Misrepresentation of surface rendering of pediatric brain malformations performed following spatial normalization.

PURPOSE: To evaluate the effects of spatial normalization on volume rendering in cases of pediatric brain malformation.

MATERIAL AND

METHODS

Three-dimensional (3D) T1-weighted volume datasets were acquired in three children, one with pachygyria, one with a Dandy-Walker malformation associated with polymicrogyria, and one with dysgenesis of the corpus callosum. On the non-normalized datasets, the skull margins were cropped and the remainder stripped with the brain extraction technique (BET). The data were also normalized into standard anatomic reference space using pediatric templates prior to the BET script. The surface constructions obtained by both techniques were then compared for geometric distortions.

RESULTS

Normalization of 3D datasets resulted in significant distortions in the shape of the brain, with increased anterior-posterior dimensions and narrower transverse diameter in all three cases. In two cases, there were alterations in the appearance of the gyri and sulci, leading to a potential misinterpretation of the volume-rendered surface when the gyri and sulci were in fact normal.

CONCLUSION

In pediatric brain, particularly those with congenital brain anomalies, normalization as a post-processing step should be avoided as this may lead to misrepresentation of brain morphometry.