Cerebral Atrophy after Traumatic White Matter Injury:
Correlation with Acute Neuroimaging and Outcome
Kan Ding,1Carlos Marquez de la Plata,2,3Jun Yi Wang,4Marysa Mumphrey,1Carol Moore,1
Caryn Harper,1Christopher J. Madden,5Roderick McColl,6Anthony Whittemore,6
Michael D. Devous,6and Ramon Diaz-Arrastia1
Traumatic brain injury (TBI) is a pathologically heterogeneous disease, including injury to both neuronal cell
bodies and axonal processes. Global atrophy of both gray and white matter is common after TBI. This study was
designed to determine the relationship between neuroimaging markers of acute diffuse axonal injury (DAI) and
cerebral atrophy months later. We performed high-resolution magnetic resonance imaging (MRI) at 3 Tesla (T)
in 20 patients who suffered non-penetrating TBI, during the acute (within 1 month after the injury) and chronic
stage (at least 6 months after the injury). Volume of abnormal fluid-attenuated inversion-recovery (FLAIR) signal
seen in white matter in both acute and follow-up scans was quantified. White and gray matter volumes were
also quantified. Functional outcome was measured using the Functional Status Examination (FSE) at the time of
the chronic scan. Change in brain volumes, including whole brain volume (WBV), white matter volume (WMV),
and gray matter volume (GMV), correlates significantly with acute DAI volume (r¼?0.69, ?0.59, ?0.58, re-
spectively; p<0.01 for all). Volume of acute FLAIR hyperintensities correlates with volume of decreased FLAIR
signal in the follow-up scans (r¼?0.86, p<0.001). FSE performance correlates with acute hyperintensity volume
and chronic cerebral atrophy (r¼0.53, p¼0.02; r¼?0.45, p¼0.03, respectively). Acute axonal lesions measured
by FLAIR imaging are strongly predictive of post-traumatic cerebral atrophy. Our findings suggest that axonal
pathology measured as white matter lesions following TBI can be identified using MRI, and may be a useful
measure for DAI-directed therapies.
Key words: MR imaging; post-traumatic atrophy; TBI
approximately 1.5 million Americans suffer from TBI with 1.2
million emergency room (ER) visits and more than 50,000
deaths (Langlois et al., 2004). Severe diffuse axonal injury
(DAI) is clinically associated with loss of consciousness and
extended deep coma after brain injury, and contributes to
disability in approximately 40% of closed head injuries (Buki
et al., 2006; Meythaler et al., 2001). The pathology of DAI is
characterized histologically by Wallerian-type axonal degen-
eration in the parasagital white matter, corpus callosum, and
dorsal upper brainstem (Meythaler et al., 2001; Povlishock
raumatic brain injury (TBI) is a leading cause of
mortality and morbidity in young people. Each year,
et al., 2005; Adams et al., 1989; Strich, 1956). Current neuroi-
maging techniques are insensitive for detecting DAI (Levine
et al., 2006). Computed tomography (CT), while universally
done for the management of TBI, is not sensitive to DAI le-
sions, and only rarely shows punctate hemorrhages of small
penetrating vessels. Recently, several magnetic resonance
imaging (MRI) techniques have been used to characterize
DAI, and explore the correlation between DAI lesion burden
and long-term neurological outcomes (van der et al., 1999;
Weiss et al., 2007). Pierellini et al. (2000) reported that the total
lesion volume measured 60–90 days after injury using fluid-
attenuated inversion-recovery (FLAIR) scans correlated sig-
nificantly with 1-year Glasgow Outcome Score (GOS), while
volume of shear hemorrhage detected on fast field-echo
1Department of Neurology, University of Texas Southwestern Medical Center, Dallas, Texas.
2Center for BrainHealth?, University of Texas at Dallas, Richardson, Texas.
3Department of Psychiatry, University of Texas Southwestern Medical Center, Dallas, Texas.
4Department of Cognition and Neuroscience, The University of Texas at Dallas, Richardson, Texas.
5Department of Neurological Surgery, University of Texas Southwestern Medical Center, Dallas, Texas.
6Department of Radiology, University of Texas Southwestern Medical Center, Dallas, Texas.
