Youri Hoogstrate

Erasmus MC

In short, Youri Hoogstrate's research entails the following:

IDH mutant glioma objective tumor grading application

IDH-mutant gliomas are relatively low incidence but inherently lethal lethal primary brain tumors. The WHO classification distinguishes two major subgroups: IDH-mutant 1p/19q-deleted oligodendrogliomas and IDH-mutant 1p/19q non-deleted astrocytomas. The aggressiveness of these brain tumors is expressed through the 'WHO grade', a classification primarily determined by histological assessment. Unfortunately, histological tumor grading remains difficult, somewhat subjective, and exhibits inter-observer variability. Additionally, the prognostic difference between WHO grades based on survival data from the period before the WHO-2021 classification is questionable. Despite these challenges, this is the current standard, which is indicative of the patient's life expectancy and is used to make decisions regarding treatment. It guides determining whether to delay or to start a treatment, as well as to choose the type of treatment. Treatment decisions are crucial for the patient's quality of life, as chemotherapy and radiotherapy can result in cognitive decline during a significant survival expectancy (~10 years). If there were more objective ways to grade these tumors, it could be used for more univocal treatment decisions. Therefore, both clinicians and patients urgently need a more objective and accurate method for grading IDH-mutant tumors.

Using large, clinically well-annotated IDH-mutant brain tumor datasets, including a randomized phase 3 clinical trial and studies on longitudinal tumor evolution, we have identified objective molecular mechanisms of tumor progression. These mechanisms are characterized by a gradual change in the DNA methylation level of specific CpG sites. This is interesting since DNA methylation profiles are commonly used in clinical practice to determine the type of brain tumor. The results suggest that combining this type of molecular data with clinical patient information could be effectively used to objectively and accurately determine the aggressive state of these tumors.

The goal of this project is to provide an objective AI/ML-driven computational prediction of the grade and prognosis of an IDH-mutant brain tumor based on DNA methylation data. Another goal is to create a practical implementation to make the most relevant information accessible to both clinicians and the patients. We expect to make a better distinction between tumors that may or may not qualify for postponing treatment.

The project is an international collaboration between research groups and consortia in neuro-oncology, bioinformatics, and glioma tumor evolution (GLASS). In the Netherlands, researchers and clinicians from Erasmus Medical Center, TU Delft, Leiden University Medical Center, and Amsterdam UMC are involved.