The incidence and mortality of primary brain cancers is rising, especially among the young population. Despite extensive research, there are
currently no effective therapeutic modalities or preventive strategies for the glioblastoma, the most common and most lethal primary brain tumor in
adults.
We aim to uncover biomarkers of glioblastoma invasiveness and treatment resistance that are important for patient classification and prognosis, as
well as for predicting patient response to treatment. Our approach addresses two major challenges that prevent efficient glioblastoma treatment:
invasive cancer cells and tumor heterogeneity that enable tumor reoccurrence. The multimodal approach combining molecular and radiographic
features of glioblastoma using artificial intelligence approaches will provide crucial insights into the tumor characteristics (tumor
microenvironment) and enable insights into the clinical potential of the identified biomarkers.
There is limited set of predictive and prognostic biomarkers for glioblastoma. Invasive glioblastoma cells that remain in the patient's brain tissue
after surgical removal and treatment are drivers of tumor regrowth. Several studies demonstrate that greater tumor resection is associated with
improved patient survival. Currently, there is a gap in knowledge on biomarkers that could identify rapidly invasive cancer cells. The invasive edge of
glioblastoma is understudied, mainly due to the limited availability of patient-derived biological material. Current knowledge about glioblastoma
biology comes mainly from the analysis of bulk tumor collected during biopsy or surgery. Nonetheless, as invasive cells, rather than bulk cells, drive
recurrence, potential differences between these populations would have profound therapeutic implications. In the proposed project, the selection of
biomarkers will be based on cell cultures obtained from patient-derived tissue biopsies that capture a unique genetic background and the invasive
nature of the tumors. Our translational platform GlioBanka allows us to conduct research using cancer cell cultures isolated from the invasive tumor
margin (rim) of patients. By establishing three-dimensional cell models of glioblastoma and treatment modelling, we will mimic the tumor
microenvironment and standard-of-care treatment. The biomarkers obtained in the project will be based on the detection of invasive and treatmentrefractory cancer cells, which represent cells that remain in the brain tissue after treatment.
The diagnosis of brain tumors is currently based mainly on histopathological examinations and molecular analysis. Removal of tumor tissue poses
a risk of morbidity and mortality, especially in the elderly and due to tumors in the affected areas. Despite the advantages of testing tumor
biomarkers using molecular biology, its wider clinical application remains a challenge due to the high costs and risks for patients due to tissue
sampling. Therefore, we aim to develop a new approach with radiogenomics that is cost-effective and risk-free for patients. Radiogenomics is based
on magnetic resonance imaging (MRI) of the entire tumor in its quantitative analysis using a machine learning approach and enables early
prediction of the tumor biological characteristics and the outcome of patient treatment without invasive procedures.
The challenges and objectives of proposed project require interdisciplinary team in the fields of cell and molecular biology, disease models,
oncology, radiology, pathology and computer science. The project will promote the advancement of all these disciplines and digital transition.
Taken together, the combination of molecular and radiographic features of glioblastoma and implementation of artificial intelligence approaches will
provide comprehensive knowledge of glioblastoma characteristics that can lead to the discovery and validation of new biomarkers signatures for
improved treatment and survival of patients.