Spot the clot: stroke diagnosis with CT angiography

Champion Name

Chris D'Esterre; Connor McDougall

Challenge ID


Project Track


Skills Needed

Matlab, Medical Imaging,

Challenge Description

Share This

Create an automated tool to "spot the clot" on computed tomography images --

Diagnosing stroke is a common challenge for any clinician, especially for those at low patient through-put centres in rural communities. It is critical that clinicians are able to identify when a stroke is occurring in order to make a timely treatment decision and potentially triage a patient to a larger urban hospital for better management. Computed tomopgraphy (CT) imaging is the standard of care for patients with stroke symptoms. The CT angiography (CTA) for which contrast is injected into the blood, improves the detection of stroke; however, many blood clots, especially those in the distal arterial tree, are missed. Therefore, our challenge is to "spot the clot" on CTA.

To spot the clot, an assortment if image analysis techniques could be used - statistical methods (machine learning, deep learning), structural methods, filter-based and model based approaches, etc. We are excited to see how the team approaches this problem!