Brain Tumour Segmentation

Type: MA thesis

Status: open

Supervisors: Luis Carlos Rivera Monroy, M. Sc., Laura Pfaff

[Master Thesis]

Glioma, especially glioblastoma and diffuse astrocytic glioma with the molecular characteristics of glioblastoma (a grade 4 astrocytoma), is the most frequent and aggressive type of brain tumour in adults. It presents extreme appearance, shape, and composition variability, and patients with glioblastoma have a poor outlook. The standard treatment consists of surgery followed by radiotherapy and chemotherapy. The methylation status of the promoter of the DNA repair enzyme MGMT has been linked to a better prognosis in newly diagnosed glioblastoma and a better response to chemotherapy. Thus, determining these patients’ MGMT promoter methylation status can impact treatment decisions. The International Brain Tumor Segmentation (BraTS) Challenges, which started in 2012, evaluate the latest machine learning methods for analyzing brain tumour images in MRI scans.

 

[Research Project]

Implementation of data augmentation pipelines for brain tumour structures. With a focus on “difficult” to-segment regions.  (Project already running)

Implementation of noise reduction schemes for MRI scans (Diversity in acquisition protocols, device brand and modality)

 

[Requirements]

Please send your transcript of records, CV, and a small motivation letter on why you would be interested in the topic.