Image quality vs. radiation dose in dental radiology: the multifold role of artificial intelligence

Recording →

Moderator: Jenia Vassileva (IAEA), Jeffery Price (IADMFR)

Presenter: Ruben Pauwels (Aarhus University, Denmark)

Organized jointly with the International Association of DentoMaxilloFacial Radiology (IADMFR)

Background

Optimization of protection remains one of the fundamental principles of radiological protection. The definition of optimization is rather straightforward but comprises several aspects of the imaging process. During image acquisition, optimization typically comes down to balancing image quality and radiation dose. However, the application of this principle is dynamic and should be adopted in the function of several (“economic and societal”) factors.

Artificial Intelligence (AI), most notably the application of deep learning (DL) using neural networks, has recently gained attention in radiology. Whereas AI has shown a lot of potential for diagnostic tasks and segmentation, this presentation will look at potential applications of AI, which may impact the optimization process considerably:

1.   Image reconstruction and image enhancement.

Image quality in radiography and tomography is hampered by noise and different types of artefacts. Whereas noise can limit ‘how low the dose can go’, artefacts can lead to non-diagnostic images and repeated exposures. Although several noise- and artefact-reduction algorithms have been developed in the past decades, a common limitation is their computational cost. DL models could change this paradigm, in the sense that they need considerable computational power and data during training but can be applied to new datasets in (almost) real time.

2.   From dose monitoring to ‘performance’ monitoring.

Quality assurance (QA), including quality control (QC) remains an essential aspect of optimization. AI can reshape the QA/QC process in various ways. In terms of dosimetry, it could lead to a more accurate estimation of patient-specific organ doses, as well as a more targeted detection of individual or collective overexposure. In terms of image quality assessment, it can allow for a continuous monitoring of radiographic/tomographic equipment using clinical images, complementing traditional phantom-based measurements.

Objectives

1. Understand the dynamic aspect of the principle of optimization of protection.

2. Understand the different ways in which AI can enhance image quality.

3. Understand the different roles AI can play in the QA/QC process.

About the presenters

Dr Ruben Pauwels

Ruben Pauwels has a Master degree in Biomedical Sciences and in Medical Imaging. His PhD project at the Catholic University of Leuven (Belgium) was part of the Euratom project ‘SEDENTEXCT: Safety and efficacy of a new and emerging dental X-ray modality’ (2008-2011). Subsequently, he worked as a postdoctoral researcher at the Department of Imaging & Pathology and Department of Mechanical Engineering at KU Leuven, as a lecturer at Chulalongkorn University, Bangkok, Thailand and as a visiting researcher at the University of São Paulo and University of Campinas, Brazil.   

Currently, he is an Associate Professor - Marie Curie Fellow at the Aarhus Institute of Advanced Studies, Aarhus University, Denmark. He has editorial/advisory roles in several journals, including Dentomaxillofacial Radiology, British Journal of Radiology, Sensors and Journal of Endodontics. During the past two decades, major developments have taken place in dentomaxillofacial radiology. Notably, intraoral imaging went digital, which reduced radiation exposure to patients. This method has been included in undergraduate education in dentistry.

Professor Jeffery Price

Professor Jeffery Price is Clinical Professor and Director of the Oral and Maxillofacial Radiology Department at the University of Maryland School of Dentistry in Baltimore, Maryland, USA and adjunct associate professor of oral and maxillofacial radiology at the University of North Carolina (UNC) School of Dentistry. He is currently a co-principle investigator for a National Science Foundation-funded research project entitled A Machine Learning Framework for Comprehensive Dental Caries Detection. Professor Price is a USA Director for the International Association of DentoMaxillofacial Radiology (IADMFR) and Secretary of its Board. He also serves on the editorial board for the Dentomaxillofacial Radiology journal and on the Academy of General Dentistry Self-Instruction Committee.