Interview with WP4 leader: Erik van den Akker

“The collaboration between clinicians and researchers, coupled with advanced analytical tools, is immensely gratifying”

Dr. Erik van den Akker leads work package 4, the ‘Rheumatic digital twin’. As an assistant professor of Computational Biology, he initially conducted research in laboratory settings before transitioning to focus on computational analysis. Presently, he holds positions at both the Leiden University Medical Centre (LUMC) and Delft University of Technology (TU Delft), specializing in pattern recognition. 

Van den Akker: “Within SPIDeRR, the collaboration between clinicians and researchers, coupled with advanced analytical tools, is truly remarkable and immensely gratifying. Clinicians are dedicated to pioneering studies and tirelessly strive to reach unique patient populations. It is very motivating and a great recognition to be asked to contribute ideas on how to analyse and utilize the data obtained from such studies to improve healthcare”.

Challenges and milestones

 The primary goal of work package 4 is to systematically map the diversity across patients with rheumatic diseases. Van den Akker explains that rheumatologists generally acknowledge the vast patient-to-patient variability amongst patients in their clinic and often indicate that the current diagnostic standards might be insufficient to fully capture this diversity in rheumatic diseases. It is anticipated that this unaccounted patient heterogeneity currently obscures the effectiveness of some of the most-widely used treatments, as these treatments might be prescribed to a too broad or too unspecified patient group. Hence, the aim of work package 4 is to make more coarse and thereby more homogeneous patient definitions, which then may be leveraged to improve the matching between patients and their prescribed treatments. Consequently, he and his colleagues aim to leverage large collections of routine diagnostic data and use artificial intelligence to create concise, informative patient profiles. This aids in refining patient categorizations within clinical settings. “Physicians often possess an intuitive understanding of their patients yet lack the computer skills to distil these patterns. Our aim is to identify and document these patterns”, asserts Van den Akker.

 
A crucial milestone within work package 4 is the identification of a robust framework for patient categorization and analysis. “We intend to leverage existing diagnostic methods as a starting point, supplementing them with additional insights. Drawing from lessons learned in cancer research, we aim to apply these principles to the study of rheumatic conditions. Our objective is to define new patient subgroups and gain deeper insights into their responses to treatment.” Van den Akker furthermore says that although they enjoy a robust collaboration with LUMC, the true test lies in devising a solution adaptable to numerous hospitals. “Implementation, or translating concepts into actionable solutions, presents a significant hurdle. It necessitates trust and collaboration across disciplines, as we navigate a multidisciplinary project where effective communication is essential.”

The significance of dissemination

Within SPIDeRR, Van den Akker primarily collaborates with other data-driven work packages, such as work packages 2 (‘Data integration, harmonisation and management’), 3 (‘Modular diagnostics’), and 5 (‘Novel risk factor identification’). However, he emphasizes the importance of dissemination, stating that the solutions they develop must have tangible real-world impacts. “Currently, our focus is on organizing and enhancing e-collaboration. Over the next five years, our focus may shift towards implementation, analysing test results, and potentially updating guidelines based on insights gleaned from data.”

Multidisciplinary approach

Van den Akker plans to focus on creating patient overviews and enhancing the integration between different disease profiles, particularly in terms of treatability. For instance, he and his team have developed an overview for leukaemia, and he now aims to transfer the methodology used to construct this overview to rheumatology to better understand the heterogeneity among patients with rheumatic diseases. He explains that certain data in this overview raise new questions and prompt further research. “My philosophy is to first thoroughly understand the patients, categorize them into subtypes, and then pose new scientific questions. This approach leads to innovative insights. As a generalist, I believe the lessons we learn in cancer therapy could potentially be applied to other fields as well.”