JOURNAL OF NEUROTRAUMA 25:1–8 (December 2008)
ª Mary Ann Liebert, Inc.
images did not (Pierallini et al., 2000). Our group extended
these findings showing that DAI lesion volume on acute
(within 14 days after injury) FLAIR scans significantly corre-
lated with 6-month Glasgow Outcome Scale–Extended (GOS-
E) (Marquez de la Plata et al., 2007). These data indicate that
the quantification of lesion volume using FLAIR scans is a
useful tool to measure acute DAI lesion burden.
Cerebral atrophy is common after TBI (Bigler et al., 2006;
MacKenzie et al., 2002; Tomaiuolo et al., 2004, 2005; Trivedi
et al., 2007). Blatter et al. (1997) studied cerebral atrophy cross-
sectionally in 123 TBI patients grouped according to time
between injury and MRI scanning. They concluded that there
was a progressive decrease in total brain volume starting 3
weeks after moderate to severe TBI, and reaching significance
at 8–12 months later. Subsequent brain volume loss continued
at a rate greater than that seen with normal aging for up to 3
years after injury. A similar finding was reported in patients
with mild TBI (Hofman et al., 2001). Numerous studies have
focused on the relationship between post-TBI atrophy and
et al., 1997) or functional outcome (Gale et al., 2005; Himanen
et al., 2005; Hofman et al., 2001; Tomaiuolo et al., 2004, 2005).
Summarizing this extensive literature, global brain volume
loss correlated with several indicators of injury severity,
including admission GCS, duration of coma, and post-
traumatic amnesia. The association between post-traumatic
global atrophy and functional outcome is more complex, but
in general, cerebral atrophy in the chronic phase is better re-
lated to the injury severity than functional outcome (Bigler,
2001). Levine et al. (2008) demonstrated a stepwise, dose-
response relationship between TBI severity and parenchymal
volume loss using Partial Least Squares (PLS) technique (Le-
vine et al., 2008). Although cerebral atrophy has been widely
used as a gross neuropathological hallmark of diffuse injury,
especially in the subacute (>30 days after injury) and chronic
defined by abnormal FLAIR signals has not been addressed.
The purpose of the present prospective longitudinal pilot
study is to determine whether global cerebral atrophy is as-
sociated with DAI lesion load in adult patients with TBI, by
performing quantitative analysis of abnormal FLAIR signals
and global brain volumes over time. Additionally, this study
aims to determine whether TBI related cerebral atrophy cor-
relates with long-term functional outcomes.
All TBI patients were referred from the Department of
Neurological Surgery at the University of Texas (UT) South-
western Medical Center at Dallas. The inclusion criteria for
TBI patients included (1) involvement in a non-penetrating
TBI that required hospitalization; (2) age between 16 to 65
years old; (3) victims of high-speed motor vehicle=motorcycle
collision. Exclusion criteria consisted of (1) patients with focal
lesions (including contusion, extraaxial hematoma, and in-
traparenchymal hemorrhages) greater than 10ml; (2) a prior
history of TBI; (3) other conditions which may result in ab-
normal MRI findings and compromise cognitive functions
(i.e., prior brain tumor, multiple sclerosis, encephalitis=men-
meningitis, Alzheimer’s disease=mild cognitive impairment,
brain abscess, HIV encephalitis, vascular malformation,
stroke, and psychiatric disease); (4) prisoners, homeless pa-
tients, and pregnant women; and (5) patients who were not
medically stableenough tohave MRI within a month after the
injury. Twenty age- and gender-matched healthy volunteers
were recruited as controls.
was obtained from patients (or their legally authorized rep-
resentative) prior to participation in the study.
Magnetic resonance imaging
Structural MRI was performed using either a General
WI) or Siemens Trio 3T MR scanner (Siemens AG, Erlangen,
Germany). Each patient was scanned using the same scanner
for both acute and chronic time points to eliminate the influ-
ence of the different slice thickness=interslice gaps on mea-
as follows: three-dimensional (3D) T1-weighted structural
FSPGR images were obtained with slice thickness 1.3mm, the
field of view (FOV) 240mm, TR=TE 8.0=2.4ms, flip angle 25,
and NEX 2; T2 FLAIR images were acquired in the axial plane
with 3-mm slice thickness, interslice gap of 3.5mm, FOV 200–
210mm, and TR=TE=TI 9500=136.6=2500ms. For the Siemens
3T scanner: 3D T1-weighted MP-RAGE structural images
were obtained with slice thickness 1.0mm, FOV 240mm, and
TE=TI=TR 4=900=2250ms; T2 FLAIR images were acquired in
the axial plane with 2-mm slice thickness, no gap, FOV
210mm, and TR=TE=TI 9800=78=2500ms.
