Optimizing the use of digital data and medical imaging in the clinical workflow

How can artificial intelligence and medical imaging techniques improve our clinical workflow and improve the treatment process for patients? This is a fundamental research question for associate professor Ruben Pauwels, born 1985.

Ruben Pauwels is interested in the use of digital data in general, and medical imaging in particular. Specifically, his research comprises diagnostic medical physics topics, e.g., the optimization of X-ray based imaging technologies, as well as artificial intelligence (AI) centered around deep learning (DL), which is a subtype of machine learning.

According to Ruben Pauwels, the importance of this research field is multifold: “From a patient-centric perspective, my aim is to ensure the optimal use of imaging and other data throughout the patient’s treatment process – from balancing the imaging quality and radiation dose to computer-aided treatment planning. From the clinician’s perspective, my research aims to augment their work by automating or facilitating certain laborious task, ranging from segmentation to radiological interpretation.”

Ruben Pauwels's journey into medical imaging research began, when he at the age of 20 chose the topic for his Master’s thesis in Biomedical Sciences. He appreciated the deterministic nature of the research, the physics and mathematics behind imaging techniques like CT and MRI, and the variety in the day-to-day research activities.

Currently, Ruben Pauwels is using deep learning in several research projects for various types of image processing, e.g., enhancement and segmentation. Through collaboration with researchers from different dental subspecialities he works on diagnostics applications of deep learning, e.g., lesion detection and bone quality assessment.

“My long-term goal is to establish an enhanced clinical workflow, in which artificial intelligence does not take over for the clinical expert, but assists them in their work,” Ruben Pauwels states and elaborates: “There is no specific point at which we can consider an AI-based model to be ‘finished’; there is always potential to improve its performance as more data and new DL methods become available. Thus, my secondary goal is to set up an infrastructure in which the ‘continuing education’ of AI models can be streamlined.”

Improving the use of digital data and medical imaging techniques for the good of clinicians and patients is the overall driving force for Ruben Pauwels, but a pivotal moment in his early career made clear to him that this research path was worth following.

“Just a few weeks before my PhD defense in 2012, I had the honor of receiving the Research Award as well as the Research Fellowship from the European Academy of Dentomaxillofacial Radiology. Despite the fact that I do not perform research for the sake of awards, rewards or other types of recognition, the acknowledgment of my research by my peers within the EADMFR definitely strengthened my belief that I should continue pursuing new research challenges within this field."