I'm using freesurfer ver 6.0.0.

If there are defects with the pial surface (i.e. the pial surface does not cover all the gray matter that it should), can you also assume that the eTIV has not been estimated incorrectly, or do pial surface inaccuracies not indicate eTIV inaccuracy? I know eTIV is estimated based on the linear transform to the talairach.xfm, but I'm not sure if there's any way to determine eTIV accuracy directly. 

I'm quality checking my images and am curious if I should remove subjects based on poor pial surfaces or if that does not matter.

Important to note, at the moment, I'm only interested in pulling hippocampal volumes and eTIV (to get relative hippocampal volumes). Similarly, would poor pial segmentation affect the reported hippocampal volumes at all?

What do you recommend?

Thanks,
Ryan
--