Digital medicine – research for the patient’s sake

Digital medicine – research for the patient’s sake
An innovative artificial heart, “ReinHeart”, pumps according to the situation – thanks to digitisation.

Online medical consultations, apps to track infections, simulations of the spread of infectious diseases – these digital tools are now familiar to all of us thanks to the corona crisis. The pandemic sparked by the novel Sars-CoV-2 virus that emerged at the start of the year not only reveals how vulnerable humankind is. It also shows how much potential digitisation has when it comes to medical treatment.

The possibilities offered by big data, digitisation and artificial intelligence will have a huge influence on medicine”, says Thomas Schmitz-Rode, director of the Institute of Applied Medical Engineering (AME) at RWTH Aachen University. “And we will succeed in deriving benefits from this, both for the patients themselves and for medicine as a whole”, promises the doctor and engineer. Schmitz-Rode himself develops cardiovascular support systems. For patients who have suffered a heart attack, the system’s algorithms ensure that the microcontroller – the digital “heart” of the implant – eases the burden on the heart according to the patient’s specific needs in any given situation. “How the implants pump must reflect the patient’s current state, which will depend for example on whether the patient is currently at rest or climbing stairs”, explains Schmitz-Rode.

Patient records contain a wealth of data

Algorithms are also an important tool when it comes to the optimal use of information in patient records with a view to making diagnosis quicker and improving treatment. Data is being generated all the time in hospitals and doctors’ surgeries: when doctors talk to patients about their medical histories, when lab results are obtained, and when examinations such as X-rays and other imaging techniques are conducted. “Much of this data is recorded in all kinds of different formats, and not all of it is digitised. This alone makes transferring the data to a central pool quite a challenge”, says Schmitz-Rode. With a view to leveraging this wealth of data, the hospital affiliated to RWTH Aachen University established the Comprehensive Diagnostic Center Aachen (CDCA) in the summer of 2018. Scientists at the CDCA are working on IT tools to “automatically evaluate, merge, and efficiently harness diverse diagnostic data for the benefit of the patient”, as the website explains.

Controlling infections in the hospital

The university in Aachen contributes its findings to SMITH - Smart Medical Information Technology for Healthcare. SMITH is a research consortium comprising nine universities and their hospitals, plus companies and research institutions. This and three other research consortia are receiving 160 million euros in funding until 2021 within the framework of the Medical Informatics Initiative set up by Germany’s Federal Ministry of Education and Research. The HIGHmed project, which is focusing among other things on controlling infections, is also being funded by the initiative. Its goal is to develop a software system that will allow transmission paths and microbial infections in the hospital to be detected at an early stage.

Avoiding misinterpretation

Many digital applications in medicine are based on self-learning algorithms. This is also true of image analysis software that is designed, for example, to detect pathological changes such as tumour tissue in microscopic images of tissue sections. To allow doctors and researchers to retain control even when using such automated interpretations, Klaus-Robert Müller, a professor of machine learning at Technische Universität Berlin, is working on an approach scientists call explainable artificial intelligence. “The software generates ‘heatmaps’ that reveal precisely which cells or imaging areas were used by the algorithm to decide whether to show, for example, a cancer or non-cancer classification”, says Müller. This allows pathologists to assess subsequently whether the automated analysis makes sense. “We must take the concerns about possible misinterpretation of data, and about data protection, seriously”, adds his Aachen colleague Schmitz-Rode. “However, this will not stop us from continuing our research with the utmost effort.”

More information:

Medical Proteome Center

Even the tiniest changes to proteins in the human body can cause diseases. Mass spectrometry is the standard method of analysing such changes. To this end, the chemical composition of elements of these proteins, known as peptides, is compared with the information in a database. The Medical Proteome Center at the Faculty of Medicine of Ruhr-Universität Bochum is developing tools to make this method of analysis even more precise. This involves the researchers using learning algorithms, among other things. It is hoped that they will ensure that mass spectrometers in the future will also be able to detect peptides that are not yet described in the database.