Hope that blood test 'could diagnose five types of cancer'

“A new blood test that detects five different forms of cancer is one step closer to becoming a reality and could save millions of lives around the world,” the Mail Online reports. The test looks for abnormal changes in DNA – what is described as a DNA signature.

This laboratory research looked at ways to identify tumour DNA – DNA affected by abnormal cell growth – in blood samples and distinguish it from normal cellular DNA.

The researchers used tissue samples from five cancers – womb, lung, stomach, colon and breast tumours – and compared it with normal healthy tissue.

In brief, they found they could identify the cancerous tissue from a particular DNA signature around a certain gene (ZNF154).

Their tests reveal this test could be fairly accurate at detecting cancer at a concentration of 1% tumour DNA on a background of 99% normal DNA in a blood sample.  

There are many things to consider before any new screening or diagnostic test for cancer is introduced, especially with a “blanket screen” like this.

These issues include how and whether the test improves on current screening or diagnostic methods, as well as looking at the possible harmful effects, such as getting an incorrect positive screen result when you’re in fact cancer free, or getting an incorrect negative screen result when you have cancer.  

Where did the story come from?

This study was carried out by researchers from the National Human Genome Research Institutes in the US, and published in the peer-reviewed Journal of Molecular Diagnostics.

The researchers report no sources of financial support and no conflicts of interest.

The Mail Online’s reporting of the study is accurate, although its claim that, “a new blood test … could save millions of lives around the world” is prematurely optimistic: this research is in its early stages and has not been tested at a significant population level.

The Daily Telegraph’s headline is slightly more restrained: “Blood test to spot five deadly cancers could prevent thousands of deaths”. 

What kind of research was this?

This laboratory study examined a possible way of detecting DNA markers for cancer. The researchers report that work on cancer prevention, early diagnosis and treatment has reduced overall cancer death rates by 20% over the past 20 years.

Further advances in screening and diagnosis are, they say, where improvements in survival rates are likely to come. In many cases, the earlier a cancer is diagnosed, the better the outcome tends to be.

Tests that are able to detect genetic information coming from cancerous cells are a possible area for development. Previous research has shown how DNA from a tumour can be found freely circulating in the blood or in saliva, urine and stool samples, for example.

One approach is to look for what is called DNA methylation. This is a signalling method that controls gene activity in a cell, and genes are effectively “switched off”.

There are a few specific cancer tests that have already been developed that involve detecting DNA methylation – for example, detecting specific genetic markers for lung cancer in lung fluid, or bowel cancer in stool samples. However, this is still an area of development.

This study builds on the researchers’ previous work, where they identified a possible hypermethylation signal near a particular human gene (ZNF154).

This signal was found to come from ovarian and womb cancers and may be found in other cancers, too. This study measured the ZNF154 methylation signal across five different cancers. 

What did the research involve?

The researchers examined cell samples from womb, lung, stomach, colon and breast tumours, and comparison samples of normal tissues from the same organs.

In total, they examined 184 tumour samples and 34 normal tissue samples. They used complex laboratory techniques to analyse cancerous DNA methylation patterns and examine them on a background of normal DNA methylation patterns.      

The researchers then used their findings to identify possible classification methods that could be used in cancer screening. They looked at different ways to characterise methylated bases – the “letters” of DNA (A, C, G and T) – and identified features that could be used to distinguish cancerous tissue from normal tissue.

They then used computational simulation to indicate how reliable these features could be for classifying samples as tumours or normal tissue at various concentration levels, given that in a blood sample, for example, tumour DNA may be present at quite dilute levels.   

What were the basic results?

The researchers found all of the tumour types tested demonstrated hypermethylation at the ZNF154 gene site compared with the normal tissue.

The classification method with the best performance had almost perfect accuracy for distinguishing between normal and cancerous tissue.

Their computational simulation indicated circulating tumour DNA could be detected at a dilution of only 1% tumour DNA on a background of 99% normal DNA. 

How did the researchers interpret the results?

The researchers concluded their findings “suggest that hypermethylation of the ZNF154 [gene site] is a relevant biomarker for identifying solid tumour DNA and may have utility as a generalisable biomarker for circulating tumour DNA”. 

Conclusion

This is very early-stage laboratory research that aimed to explore new avenues that could detect and diagnose cancer earlier – and hopefully ultimately lead to earlier and more successful treatment, and so better cancer survival rates. 

The study indicates taking blood samples and detecting DNA methylation from tumours could be one possible early screening or diagnostic method, and shows this technique’s use for indicating womb, lung, stomach, colon and breast tumours.

However, there are likely to be many more stages of research necessary to build on these findings and check how reliable the test could be for different subtypes of these cancers, and also whether it could be used for other types of cancer.

Even then, there are many things to be taken into account before considering introducing any new screening or diagnostic test for cancer, including how and whether it improves on current screening or diagnostic methods.

For example, the media has highlighted the benefits of a blood test being “non-invasive”, but current screening tests for bowel and breast cancer – taking stool samples and using mammograms, for instance – are also non-invasive.

Possible harmful effects also need to be considered, such as getting an incorrect positive screening result when you’re in fact cancer free (false positive), or getting an incorrect negative screen result when you do have cancer (false negative). There is also the question of whether screening for certain cancers could seem to lead to improved survival time.

While early diagnosis often leads to a better prognosis, this is not the case for all cancers. Some people, for example, may face the emotional trauma of living with the knowledge that they have cancer for longer, but there still isn’t an effective treatment to cure them.

In this situation, longer survival time might not actually mean better survival – it just means longer survival with a cancer diagnosis.

Ultimately, screening for any disease is no magic bullet, especially a potential “blanket screen” like the method outlined in this study.

Links To The Headlines

New blood test could diagnose FIVE different cancers ‘without the need for invasive biopsies, saving millions of lives’. Mail Online, February 5 2016

Blood test to spot five deadly cancers could prevent thousands of deaths. The Daily Telegraph, February 5 2016

Links To Science

Margolin G, Petrykowska HM, Jameel N, et al. Robust Detection of DNA Hypermethylation of ZNF154 as a Pan-Cancer Locus with in Silico Modeling for Blood-Based Diagnostic Development. The Journal of Molecular Diagnostics. Published online February 5 2016