The purpose of this paper is to compare volume- and boundary-based morphometry methods by applying them in the statistical analysis of 2-D shapes. The methods discussed in the first category include voxel-, deformation-, and tensor-based morphometry. The active shape model is demonstrated as an example of the second category of methods. The test data are 2-D shapes of the corpus callosum (CC) obtained in patients with left-thoracic adolescent idiopathic scoliosis (AIS), as well as age and sex matched healthy participants. The features of both categories of methods and the complementarily of them are demonstrated, which may provide guidelines for their applications in medical-image analysis. The morphometric abnormality in the splenium of the CC cross-validated by different methods has potential value in the prognosis and curve prediction of the left-thoracic AIS.