Diffuse axonal lesion measurements.
were converted from DICOM to ANALYZE format for fur-
ther analysis. They were then analyzed via a semi-automatic
quantification tool developed in our lab on MATLAB (Math
Works, Inc., Natick, MA). The detailed processing steps have
been described previously (Marquez de la Plata et al., 2007).
Briefly, the volume of hyperintense FLAIR signal seen in
white matter was quantified for the acute scans. The hyper-
lesion volume to whole brain volume (WBV). Both high and
low FLAIR signal seen in white matter in the follow-up scans
were quantified to create a hyperintensity lesion index and a
hypointensity lesion index, respectively. The volume of total
abnormal FLAIR lesions, including both hyperintense and
hypointense signals, on the follow-up scans were normalized
to the WBV in order to generate total lesion index.
Whole brain volume measurements.
Evaluation, using Normalization of Atrophy, Single-Time-
Point Estimation (SIENAX) within the FMRIB Software Li-
brary (FSL) was used for the automated assessment of total
brain volume, and the volume of gray and white matter. The
high-resolution T1imageswereconverted fromANALYZEto
NIFTI format. SIENAX starts the quantification process by
extracting brain from skull (Smith, 2002; Smith et al., 2002,
2004). The brain image is then affine-registered to MNI 152
space (Jenkinson et al., 2001, 2002). Next, tissue-type seg-
mentation with partial volume estimation is taken in FMRIB’s
Automated Segmentation Tool (FAST) to calculate total vol-
ume of brain tissue (including separate estimates of volumes
2DING ET AL.
of gray matter and white matter). The WBV change (WBV %)
was obtained as the ratio of the difference of WBV between
two-time points and the initial WBV. The white matter vol-
ume change (WMV %) and the gray matter volume change
(GMV %) are volume changes as percent of initial WMV and
Functional outcome measure
Functional outcome was determined at the time of the
(FSE). FSE is a semi-structured interview conducted in person
or via telephone, and typically takes 20min to complete. The
10-item interview covers a broad range of everyday activities
within physical, social, and psychological domains to deter-
mine the nature and degree of limitations that have occurred
as a result of TBI. For each category, outcome is rated along a
four-point ordinal scale:1 signifies no change from pre-injury;
2 signifies difficulty performing the activity, but maintains
total independence; 3 signifies dependence upon others to
perform the activity sometimes or a significant decrease in the
frequency of performing the activity; 4 signifies complete
dependence on others or not performing the activity at all.
Total FSE scores range from 10 to 40, with lower scores as-
sociated with better outcome. Score 41 is assigned to subjects
who die before follow-up. Dikmen et al. (2001, 2003) dem-
onstrated that FSE is more sensitive to recovery at 1–6 months
compared to GOS. Our previous study also showed that FSE
and GOS-E scores correlated well with each other (r¼?0.83),
while FSE is a more sensitive outcome measurement (Hudak
et al., 2005).
FSE and GOS-E were administered by clinical research
coordinators who were blinded to circumstances of each pa-
tient’s injury and imaging results. Inter-rater reliability for
scoring FSE was assessed byauditing 20% of the scoring sheet
every 3 months. Reproducibility was >99%. Whenever pos-
sible, the GOS-E and FSE were completed by the survivor,
although completion of the questionnaires by the caregiver
was also acceptable. Published data from our group and
others (Hudak et al., 2005; Dikmen et al., 2001) support the
validity of survivor or caregiver completion of the outcome
D’Agostino-Pearson omnibus normality test was used to
volume, and outcome measurement. Independent group
t-tests were used to compare brain volumes of normal vol-
unteers, patients in the acute phase, and patients in the
chronic phase. Paired t-tests were used to detect the change in
brain volume among patients over the time. Spearman’s
nonparametric correlations were performed to determine the
relationship among DAI lesion volume, the change of brain
volume, and outcome measure. Hierarchical regression
analysis was performed to determine factors of importance in
predicting outcome. Statistical analyses were performed
using SPSS (SPSS Inc., Chicago, IL) and GraphPad Prism
5 (GraphPad software, Inc., La Jolla, CA).
