Analysing 1,600 human hearts with Big Data
Medical researchers are increasingly turning to analytic tools to undertake ever bigger projects, while several big data analyses have already delivered significant results for healthcare.
U.S. President Barack Obama recently described it as “one of the greatest opportunities for new medical breakthroughs that we have ever seen.” Google’s Larry Page believes it will even save “100,000 lives.” Right now in medical research, big data is a big deal.
In one of the most striking uses of data analytics yet, Dr. Declan O’Regan, head of the Robert Steiner MRI unit at Hammersmith Hospital in London, has stored 1,600 beating human hearts on a computer and is using advanced magnetic resonance imaging to observe the organs and structures of the human body.
MATLAB and RStudio
When it comes to analysing images, the research team uses the old scientific favourite MATLAB for its 3D visualisation. Developed in the late 1970s, MATLAB has caught up on the big data trend and added several data analytics features such as machine learning and a MapReduce functionality.
The team’s statistical modelling is carried out using RStudio server, a browser-based open-source IDE built solely for the programming language R. The server edition of the IDE renders the researcher’s data accessible from any web browser that can communicate with the server.
The team, which will publish its findings later this year, are already discovering how rising blood pressure causes thickening of the heart muscle and this will help to understand how hypertension shortens lifespan, O’Regan told JAXenter:
We take moving 3D images of the heart from an specialized MR scanning sequence and then use a “cardiac atlas” to identify every anatomical part. This builds a digital model of how the beating heart varies from one individual to another and we can start to explore how genetic and environmental effects play their part.
But the research doesn’t stop at data analysis. Medical researchers like O’Regan are also looking at machine-learning approaches in order to predict heart disease by observing how the heart functions in 3D.
Meanwhile, the European Bioinformatics Institute in Cambridge is conducting a similar analysis of genetic codes in plants to gain a better understanding of genetics. Big data research has also been used to alert patients at risk of potentially fatal cholesterol disorders.
Programmers with medical knowledge wanted
With all the potential that data analytics holds for medical research, there’s no denying that programming skills are becoming more and more essential for medical researchers. However O’Regan says this interdisciplinary overlap makes it difficult to find individuals for technically advanced medical research.
SEE ALSO: What is Big Data?
“Knowing what programming can achieve is essential, but it’s hard to find programmers with the right combination of skills and knowledge who can work with both imaging and genetic data,” O’Regan told JAXenter.
Traditionally, programmers interested in scientific research require some experience in MATLAB, R, C or C++.
One of the tech world’s biggest advocates of opening up medical data is Google’s Larry Page, who, although admittedly ‘disappointed’ in the U.S. government’s use of its citizen’s data, claims that “tremendous good can come from people sharing the right information with the right people in the right way.” Page himself benefited from big data analytics in medical research in the treatment of his vocal cord nerve disorder.
However there is one major hurdle that the big data world must overcome if medical analytics is to succeed in improving healthcare: the fear of sharing private data. Unlike software end users who, often unwittingly, exchange personal data for free services, the genetic information required for projects like Declan O’Regan’s needs to come from individuals willing to hand over personal data.
Prying governments, global privacy scandals, corporate data collection – the relationship between individual (whether consumer or patient) and the countless parties that want access to his or her data, is strained.
For this purpose, leading British bioethicists have released a report on the importance of governance in big data analysis for medical research. The only way to (re)gain the trust of hospital patients is to achieve more transparency in data analysis and introduce criminal penalties for the abuse of private medical data, say the team of bioethicists.
“We are generating more data about people’s health and biology, from more sources, than ever before, including from GP records (e.g. care.data), hospital notes, laboratory tests, clinical trials, monitoring devices and health apps,” the authors of the medical data report write. However the council for bioethics claims that the risk of inhibiting potentially valuable research through loss of public trust in the medical profession has not yet been fully recognised.