Affordable Access

Access to the full text

Automated Detection and Quantification of Brain Lesions in Acute Traumatic Brain Injury Using MRI

Authors
  • Hillary, F. G.1
  • Biswal, B. B.2
  • 1 Pennsylvania State University, Psychology Department, University Park, PA, USA , University Park (United States)
  • 2 University of Medicine and Dentistry-New Jersey Medical School, Department of Radiology, Newark, NJ, USA , Newark (United States)
Type
Published Article
Journal
Brain Imaging and Behavior
Publisher
Springer-Verlag
Publication Date
Jan 24, 2009
Volume
3
Issue
2
Pages
111–122
Identifiers
DOI: 10.1007/s11682-008-9053-0
Source
Springer Nature
Keywords
License
Yellow

Abstract

The purpose of this study is to develop a reproducible method for quantifying brain lesions in traumatic brain injury (TBI). Quantifying the effects of neuropathology is an important goal in the study of brain injury and disease, yet examiners have encountered significant difficulty quantifying brain lesions in neurotrauma where there may exist multiple, overlapping forms of injury including large focal lesions and more subtle, diffuse hemorrhage and/or shear injury. In the current study, we used conventional MRI to quantify brain lesion volume at separate time points in individuals with severe TBI. We present an automated method (ISODATA) for quantifying brain lesions that is compared against a standard semi-automated volumetric approach. The ISODATA method makes no assumptions about the location or extent of brain lesions, instead identifying areas of neuropathology via voxelwise comparisons of MRI signal intensity. The data reveal that ISODATA overlaps significantly with a semi-automated approach, is reliable across multiple observations, and is sensitive to change in lesion size during recovery from TBI. This study validates a reproducible, automated lesion quantification method used here to determine the location and extent of brain pathology following TBI. This approach may be used in conjunction with advanced imaging techniques to characterize the relationship between brain lesions and neurometabolism and function.

Report this publication

Statistics

Seen <100 times