Transforming Pancreatic Cancer Data into Clinically Useful AI Representations

PhD defence by Reza Karimzadeh Mostafaabadi

Assessment Committee

Associate Professor Melanie Ganz-Benjaminsen, Computer Science, University of Copenhagen (Chairperson)
Lektor Esben Bolvig Mark, Aalborg Universitetshospital
Professor Žiga Špiclin, University of Ljubljana

Supervisors

Associate Professor Bulat Ibragimov

Department

Department of Computer Science

Place

Building: SCI-DIKU-VG5, Room: 01-0-074, Vermundsgade 5, 2100, Copenhagen

Email address to gain access to the thesis: reza.karimzadeh@di.ku.dk.
You will either receive a copy of the thesis or be informed where you can read a physical copy.
Recipients of copies of the thesis are not allowed to share or distribute it due to copyright compliance.

Short description of the thesis

How can artificial intelligence turn complex cancer data into useful information for doctors? This thesis explores that question with a focus on pancreatic cancer. Medical scans and patient records contain valuable clues, but the information is often incomplete, scattered, and difficult to interpret. The research develops methods to learn from scans with limited expert guidance, estimate missing clinical information, find relevant details in written reports, and explain AI predictions. It also investigates how information from multiple tumors can inform predictions for individual patients. These contributions lay the foundation for more transparent and useful AI tools to support future cancer care.