The 20 patients included in this study had a median age of
20 years (range, 16–62 years) and were predominantly male
(65%). Four patients (20%) had complicated-mild TBI, as their
first post-resuscitation GCS was above 13, but they required
hospitalization; four patients (20%) had moderate TBI with
GCS of 9–12; and 12 patients (60%) suffered severe TBI with
GCS lower than 8. Initial MRI was performed in the acute
period (median, 7 days; range, 1–35 days). The follow-up MRI
was performed with a median of 8 months after the injury
(range, 6–11 months). Age- and gender- matched normal
healthy volunteers were scanned one time using the same
protocol based on the assumption that no significant brain
volume change during a period of 1 year in healthy young
adults. Patients’ age, initial GCS, and time to follow-up scan
were normally distributed, while as expected, time to initial
scan was not. Age was normally distributed among the
sample of controls. Demographic information for both the
patient and control groups is summarized in Table 1.
Themedianhyperintensity lesion volumefor all20patients
was 1.3mm3(mean?SD, 4.5?5.5mm3). Fewer hyperintense
signal lesions were found on follow-up FLAIR scans with a
median volume of 1.2mm3(mean?SD, 1.4?1.3mm3).
Several areas of hyperintense signal seen on acute FLAIR
scans were identified as areas of decreased attenuation on
follow-up FLAIR scans, and were measured as hypointensity
lesions with a median volume of 0.19mm3(mean?SD,
0.66?1.1mm3; Fig. 1). There are no hypointense lesions large
enough to measure in the acute scans.
Table 1. Demographic Information
nMedianRangeMean SDNormality distribution
Time to initial scan (days)
Time to follow-up scan (months)
GCS at ER
Normality test was performed using D’Agostino-Pearson normality test with alpha¼0.05.
CEREBRAL ATROPHY AFTER TRAUMATIC WHITE MATTER INJURY3
The results of the brain volume measurements can be seen
in Table 2. No significant difference was found between acute
scans and normal healthy controls; however, paired t-tests
revealed significant differences between acute and follow-up
scans in terms of WBV, GMV, and WMV. Unpaired com-
volume of brain injured patients showed a non-significant
trend toward smaller volumes for TBI patients. This is likely
due to limited sample size. Seven of 20 patients (35%) had at
least 5% WBV loss.
Correlations among volumetric measurements
and functional outcomes
Spearman correlations demonstrated that WBV change,
WMV change, and GMV change were all significantly corre-
lated with the acute hyperintensity lesion index (r¼?0.69,
p¼0.001; r¼?0.59, p¼0.006; r¼?0.58, p¼0.008, respec-
tively), such that greater amounts of acute hyperintensity
FLAIR lesions correlate with greater brain volume loss after
injury (Fig. 2).
Acute hyperintensity lesion index was strongly correlated
with chronic hypointensity lesion index and total lesion index
in follow-up FLAIR scans. Additionally, the acute hyper-
intensity index was only moderately correlated with the hy-
perintensity index on follow-up FLAIR scans. Similar results
WBV. These data show that greater amounts of acute DAI
lesions are associated with greater amounts of chronically
abnormal FLAIR signal. Furthermore, the acute hyper-
intensity index was moderately correlated with first post-
resuscitation GCS score in the ER, such that greater amounts
of acute DAI are associated with lower GCS scores (Table 3).
Patients with greater degree of brain volume loss had
poorer FSE performance. Likewise, acute hyperintensity
man involved in a motorcycle collision accident. The first post-resuscitation Glasgow Coma Scale (GCS) score was 8. The first
MRI scan was performed 2 days after his injury: hyperintense fluid-attenuated inversion-recovery (FLAIR) signal lesions
were seen at corpus callosum, fornix, and subcortical white matter. The follow-up scan was performed 6 months later:
decreased attenuation FLAIR lesions were found in corpus callosum and fornix, where hyperintense lesions were noted in the
acute scans. The volume of bright FLAIR lesions in corpus callosum decreased compared to the acute scans. He lost 13% of
whole brain volume in 6 months. Functional Status Examination (FSE) score at the time of his follow-up scan was 32, and
Glasgow Outcome Scale—Extended (GOS-E) score was 5.
Magnetic resonance imaging (MRI) in a case of severe traumatic brain injury (TBI). The patient was a 20-year-old
Table 2. Brain Volumetric Measurement
Control M (SD)
Acute M (SD)
Follow-up M (SD)
(acute vs. control)
(acute vs. follow-up)
aPass D’Agostino-Pearson normality test (alpha¼0.05).
