Generously funded by Donors of Brain Tumour Foundation of Canada

Garnette Sutherland, University of Calgary, Calgary, Alta.

Project title: “Artificial intelligence for early diagnosis and prognostication for meningioma”

Description of project:

Meningioma is the most common primary brain tumour in adults. Although a large percentage are classified as benign, the more aggressive types of meningioma impose a significant morbidity and mortality burden on patients. Molecular analyses of these tumours aid accurate prognostication; however, performing such analyses is costly and difficult for resource-challenged centres, even in developed nations. A tool to identify the genetics and aggressiveness of meningiomas before surgery can help with surgical planning and postoperative management, leading to better patient-specific treatment strategies. Artificial intelligence (AI) offers great promise for this purpose. Although magnetic resonance imaging (MRI) produces a wealth of pixel data invisible to the naked eye, AI algorithms can rapidly process and interpret thousands of pixels.

What receiving this award means:

The support by Brain Tumour Foundation of Canada will be a catalyst towards an online prediction model, built on population-based molecular characterization of meningioma, that includes imaging and patient demography. Built on a unique dynamic digital database (using a knowledge graph) and custom machine learning algorithms, this tool will provide valuable insights for clinicians, scientists, and patients.

Midpoint Report – February 2026

Thanks to support from BTFC, our project, titled Artificial Intelligence for Early Diagnosis and Prognostication for Meningioma, continues to advance, with early success in clinical imaging data acquisition and analysis. Clinical correlates and pre-surgery MR imaging data from over 500 patients (under institutional ethics approval) have served as ground truth for training a unique AI model. It is anticipated that these pre-surgery imaging data output and corresponding patient-specific molecular diagnoses of meningioma will help predict patient outcomes.

For the latter, efforts are underway to establish a correlative link to molecular genetics and to apply it to the evolving AI model. Additional use of imaging and molecular data in the AI model will also be valuable for evaluating meaningful clinical output. More effort and resources will be allocated to the preparation and publication of new manuscripts and to presentations at brain tumour conferences and patient-centric educational platforms.

Of note, the BTF Canada support has enabled doctoral thesis work that interrogates molecular-genetics and radiomics characteristics of meningioma using novel AI/machine-learning tools towards prediction of disease progression/recurrence/conversion to higher grade, such that patient-specific precision management strategies could be tailored while devising novel models to predict disease course during their lifetime.

The project will also provide valuable input to the ongoing development of a comprehensive brain tumour-specific dynamic digital database for Southern Alberta that Dr. Garnette Sutherland is spearheading.