They are the surface areas for each respective surfaces: midcortex, inner cortex, and the pial cortex.
There exists some nice studies that show that cortical surface area, pial cortex in particular, can be more sensitive to detecting differences/changes in disease in comparison to cortical thickness.
Additionally, this measure is highly correlated to grey matter volume. In my opinion, this measure may be really useful if in your dataset does not have great contrast and great tissue classification but are getting more reliable surface extractions.
Midcortex is essentially the average of the inner and pial cortex. The inner cortex would sit along the grey-white junction while the pial cortex lines the surface of the brain. Therefore, the midcortex runs along the middle of the cortical ribbon.