Traditionally, when examining bundles, comparative methods between measures such as volume, streamline density and mean fractional anisotropy are often used. However, these measures can be biased and do not inform about shape of the white matter bundle. In this work, a new method is proposed to compare bundles based on its surface. Indeed, three measures (area, elongation, irregularity) applied along the cross-sections of the surface enable to highlight shape differences. Here, we illustrate the potential of these surface-based shape measures on the right and the left pyramidal tract in a healthy group of 30 datasets.