Spinal cord atrophy is a major determinant of physical disability in multiple sclerosis (MS) and other diseases with neurodegeneration. The upper spinal cord cross-sectional area (CSA) is therefore a clinically important measurement reflecting global spinal cord atrophy. New image analysis software enable semi- and fully-automatic quantification of spinal cord atrophy. This study characterizes the repeatability and reproducibility of semi-automatic CSA measurements of the spinal cord in healthy subjects and in patients with multiple sclerosis, using the Spinal Cord Toolbox (SCT). Results demonstrated the high repeatability and reproducibility of CSA measures using SCT in both healthy persons and in MS.