WBV, white brain volume; GMV, gray matter volume; WMV, white matter volume; M, mean; SD, standard deviation; NS, non-significant
4 DING ET AL.
lesion volume and chronic total lesion volume were signifi-
cantly correlated with FSE score, while the correlation be-
tween chronichypointensity lesionandFSEscore approached
statistical significance (p¼0.08). Similar correlations were
the ER and age were not significantly associated with func-
tional outcome (Table 4).
Multiple regression analysis
A hypothesis-driven hierarchical regression analysis was
performed to determine factors of importance in predicting
outcome. Age and first post-resuscitation GCS, which are
commonly used outcome predictors in clinical practice
(Flanagan et al., 2005; Katz et al., 1994; Livingston et al., 2005;
Marquez de la Plata CD et al., 2008; Stuss et al., 2000), were
entered in the first step of this analyses as the base model
with which all subsequent models are compared. This base
model explained 34% of the variance in FSE performance.
Subsequently, three alternative models were created by en-
tering three different variables (i.e., acute DAI lesion volume,
chronic hypointensity lesion, and brain volume change) into
the base model separately to determine whether these vari-
ables account for significant amounts of variance in FSE per-
formance. The three alternative regression models (Table 5)
each added a significant amount of explained variance in
functional outcome after TBI.
The present study is the first longitudinal study to directly
demonstrate the association between acute FLAIR hyper-
intensity lesion and post-traumatic atrophy. We excluded
patients who suffered large ormedium focal lesions including
contusions, extra- or intra- axial hematomas from the present
study in order to focus on the relationship between diffuse
can be visualized indirectly through shear hemorrhages
caused by tearing lesions of blood vessels (Scheid et al., 2003;
Tong et al., 2003) or more directly by analyzing white matter
hyperintensities on FLAIR MRI (Marquez de la Plata et al.,
Table 3. Univariate Correlations with Acute
index on F=U
volume on F=U
index on F=U
volume on F=U
Total lesion index on F=U
Total lesion volume on F=U
Correlations among acute lesion volumes and chronic lesion
Significant correlation in bold.
Table 4. Correlations Among Injury Severity
Indicators and FSE
Acute hyperintensity lesion
Hypointensity lesion on F=U
Hyperintensity lesion on F=U
Total lesion on F=U
Correlations between potential injury severity indicators and
Significant correlations in bold.
F=U, follow-up; WBV%, whole brain volume change (%); WMV%,
white matter volume change (%); GMV%, gray matter volume
change (%); total lesion on F=U¼(Hypointensity lesion on F=Uþ
Hyperintensity lesion on F=U).
Table 5. Hierarchical Regression Model
for Predicting FSE Performance
Models R square R square change Significant F change
Alt model 1
Alt model 2
Alt model 3
Base Predictors: ER_GCS, AGE.
1 Predictors: ER_GCS, AGE, acute hyperintensity lesion volume.
2 Predictors: ER_GCS, AGE, follow-up hypointensity lesion
3 Predictors: ER_GCS, AGE, whole brain volume change.
Relative contributions of explained variance in functional out-
attenuated inversion-recovery (FLAIR) hyperintense lesion
volume and 6-month post-traumatic whole brain volume
(WBV) change. The WBV change (%) is a ratio of the differ-
ence of WBV between two time points and the initial WBV.
Hyperintensity lesion index is a ratio of the volume of hy-
perintensity lesion to the WBV. Correlation was performed
using Spearman’s nonparametric correlation coefficient (r).
Scatter plot of the association between acute fluid-
CEREBRAL ATROPHY AFTER TRAUMATIC WHITE MATTER INJURY5
2007; Pierallini et al., 2000; Takaoka et al., 2002). We found
that acute hyperintensity FLAIR lesions are strongly predic-
tive of post-traumatic cerebral atrophy. Although the patho-
physiology of the post-TBI cerebral atrophy remains
unknown, axonal injury and subsequent Wallerian degener-
ation may be a possible mechanism. The current working
hypothesis, based on animal data (Povlishock et al., 2005;
non-penetrating brain injury), the axons swell due to the local
ionic homeostatic disruption, increased permeability of the
axolemma, with immediate mechanical damage to the axonal
cytoskeleton (primary axotomy) seen in severe cases. Days
to months after the injury, pathological changes of DAI are
believed to consist of progressive disorganization of the ax-
onal cytoskeleton and progressive protein accumulation,
leading to disconnection of axons (secondary axotomy). The
primary and secondary axotomy also triggers the local or
even global metabolic changes which would lead to further
phase after TBI. Our data is consistent with this working
hypothesis, and also indicates that tracking of macroscopic
lesions visible in FLAIR scans may be a useful method to
monitor progressive tissue pathology associated with DAI.
