Artificial Intelligence in Hepatobiliary Imaging

medical imaging Radiology Artificial Intelligence

Dr. T.C. Kwee
Dr. R.J. de Haas
Dr. D. Yakar

Nature of the research:
Retrospective evaluation

Fields of study:
surgery gastroenterology radiology

Background / introduction
Artificial intelligence (AI) is an exciting research field which investigates the creation of computer algorithms as a potential substitute for specific human tasks (1). Sometimes referred to as machine learning, is a very dynamic field of research and has gained a lot of attention in the media recently (Figure 1). Tech giants such as Google have investigated billions in AI. For the medical imaging field it is to be expected to have major consequences in the future.
At the UMC Groningen, more than 200,000 medical imaging procedures are performed annually. These medical imaging procedures include among others CT, MRI, and nuclear medicine exams. Virtually any part of the body and numerous diseases can be imaged with a medical imaging technique. It is the radiologist’s task to interpret these exams and provide an adequate diagnosis, to guide further follow up, or asses treatment planning possibilities. AI in radiology can potentially increase diagnostic accuracy, realize faster turnaround, and perhaps even improve quality of work experience for the radiologist (2,3). The UMCG is considered an expert centre for hepatobiliary diseases and each year many patients undergo imaging exams for liver, pancreatic or bile duct diseases. In the current study we would like to investigate how AI can aid in the evaluation of imaging exams of patients with hepatobiliary diseases.
Research question / problem definition
What is the value of AI in the evaluation of imaging studies in patients with hepatobiliary diseases?
• Can AI determine the level of fibrosis in liver cirrhosis?
• Can AI classify different types of liver tumours?
• Can AI determine patient prognosis with a certain type of hepatobiliary disease?
This project is suited for a medical student in his/her senior Bachelor (2nd or 3rd year) or Master phase for a period of 3 (minimum) to 6 months. Students with an interest in radiology and AI are very welcome, but students with interests in other specialties (surgery, internal medicine, etc.) are also invited to apply, because the subject interconnects with all fields of medicine and AI. This project will be intensively supervised by a team of radiologists and computer scientists with ample scientific experience. The final product will be a co-authored scientific article. Time schedule: part 1: familiarisation with the topic; part 2: data collection; part 3: data analysis and manuscript writing. During this research internship, there is also the possibility to accompany the radiologists with daily clinical activities and attend clinical meetings with other medical specialists.
1. Tang A, et al; Canadian association of radiologists (CAR) artificial intelligence working group. Canadian association of radiologists white paper on artificial intelligence in radiology. Can Assoc Radiol J 2018;69:120-135.
2. Aerts HJWL. Data science in radiology: a path forward. Clin Cancer Res 2018;24:532-534.
3. Thrall JH, et al. Artificial intelligence and machine learning in radiology: opportunities, challenges, pitfalls, and criteria for success. J Am Coll Radiol 2018;15:504-508.
back to toptop

Laatst gewijzigd: 23 februari 2012