latest
The MILTON tool, trained on UK Biobank data, has been found to predict an individual’s likelihood of being diagnosed with a disease based on their biomarkers
"Improving our ability to detect illnesses earlier and at more treatable stages is critical for early interventions in clinical care." Slavé Petrovski, head of the Centre for Genomics Research, AstraZeneca
A machine-learning tool developed by pharmacy company AstraZeneca could predict more than one 1,000 diseases before diagnosis, a study has found.
The study, published in Nature Genetics, reported on an AI tool called MILTON, short for MachIne Learning with phenoType associatiONs. It found that MILTON, which was trained on UK Biobank health record data, is able to predict whether individuals are more likely to have, and be diagnosed with, certain diseases based on 67 biomarkers that are routinely collected during clinical practice. These include blood biochemistry, blood counts, respiratory function scores, blood pressure variables, and other measures like age, sex, body size and fasting time.
In the study, MILTON was used to analyse 3,200 diseases and achieved a high prediction score for 1,091 of them. The results were validated in Finland’s FinnGen biobank. MILTON can be applied to any biobank, irrespective of genomic ancestry, the researchers say, and will be further developed by adding additional data such as proteomics collections. (Proteomics is the study of the proteins present in a given cell or organism.)
The AstraZeneca researchers, who are based at the firm’s genomics research centres in the UK and the US, say that MILTON has the potential to accelerate the discovery of new drug targets and biomarkers. This could pave the way for the development of more effective and targeted treatments, as well as allowing early disease detection.
MILTON was created in order to identify individuals who had been wrongly classified in population-based cohort studies where a control population is used. By correctly reclassifying some individuals in the control group as putative cases, MILTON has enhanced the statistical power of the genetic association analyses.
The researchers note that “identifying individuals at high risk of developing disease is a priority for preventative medicine,” but that traditional risk assessment tools rely on clinical parameters such as as age, sex and family history, and a “reduced set of basic biomarkers tailored to the disease under study.” They add that these tools “may not capture the full spectrum of biological processes that underlie complex diseases.” In contrast, biobanks, which integrate electronic health records with, for example, blood biomarkers and imaging data, provide an opportunity to examine the association between particular biomarkers and certain diseases.
According to lead author Slavé Petrovski, head of the Centre for Genomics Research at AstraZeneca, MILTON is a significant advance on the predictive tools currently used, and can outperform gene-based risk scoring systems.
“Our research demonstrates MILTON’s capabilities and how it is able to identify disease risk cases in large biobank datasets, which, in the future, could enable us to detect illnesses earlier and at more treatable stages,” Petrovski said. “Improving our ability to detect illnesses earlier and at more treatable stages is critical for early interventions in clinical care.”
Other experts have been more cautious. Professor Tim Frayling, professor of human genetics at the University of Geneva, praised the thoroughness of the study, but said care needs to be taken when talking about predicting disease when “we really mean ‘we can give you a slightly better idea of your chances of developing a disease, but there are still many unknown factors.’ Thus this approach will likely have more impact on improving our knowledge of how diseases develop rather than who exactly will develop them.”
Professor Dusko Ilic, a stem cell specialist at King’s College London (KCL), said MILTON represents a “significant step forward in the field of predictive medicine,” but said he also had concerns about its ethical use: “The powerful predictive abilities of this tool could, if unregulated, be misused by health insurance companies or employers to assess individuals without their knowledge or consent. This could lead to discrimination and a breach of privacy, [so] strict guidelines and oversight will be critical in ensuring that the benefits of MILTON are realised in an ethical and responsible manner.”
FCC Insight
This exciting study demonstrates the huge potential of artificial intelligence to transform health care. By training AI tools on large datasets such as the UK Biobank or FinnGen, it is possible to find associations between biomarkers and particular diseases, thus providing the opportunity for clinicians to detect the beginnings of disease long before patients display symptoms. If we are to meet the demands of an ageing society, in which more and more people are living with chronic disease, it is essential that we make good use of AI tools such as these. Experts are right to sound a note of caution, because such advances bring risks as well as benefits, but if we use the technology wisely, we can find ways to intervene and treat diseases early.