Using data science for precision medicine

Dr Lucas Fleuren introduced the night’s presentations with his regular update on the latest news from the medical data science world. It was a pat on the back for his own institution: a news story about Amsterdam UMC releasing a database of medical data which was published on the NOS website and on its television news bulletin. “It’s cool that the mainstream news outlets are picking up on this – that data science in the medical field is really taking off.”

Bringing algorithms to the bedside

Amsterdam Medical Data Science (AMDS) Associate Professor of Artificial Intelligence at Vrije Universiteit Amsterdam (VU Amsterdam) Mark Hoogendoorn was next to speak. Welcoming the full room, he emphasised how important the meetings were to help to bring medical data science applications to the bedside in clinical practice. “We see a lot of data science applications that hardly ever reach the patient. We feel it’s very important to do that to move things forward.”

Hoogendoorn is also enthusiastic about how data science is developing personalised treatment recommendations. “We see a lot of predictive modelling. But with personalisation, we’re moving one step further and starting to decide on what interventions would be successful. And that might be a series of interventions to get the patient as well as you can. And you will see more things coming that try to do this. That’s of great interest because it brings it closer to the patient.”

Model bedside manners

The frustration of the disconnect between medical data science and the bedside is also felt by intensivist Dr Patrick Thoral. “Getting a model to the bedside is a problem. We’re facing a translational gap.” As well as developing the model, you need to validate it, create a software prototype, test it on the end-user, engage stakeholders and perhaps even secure CE-certification.

In a collaboration with MedTech startup Pacmed, Amsterdam UMC developed a real-world decision support tool for discharge from the Intensive Care Unit. “But developing a model is just the beginning. Getting it to the bedside can take a lot of time. In our case, it took a year. We hope in the next few weeks we will get approval from the board so we can use this software from next year.”

Precision is everything
Radiotherapy is particularly tricky business. The radiation dose delivered to a particular tumour should be big enough to destroy the cancer cells while at the same time not cause permanent damage to normal tissue. In other words: the more precision you have, the better…

As a senior researcher at the Centrum Wiskunde & Informatica (CWI) (Center for Mathematics and Computer Science), Dr Peter Bosman is specialised in applying medical informatics to make radiation treatments as effective – and non-damaging – as possible.

Currently, big dose reconstruction data is used to model the relationship between the dose and onset of adverse effects and thereby help formulate the best treatment possible. In his talk ‘3D Radiation Dose Reconstruction’, Bosman described the incredibly complex approach of reconstructing a 3D dose distribution from 2D records of a past treatment – for example of a childhood cancer survivor.

How profiling health behaviour is like dominoes

Professor Aart van Halteren leads a team of concept designers, computer scientists and health psychologist scientists at Philips Research. Using Leiden University’s J. M. J. van Leeuwen’s domino magnification study, Van Halteren showed how a small amount of energy can have a big impact on a patient’s condition. “Through AI you can empower patients to take care of their own health. Up to 60% of the risk associated with the development of a chronic disease is to do with the behaviour of an individual.”

By profiling a patient, Van Halteren and his team are trying to find the why and the what of their health behaviour. “In healthcare we are, to some extent, already doing that. You need to use the energy in the system in the right way. That’s what we’re still learning in the behaviour of patients.” Van Halteren’s team developed a SeMaS, or a self-management screening questionnaire. They also developed a predictive model that provides a strategy to overcome barriers to medical adherence, or the degree to which a patient correctly follows medical advice. “The fact that we can build models that are predictive of people’s behaviour is a great opportunity. We can use technology to create solutions that are suitable for individuals and helps them achieve their health goals.”

About Amsterdam Medical Data Science

The Medical Data plus Pizza meeting aims to bridge the gap between health professionals and data scientists by bringing both together in an informal setting for presentations and pizza. At Amsterdam UMC data scientists, medical professionals and researchers discuss how AI (Artificial Intelligence) and medical data can be used to benefit patients and improve medical practices. Since its launch in August 2018 it has steadily grown in popularity with more and more people attending to hear the latest news and innovations in medical data and spark new collaborations.

For more reports from previous Medical Data plus Pizza Meet-ups, click here.

The Amsterdam Medical Data Science Group meetings are supported by The Right Data Right Now consortium, which includes Amsterdam UMC, OLVG, Vrije Universiteit, Pacmed and the Amsterdam Economic Board.

Copy: Edenfrost