Changes in blood cells linked to an aggressive form of leukaemia may begin decades before diagnosis, an analysis of stored samples has found.
Researchers at University College London and the University of Cambridge examined blood samples from 47 women who later developed acute myeloid leukaemia, or AML. They compared the results with samples from 46 women of similar ages who remained free of blood cancer.
The research, published on 11 November, combined detailed genetic sequencing with mathematical modelling. It offers a picture of how groups of altered blood cells can grow over time, but it is not a test that can determine with certainty whether an individual will develop leukaemia.
Looking back through blood samples
AML is a rare cancer affecting the blood and bone marrow. Although the disease itself usually develops quickly and needs prompt treatment, the genetic changes that contribute to it can accumulate over a much longer period.
The samples came from a large UK trial originally set up to investigate ovarian cancer screening. Some participants had given blood every year for up to 11 years, creating a collection that could also be used to explore other questions about health and disease.
For the women who went on to develop AML, researchers examined between two and ten samples per person. Some had been taken as long as 15 years before diagnosis. The women were aged between 53 and 83 when diagnosed, with a median age of 72. No sample was taken at the point of diagnosis for this analysis.
By comparing samples taken at different times, the team could see how the proportion of cells carrying particular genetic changes rose or fell. They then used a model to estimate when those changes first appeared, including periods long before the earliest stored sample.
In about two-thirds of the women who developed AML, the first cancer-driving mutation was estimated to have arisen by the age of 20. Across the cases, the estimated first changes appeared, on average, more than 50 years before diagnosis. Those early dates are reconstructions, not direct observations of blood taken from the women when they were young.
Why the pattern of growth matters
As people age, some blood-forming cells acquire changes in their DNA. If a cell and its descendants gain an advantage, they can make up a growing share of the blood. A group of related cells with the same change is known as a clone.
Having such a clone is not the same as having leukaemia. Similar changes were found in many women in the comparison group who did not develop blood cancer. In the samples studied, altered blood-cell populations were detected in 33 of the 46 controls and 42 of the 47 women who later developed AML.
The researchers found that some early clones expanded slowly for years. In some cases, cells then acquired further changes associated with cancer and grew more rapidly. Other groups of altered cells competed with one another, producing different patterns rather than a single route towards disease.
In at least a third of the AML cases, a clone came to dominate the blood before diagnosis, a pattern known as a somatic sweep. No such sweep was seen in the matched control group. Analysis of a separate UK Biobank dataset also found an association between these sweeps and a higher risk of AML.
That association does not mean a sweep guarantees a future diagnosis. The comparison group in the stored-sample study was small, and further work would be needed to understand how reliably the pattern could help clinicians assess risk in a wider population.
What the findings do not show
The timing and growth of altered cells varied. In around one in five of the AML cases, the pre-leukaemic clone was detectable only in the final one or two years before diagnosis, if at all. Possible explanations include changes outside the testing panel or cell populations that remained below the level the analysis could detect.
All of the stored samples studied came from women who had taken part in a trial involving postmenopausal women. The findings therefore need testing in other groups before researchers can know how broadly the same patterns apply.
The research raises the possibility that tracking high-risk clones could one day help identify a window for intervention. A model also suggests that slowing the growth of such cells over time might reduce the opportunity for further harmful changes. Neither idea demonstrates that a preventive treatment exists or that it would be safe and effective.
These findings do not mean a person’s leukaemia could already have been predicted or prevented, or that anyone was responsible for developing it. People with symptoms such as unexplained bruising, unusual tiredness or repeated infections should seek an urgent GP appointment or contact NHS 111. Such symptoms are common and can have many causes, but they need assessment.
