Trial lecture
“Anatomical atlases for medical image segmentation - state of the art and future perspectives”
Time and place: February 7, 2025 11:15 AM. Seminarrom 2 B2.U002, Rikshospitalet, Oslo Universitetssykehus.
Main research findings
Liver cancer is the sixth most common cancer globally and the third leading cause of cancer-related deaths. Laparoscopic surgery and liver parenchyma-sparing procedures are effective treatments for liver metastases. Due to the limited view and tactile feedback in laparoscopic surgery, enhanced surgical planning is crucial for success. Efficiently extracting clinically relevant information from pre-operative or intraoperative images and presenting it to clinicians as patient-specific models can greatly improve surgical planning and patient outcomes.
But extracting this patient-relevant information is a labor-intensive process. With the advent of advanced techniques, particularly in artificial intelligence and deep learning, incorporating these advancements into clinical routines would be highly beneficial. This integration is not straightforward due to challenges such as data availability, clinical validation, and the reliability of these techniques.
This thesis explores the transformative potential of deep learning in creating patient-specific 3D models from multi-modal medical images. The main challenges in developing deep learning solutions for medical images include the availability of data with standard annotations, computational resources, and clinical validation of the developed methods. Clinicians often perceive deep learning-based methods as "black boxes," so it is essential to help them understand why and how the model makes its predictions. This work introduces ways to tackle some of these challenges, particularly handling limited data availability. Additionally, through clinical validation of the developed technique, the relevance of integrating automated methods into clinical routines was tested.
Adjudication committee
- Professor Bjoern Menze, Universität Zürich, Switzerland
- Professor Ôrjan Smedby, KTH Royal Institute of Technology, Sweden
- Professor Kyrre Glette, University of Oslo, Norway
Supervisors
- Senior researcher, PhD, Rahul Kumar, Oslo University Hospital, Norway
- Professor Ole Jacob Elle, University of Oslo, Norway
- Professor Emeritus Fritz Albregtsen, University of Oslo, Norway
- Professor Frank Lindseth, Norwegian University of Science and Technology (NTNU), Norway
Chair of defence
- Professor Carsten Griwodz, Department of Informatics, University of Oslo
Contact information at Department: Mozhdeh Sheibani Harat