DIFFUSION MRI-BASED FIBER TRACKING IN HEALTHY AND BRAIN INJURY PATIENTS: A COMPARISON OF DIFFERENT SOFTWARE TOOLS

Abstract

In this work, we compared different DTI-based fibertracking software using deterministic and probabilistic approaches. DTI brain images of 35 healthy and five brain-injury patients were acquired with Philips Achieva 1.5 T scanner using an EPI-SE DTI sequence with 16 diffusion directions. Images were analyzed with Philips FiberTrack module, DTI-Studio and FSL. We studied corticospinal tract and corpus callosum, considering different termination criteria for the fibertracking algorithm. Group studies were performed to create a database of healthy patients. Results of FSL fibertracking with 1 or 2 fibers per voxel were no statistically different. T-tests between Philips and DTI-Studio led to p-values > 0.01 for corticospinal tract and < 0.01 for corpus callosum. FSL analysis led to higher ADC and lower FA values, with significative differences with the other software. In brain injury patients we measured different fibers orientation, reduced FA and increased ADC around the lesion. In conclusion, although DTI fibertracking is a promising non-invasive preoperative imaging tool, the outcome is strongly influenced by the algorithm used and the parameters chosen for the seed generation and fiber propagation.