Deep-Learning-based Image Registration and Tumor Follow-Up Analysis

Monday 10 October 2022, 12:30
PhD defence
Speaker or Ph. D. student
A.D. Hering MSc.
prof. dr. B. van Ginneken, prof. dr. ir. H.K. Hahn
dr. N. Lessmann, dr. S. Heldmann
Faculty of Medical Sciences

Computed tomography (CT) scans are taken at every stage of cancer diagnosis and treatment and must be read by a radiologist. Before treatment begins, the radiologist locates and measures the metastases. Now the patient has been treated for some time and the doctor wants to know if the metastases have shrunk as hoped. To determine this, another CT scan is taken and compared to the previous scan. However, comparing these images is quite complex. This research has shown that artificial intelligence (AI)-based image registration algorithms are capable of establishing correspondences between two images, and therefore can be used to aid in this comparison: Clicking on a specific point in the old images automatically finds the corresponding position in the new image. In addition, research has shown that clinical workflow can be further improved by using AI-based algorithms to automatically measure metastases.

Alessa Hering (1990) obtained her Master’s degree in Computational Life Science, summa cum laude, at the University of Lübeck, Germany in 2017. Afterwards, she started as a research scientist at Fraunhofer MEVIS. In 2018, she additionally joined the Diagnostic Image Analysis Group at Radboudumc as an external PhD candidate.  Currently, she works as postdoctoral researcher at Radboudumc and Fraunhofer MEVIS.

This website is still under construction. More information: 'a new website'.