A digital twin helps to understand, treat, and predict 

A digital twin is an interactive model of a patient, disease, a biological process, or environment that allows collecting, simulating, and analyzing various health data in real time. Digital twins help to better understand disease mechanisms and test how a patient might respond to treatment or how the disease might progress. For example, data about the patient's genome, behavior, and blood plasma can be input into the model to create a disease model, test different treatment plans, or predict disease progression without actual intervention.

TalTech researchers are involved in developing a digital twin to support schizophrenia treatment. Currently, schizophrenia treatment works for only half of the patients, and many experience severe side effects. We want to understand who responds to the treatment and why, and how we can predict it.

The Estonian company Protobios and the neurobiology research groups of Tõnis Timmuski and Indrek Koppel from the Technical University are involved in developing the schizophrenia digital twin. Neuroscientists contribute data from animal models, which helps to understand the disease mechanisms at the cellular level. 

Shared database, distributed infrastructure 

A platform connecting European research institutions is being created, where data and technologies are consolidated into a so-called unified catalog. We produce a lot of data but do not know how to extract enough knowledge from it. The use of data is also hindered by different data standards and data protection restrictions. It is like having a huge library, but it is impossible to find the right book because there is no catalog and access is difficult.

During the project, this missing catalog will be created – data collections and resources will be described and standardized, data exchange standards and principles of data access will be developed, and data protection will be strengthened.

Federated learning, a distributed training method, is used for training the models and is one of the core strengths of our project.

Distributed learning is a future technology that attempts to solve the problem of data sharing. Currently, it is difficult to share data between institutions. This is addressed with a distributed infrastructure where each institution has its own data node. Thus, data is not transmitted to others; instead, each trains the model with its own data.

First, of course, the model needs to be developed. So initially, information is gathered about who has what data, and a corresponding model is created. Then it is divided who trains which aspect, because the data is different. With each additional user training the model, it becomes more accurate. The model is later made available to others. 

Sensitive data requires careful handling 

A digital twin is an interactive model that looks like a set of code, but data often needs to be visualized for use. Technologies, especially virtual reality (VR) and augmented reality (AR), are used to visualize the twin, making complex data, predictions, and dynamic processes visible and understandable. The project includes partners who help decide where 3D visualization would be most beneficial. One possible use case is magnetic resonance imaging (MRI), where VR-based simulations could be used for safety training.

Health data is very sensitive, and there are many aspects that need to be carefully considered when using and sharing it. Therefore, the consortium includes data scientists, lawyers, ethics experts, and the European Health Management Association (EHMA). Ethics experts have a significant role in this field, and recently four external ethics experts were added to the project partners. One of the project outcomes will be a governance and ethics framework/guide, which will be made publicly available afterwards.

The project goals align very well with the objectives and themes of TalTech’s Health and Food Technologies Centre of Excellence. Together, we contribute to increasing the number of healthy life years by enhancing our capacity to use next-generation solutions and technologies in the prevention and treatment of chronic diseases and cancer, as well as for personalized and inclusive medicine. 

The DTRIP4H project is funded under the European Horizon program and received 11.9 million euros in support from the European Union (grant number 101188432). The views expressed within the project are those of the authors and may not reflect the official positions of the European Union or the funding agency. The European Union and the funder are not responsible for these views. 

The article was published in the Tallinn University of Technology magazine Mente et Manu.