An Rx for Healthcare: AI in Systems Medicine Sponsored Content by Dell EMC | Intel
In a workshop at Cardiff University in Wales, 15 expert speakers showcased some of the ways that machine learning applications and computational methods are driving advances in systems medicine.
Machine learning and computational methods in systems medicine are among the keys to advances in healthcare and personalized medicine. These approaches help researchers turn massive amounts of patient data into new treatments and therapies that promote better health and save lives. That’s all part of the promise of systems medicine.
But what exactly is “systems medicine”? The European Association of Systems Medicine (EASyM), an organization devoted to the topic, offers this definition: “Systems medicine is a novel approach to medicine. It is the first step on the path to personalised medicine. Systems medicine is based on computer models, in which vast amounts of clinical data are used to analyse the health of individual patients.”1
The application of machine learning and computational methods in systems medicine was a topic that was front and center this summer at a workshop at Cardiff University in Wales. The event — co-hosted by the main sponsor Dell EMC with Partners (Supercomputing Wales, Atos, Advanced Research Computing @ Cardiff, British Society of Immunology South Wales group, Cardiff Institute of Tissue Engineering and Research and the Systems Immunity Research Institute) — provided a forum that allowed 15 expert speakers from across industry and academia to share their insights into the work they are doing to drive advances in medicine and healthcare.
In this workshop, the presenters highlighted some of the ways that artificial intelligence and systems modeling can be applied to issues relating to medicine and healthcare, using best practices that capitalize on the synergies and interfaces between scientific communities and the IT industry.
A commonality to the applications and research projects highlighted at the workshop was, in a few words, massive amounts of patient data. That’s according to a report on the workshop prepared by science and technology writer Sefat Salama and Cardiff University Systems Immunity Research Institute lecturer Barbara Szomolay.
“In medicine and healthcare, it is imperative to have advanced informatics infrastructure to facilitate the process of collating, storing, sharing, integrating and analyzing the huge amounts of patient data,” the report notes. “Medicine constantly generates massive amounts of data which exceed the human capacity to process and utilise. Machine learning (ML) is a technology that can use this data to build algorithms that allow computer-based systems to generate models for meaningful understanding and potential clinical use.”2
A few examples
The real impact of AI-driven solutions in medicine and healthcare is most evident in real-world applications and research projects. With that thought in mind, let’s look at a few examples of applications and projects showcased at the data-driven systems medicine workshop. These examples were gleaned from the report by Sefat Salama and Barbara Szomolay.
- Cancer patients have many questions. Technology can help them get the answers they need, when they need them. To that end, Phil Webb, associate director at Velindre NHS University Trust, highlighted innovative applications that Velindre is doing in Wales, including the world’s first real-world virtual assistant trained in oncology. This AI-supervised chatbot — called Realtime information Technology Towards activation (RiTTa) — facilitates patient communication designed to provide relevant information at any time and in any place.
- Clinical trials are one of the gateways to the discovery new treatments, but it can be difficult to identify and recruit patients for studies. AI could help here. Irena Spasic, a professor at the Cardiff University School of Computer Science, showed how natural language processing (NLP) systems can be used to identify patients eligible for clinical trials from narrative medical records. She explained that text mining is an effective and rapid approach to search electronic health records to identify patients for trials using binary classification against certain eligibility criteria.
- Many research initiatives require access to patient medical records, but gaining that access can be tricky, given the need to protect confidential patient data. Simon Elwood-Thompson, chief technology officer of the Secure Anonymised Information Linkage (SAIL) databank, presented a new way forward for researchers. Over the past 12 years, the SAIL databank has amassed more than 32 billion medical records that are made available to researchers via a publicly funded research cloud that allows researchers store, access, share, analyse and link to data safely and anonymously for research purposes.
These examples are just a glimpse of the amazing things people are doing with AI technologies and systems modeling in medicine and healthcare. See detailed information about the individual speakers and their talks.
Dell EMC, Intel® and SurfSara data scientists worked together to demonstrate how high levels of model accuracy’s can be achieved when looking at pathologies in chest X-ray images, including pneumonia. Achieving high levels of accuracy, scalability and performance on AI models with Intel® Xeon® processors makes these capabilities more accessible to the masses.
The bottom line
A key takeaway here is that we now have data-driven technologies we need to solve some really big challenges in medicine and healthcare. That’s a point that was underscored in a 2018 report by the London-based Academy of Medical Sciences.3
The academy’s report noted, for example, that data-driven technologies could:
- Help prevent illness by identifying those at higher risk of certain diseases or by enabling earlier diagnoses
- Empower patients with long-term conditions in self-management, including through self-monitoring
- Improve outcomes by personalizing, fine-tuning or automating treatment
- Potentially increase the efficiency of healthcare services
- Drive research and innovation to develop new interventions, with potential patient benefits in both the near- and longer-term
The bottom line: All the machine learning and computational methods are in place for the application of systems medicine. Now we just need to make the most of them.
To learn more
Find out more about Supercomputing Wales’ high-powered HPC and AI resources that enable world-class science and innovation. Visit dellemc.com/hpc and join the conversation @dellemcservers.
1. European Association of Systems Medicine, “What Is Systems Medicine,” accessed August 16, 2019.
- Sefat Salama and Barbara Szomolay, “https://www.slideshare.net/mntbs1/data-driven-systems-medicine-article-167410522,” 2019.
- The Academy of Medical Sciences, “Our data-driven future in healthcare,” November 2018.