AI & future brain tumour treatment: Development of a panel of predictive immunotherapy biomarkers using MRI-based radiogenomic analysis of glioblastomas

Lead Supervisor
Dr Thomas Booth
Senior Lecturer in Neuroimaging
Biomedical Engineering & Imaging Sciences and King’s College Hospital
King’s College London

Dr Marc Modat, Dr Igor Vivanco, Prof Keyoumars Ashkan, Prof Richard Houlston

Project Details

Research question 

We hypothesise that in patients with glioblastoma, a phenotype from imaging data can be correlated with the underlying genotype, allowing these radiogenomic biomarkers to guide more precise immunotherapy treatment options based on each patient’s molecular profiles.


Glioblastoma is the most common primary brain cancer. It is an aggressive tumour and significantly influences the survival of patients despite surgery/biopsy and chemoradiotherapy treatments. Standard-of-care glioblastoma treatment has not resulted in satisfactory improvements in patients’ survival outcomes. Therefore novel therapies, including immunotherapy, have been investigated in randomized settings. However, not all glioblastomas respond to these immunotherapies with significant variability in patients’ clinical responses.

Currently, most potential biomarkers for predicting immunotherapy response are genetic and molecular biomarkers, for which tissue sampling is needed. However, repeat tissue sampling has a high risk of procedure-​related morbidity in glioblastoma patients, unlike other systemic cancers. Besides, biopsy has a potential drawback of sampling bias. In contrast, radiogenomic models, which are non-invasive and can include the entire tissue volume, use machine learning to associate quantitative medical imaging data to the underlying genome.


To develop and validate a panel of radiogenomic biomarkers that will serve, ultimately, as immunotherapy biomarkers during trials. If proven these will guide individualised treatment decisions and plausibly improve clinical outcome. 


To achieve this, we will use imaging-based and genomic data during the study, facilitated by a broader supervisory team with expertise in bioinformatics, cancer genomes and pharmacology. Please contact the supervisors for more information.


  1. Ostrom Q. T. et al. CBTRUS statistical report: primary brain and central nervous system tumours diagnosed in the United States in 2008–2012. Neuro. Oncol. 17 (Suppl. 4), iv1–iv62 (2015).
  2. Weller M. et al. European Association for Neuro-Oncology (EANO) guideline on the diagnosis and treatment of adult astrocytic and oligodendroglial gliomas. Lancet Oncol. 18, e315–e329 (2017).
  3. Stupp R. et al. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med 2005; 352: 987–96
  4. Stupp R. et al. Effects of radiotherapy with concomitant and adjuvant temozolomide versus radiotherapy alone on survival in glioblastoma in a randomised phase III study: 5-year analysis of the EORTC-NCIC trial. Lancet Oncol 2009; 10: 459–66
  5. Stupp R. et al. Effect of tumour-treating fields plus maintenance temozolomide vs maintenance temozolomide alone on survival in patients with glioblastoma: a randomized clinical trial. JAMA. (2017) 318:2306–16. doi: 10.1001/jama.2017.18718
  6. Cloughesy TF. et al. Neoadjuvant anti-PD-1 immunotherapy promotes a survival benefit with intratumoural and systemic immune responses in recurrent glioblastoma. Nat Med. 2019 Mar;25(3):477-486
  7. Lim, M. et al. Current state of immunotherapy for glioblastoma. Nat Rev Clin Oncol 15, 422–442 (2018).
  8. FDA-NIH Biomarker Working Group. In: BEST (biomarkers, EndpointS, and other tools) resource. Institutes of Health (US) (2016).
  9. Müller DMJ. et al. Comparing glioblastoma surgery decisions between teams using brain maps of tumour locations, biopsies, and resections. JCO Clin Cancer Inform. 2019;3:1-12
  10. Gillies RJ. et al. Radiomics: images are more than pictures, they are data. Radiology. 2016;278(2):563–577
  11. Rushing, E.J. WHO classification of tumours of the nervous system: preview of the upcoming 5th edition. memo (2021). 
  12. Hestness J & al. Deep Learning Scaling is Predictable, Empirically. arXiv preprint arXiv:1712.00409v1 2017
  13. Sun C & al. Revisiting unreasonable effectiveness of data in deep learning era, arXiv preprint arXiv:1707.02968v2 2017
  14. Rolnick D & al. Deep Learning is Robust to Massive Label Noise arXiv preprint arXiv:1705.10694v3 2017
  15. Joulin A & al. Learning visual features from large weakly supervised data. arXiv preprint arXiv:1511.02251v1 2015
  16. Lei S & al. How training data affect the accuracy and robustness of neural networks for image classification. ICLR, New Orleans, USA, 2019
  17. Lemaitre GF & al. Imbalanced-learn: python to tackle the curse of imbalanced datasets in machine learning. arXiv:1609.06570v1 2016
  18. Yang K et al. Cancer genetic markers according to radiotherapeutic response in patients with primary glioblastoma – Radiogenomic approach for precision medicine. Radiother Oncol. 2019 Feb;131:66-74.
  19. Dorent R et al, Learning joint segmentation of tissues and brain lesions from task-specific hetero-modal domain-shifted datasets, Medical Image Analysis 2021; 67.
  20. Yong R et al, Self-Supervised Graph Transformer on Large-Scale Molecular Data, Advances in Neural Information Processing Systems 33 (NeurIPS 2020).