The current study confirmed our prior finding that acute
hyperintensity FLAIR lesion was an important factor for
predicting functional outcome, though only moderately cor-
related with FSE (Marquez de la Plata et al., 2007). Combined
with GCS in the ER and age, the acute DAI lesion volume
could be used to stratify injury severity when selecting
patients for TBI clinical trials. Our study indicates that the
follow-up MRI, especially on high magnetic field, could offer
useful information on the pathological change of DAI, which
would be potentially useful in DAI-directed therapies.
While FLAIR is a commonly used clinical sequence which
all physicians are familiar with, it may not be the most sen-
sitive MR sequence to detect DAI. Several other MR tech-
niques have been proposed to increase the sensitivity of
detecting DAI in vivo. T2-weighted gradient echo is excellent
in detecting acute small punctuate hemorrhagic lesions. The
number of traumatic microbleeds detected on T2-weighted
gradient echo sequence at chronic stage ($3 months after TBI)
correlated significantly with GCS, but not with long-term
outcome measured by GOS-E (Pierallini et al., 2000; Scheid
et al., 2003). A new high-resolution 3D gradient-echo MR
imaging technique, known as susceptibility-weighted im-
aging (SWI) is much more sensitive than conventional T2-
weighted gradient-echo sequences in detecting hemorrhagic
DAI. Number of traumatic microhemmorhagic lesion de-
tected bySWIcorrelated better withGOS-E than that detected
by gradient echo (Tong et al., 2003, 2004, 2008). No longitu-
dinal data is available to assess whether SWI lesions correlate
with post-traumatic brain atrophy. Diffusion-weighted im-
aging (DWI) has proven to be highly sensitive for the detec-
tion of early cytotoxic edema in the setting of acute stroke.
DWI has not been widely used in clinical TBI, though the
sensitivity of DWI to identify DAI lesions is similar to that of
FLAIR. It is less sensitive than T2 gradient echo for detecting
hemorrhagic lesions (Huisman, 2002; Huisman et al., 2003;
Kinoshita et al., 2005). The volume of DWI lesions in white
matter is moderately correlated with functional outcome
imaging (DTI) characterizes the diffusion of water along
white matter tracts. Our group and others are studying DTI in
the hope that it would be more sensitive to axonal pathology
after traumatic injury (Bazarian et al., 2007; Kim et al., 2008;
Sidaros et al., 2008; Wang et al., 2008). DTI is able to visualize
changes that were not seen on conventional scans and
strongly correlated with functional outcome. However the
analysis of DTI data is time-consuming, experience-depen-
dent and may not be ideally suitable for routine clinical
practice. Overall, the combination of FLAIR and other MRI
sequences may provide additional information about injury
severity and correlate with the functional outcome better.
Our findings of reduced WBV, WMV, and GMV at ap-
proximately 8 months after TBI are consistent with previous
volumetric studies (MacKenzie et al., 2002; Trivedi et al.,
2007). This indicated that our method of brain volume mea-
surement is reliable and consistent. In our study, SIENAX, a
fully automated method, has been used to measure atrophy.
This program has been shown to be an accurate approach to
measure cross-sectional normalized brain volume with 0.5–
1% brain volume accuracy for a single-time point. It has
successfully coped with both de-skulling and tissue segmen-
tation, and can be used for the subjects with extreme paren-
chymal loss (Smith et al., 2002). Recently, its reliability has be
demonstrated in other neurological conditions, such as mul-
tiple sclerosis (MS) and Alzheimer’s disease (Anderson et al.,
2006, 2007; Smith et al., 2007). A particular advantage of
SIENAX is that it is relatively insensitive to different scanning
parameters. For MS-related brain atrophy, the inter-center
agreement assessed with the concordance correlation coeffi-
cient was 0.94 between two centers regardless of the differ-
ence of magnetic field strength and scanning parameters
(Jasperse et al., 2007). SIENA is a similar fully automated
program developed to measure longitudinal atrophy rate.
