Rosebrugh Bldg, Toronto, ON M5S 3G9
Room: RS 211
Cerebral vasculitis, also known as primary angiitis of the central nervous system (PACNS), is a disease characterized by vessel inflammation of the brain and/or spinal cord. Untreated, the disease can lead to severe neurological deficits and even death. Neuroimaging using magnetic resonance (MRI), which is a non-invasive procedure, is a key diagnostic tool that is sensitive to detecting the changes associated with PACNS. However, these changes may be similar in presentation with other conditions such as multiple sclerosis (MS), making the distinction between these diseases difficult, which is important as treatments are very different. To date, there is no clear and agreed upon way for assessing neuroimaging in PACNS.
Our first aim is to develop machine learning algorithms which would automatically segment PACNS lesions based on neuroimaging features. This will then be used in conjunction with clinical presentation to develop a comprehensive PACNS scoring system. We then aim to test our algorithm’s performance against expert clinician diagnosis.
We expect to transform the current MRI review process by developing an algorithm to automatically detect MRI features indicative of PACNS, and use these features to augment clinical features in a combined probability score, which will enable physicians to rapidly diagnose and initiate life-saving therapy.
The establishment of a structured method of analyzing MRI is of clinical importance as treatment and diagnosis often rely heavily on this modality. This study will help physicians around the world diagnose CNS vasculitis, start therapy and prevent brain damage.