Master Thesis: Deep Learning for Brain Tumor Modeling

26.04.2024, Studentische Hilfskräfte, Praktikantenstellen, Studienarbeiten

We model brain tumor growth for improved radiotherapy planning based on 3D magnetic resonance images (MRI). Partial differential equation (PDE) based models have the potential to personalize glioma therapy. However, calibrating these models to individual patients is computationally expensive using traditional numerical solvers and classical optimization techniques. First tests with simple tumor models show promising results using differentiable deep learning based solvers for fast, gradient based optimization. Thereby a neural network inputs the specific tumor parameters like growth and diffusion rate and predicts a tumor cell concentration inside the brain [1]. To calibrate this model for a specific patient, the input parameters are optimized utilizing the differentiability of the network with respect to the input. Your role would be to utilize state-of-the-art network architectures to develop a deep learning-based solver for more realistic brain tumor growth models.

Your qualifications:
• Deeply motivated by the opportunity to make a meaningful impact in medical applications and real-world AI.
• Currently enrolled in a computer science or a related field.
• Preferably prior work experience using deep learning for image processing
• Strong programming skills in Python as well as PyTorch

What we offer:
• A chance to be part of a project with profound implications for clinical practice.
• Close supervision and state-of-the-art hardware
• An environment of collaboration with leading experts in image processing, deep learning, biomedical engineering, and medicine.

Our Group:
PD Dr. Benedikt Wiestler https://compimg.github.io/

How to apply:
Send an email to j.weidner@tum.de with your CV and transcript (Bachelor's and Master's).

References:
[1] Ezhov, I., Mot, T., Shit, S., Lipkova, J., Paetzold, J. C., Kofler, F., ... & Menze, B. (2021). Geometry-aware neural solver for fast Bayesian calibration of brain tumor models. IEEE Transactions on Medical Imaging, 41(5), 1269-1278.
n

Kontakt: j.weidner@tum.de