Meeting Banner
Abstract #2897

Improving the Reliability of Between Group Analyses in DTI-FA Analyses by Detecting and Removing Anatomical Anomalies

Ramtilak Gattu1, Zhifeng Kou2, Robert Welch3, Valerie Mika4, Hardik Doshi4, Ewart Mark Haacke2, Randall Benson5

1Radiology, Wayne State University, Detroit, MI, United States; 2Biomedical Engineering, Radiology, Wayne State University, Detroit, MI, United States; 3Emergency Medicine, Wayne State University School of Medicine, Detroit, MI, United States; 4Biomedical Engineering, Wayne State University, Detroit, MI, United States; 5Center for Neurological Studies, Novi, MI, United States

A method of removing anomalous anatomy is described to be applied to diffusion tensor image processing involving between subject comparison. The method first attempts to spatially register all images into a common space but in some cases the registration is strained by anatomical anomalies which would result in artifactual results if not accounted for. The method described utilizes sampled voxelwise FA variance from controls in order to exclude or mask outlier values over spatial scales consistent with anatomical anomalies. The results reported indicated the value of the method in reducing false positive artifacts in DTI analysis.

Abstract PDF Presentation Video

Keywords

aberrant abnormal accounts acquisition advanced affect analyses analytically analyze anatomical anatomy anomalies anomalous anomaly apparent applied applying around arrow axons biomedical blue brain cause caused central chronic clinical control controls corpus count create death describe detect detected detecting detection determine deviance deviations difficult diffuse diffusion disability distinguishes done eighteen emergency engineering errors even evident exclude extent failure false final flow functional general head healthy hemorrhages herein hours identify includes indicated injured injury insensitive inspection introduce largely larger lead linear maps mask masking masks match materials maximize medicine mild million mortality native neurological nonlinear normalization note notoriously overlaid partial pathology patient patients positive prevalent prolonged quite radiology recruited register registration reliability reliable remove removing respectively salvage scale scanned school score script segmentation since skeleton slice smith smoothed spatial specificity stage stat statistical studies subject subjects susceptibility system template thalamus thinning threshold thresholded tracts traumatic utilizing variance visits visual visually warping white years