Abstract #4293

Detection of Spinal Cord Abnormality on Diffusion Tensor Imaging (DTI) in Patients with Unilateral Deficit using Pattern Classification

Arturo Cardenas-Blanco¹, Santanu Chakraborty², Fahad Alkherayf³, Eve Tsai³, Mark Schweitzer², Thanh Nguyen²

¹Diagnostic Imaging Department, the Ottawa Hospital, Ottawa, Ontario, Canada; ²Radiology, the Ottawa Hospital, Ottawa, Ontario, Canada; ³Neurosurgery, the Ottawa Hospital, Ottawa, Ontario

Conventional magnetic resonance imaging in the spinal Cord (SC) is often insufficient to diagnose, asses the stage and progression of disease. Abnormalities seen on conventional MRI are often unreleated to clinical findings. During the last years, Diffusion Tensor Imaging (DTI) has become the preferred tool to analyze white matter properties, fibre organization and mobility of the water molecules, reflected by Fractional Anisotropy (FA) and Mean Diffusivity (MD) respectively. To reduce the discrepancies between MR findings and clinical presentations we introduce a new model to analyze spinal cord images based on pattern classification. Looking at the interrelationship of quantitative MR parameters in healthy spinal cord, a pattern classification algorithm was trained in healhty subjects and tested in patients with unilateral deficits.