vast number of different methylations combined into a "DNA methylation signature" that could possibly predict risk of breast cancer.
However, in the current study, the predictive power of this test was not very good – it was better than guessing, but not much more. Such a poor predictive accuracy could never be used as the basis of decisions about treatment. It has also only been tested in fairly small cohort samples.
As it currently stands, a test that could predict breast cancer development in people without hereditary BRCA gene mutations does not seem to be on the cards any time soon. As the researchers acknowledge, given the low predictive accuracy when using the blood samples, they next need to test this using other body cells.
Much work is needed to improve the accuracy of this test. And even then, if researchers can one day develop an accurate test using methylation profiles, there are great considerations to make in weighing up the risks and benefits before a screening test could be introduced for widespread use.
For now, the most well-established lifestyle factors that reduce the risk of non-hereditary breast cancer are a healthy balanced diet, regular physical activity to avoid becoming overweight or obese, avoiding smoking, and limiting your alcohol intake.
Analysis by Bazian. Edited by NHS Choices. Follow Behind the Headlines on Twitter. Join the Healthy Evidence forum.