Our finding of cerebral atrophy measured using SIENAX was
commensurate to the results of Trivedi et al. (2007) using
SIENA. The development of SIENA=SIENAX makes it prac-
tical to use atrophy rate=state as an index of disease pro-
gression in clinical studies.
The present investigation was a pilot study, and a larger
scale investigation is under way in our group to confirm these
pilot results. While many initial scans took place within the
first week, many occurred several days after injury. The het-
erogeneity in the interval from injury to first scan may con-
found results, as it is possible that FLAIR lesions increase in
conspicuity over time. In order to establish the utility of MRI
as a biomarker in clinical trials, scans obtained within 24h of
injury must be studied.Furthermore, neuropathologic studies
are needed to correlate the cellular and tissue abnormalities
with lesions detected by MRI.
The present study provides in vivo evidence of pathological
change of DAI after TBI. It demonstrates a strong association
presented in this study could be a practical method to monitor
the efficacy of DAI-directed therapies in future clinical trials.
The present study was supported by the U.S. Department
of Education (grant NIDRR H133 A020526) and the National
Institutes of Health (grants NIH R01 HD48179 and NIH U01
6 DING ET AL.
Author Disclosure Statement
No competing financial interests exist.
Adams, J.H., Doyle, D., Ford, I., Gennarelli,T.A., Graham, D.I.,
and McClellan, D. (1989). Diffuse axonal injury in head injury:
Definition, diagnosis, and grading. Histopathology 15, 49–59.
Anderson, V.M., Fernando, K.T., Davies, G.R., Rashid, W., Frost,
C., Fox, N.C., and Miller, D.H. (2007). Cerebral atrophy
measurement in clinically isolated syndromes and relapsing
remitting multiple sclerosis: a comparison of registration-
based methods. J. Neuroimaging 17, 61–68.
Anderson, V.M., Fox, N.C., and Miller, D.H. (2006). Magnetic
resonance imaging measures of brain atrophy in multiple
sclerosis. J. Magn. Reson. Imaging 23, 605–618.
Ayala, R.R., Redfern, S., Ding, K., Moore, M., Harper, C., Mad-
den, C.J., and Diaz-Arrastia, R. (2008). Diffusion-weighted MR
imaging in closed head injury: identification of distinct injury
subtypes and correlation with functional outcome. Neurology
Bazarian, J.J., Zhong, J., Blyth, B., Zhu, T., Kavcic, V., and Pe-
terson, D. (2007). Diffusion tensor imaging detects clinically
important axonal damage after mild traumatic brain injury: a
pilot study. J. Neurotrauma 24, 1447–1459.
Bigler, E.D. (2001). Quantitative magnetic resonance imaging in
traumatic brain injury. J. Head Trauma Rehabil. 16, 117–134.
Bigler, E.D., Ryser, D.K., Gandhi, P., Kimball, J., and Wilde, E.A.
(2006). Day-of-injury computerized tomography, rehabilita-
tion status, and development of cerebral atrophy in persons
with traumatic brain injury. Am. J. Phys. Med. Rehabil. 85,
Blatter, D.D., Bigler, E.D., Gale, S.D., Johnson, S.C., Anderson,
C.V., Burnett, B.M., Ryser, D., MacNamara, S.E., and Bailey,
B.J. (1997). MR-based brain and cerebrospinal fluid measure-
ment after traumatic brain injury: correlation with neu-
ropsychological outcome. Am. J. Neuroradiol. 18, 1–10.
Buki, A. and Povlishock, J.T. (2006). All roads lead to discon-
nection?–traumatic axonal injury revisited. Acta Neurochir.
(Wien.) 148, 181–193.
Dikmen, S., Machamer, J., Miller, B., Doctor, J., and Temkin, N.
(2001). Functional status examination: a new instrument for
assessing outcome in traumatic brain injury. J. Neurotrauma
Dikmen, S.S., Machamer J.E., Powell, J.M., and Temkin, N.R.
(2003). Outcome 3 to 5 years after moderate to severe trau-
matic brain injury. Arch. Phys. Med. Rehabil. 84, 1449–1457.
Flangan, S.R., Hibbard, M.R., and Gordon, W.A. (2005). The
impact of age on traumatic brain injury. Phys. Med. Rehabil.
Clin. North Am. 16, 163–177.
Gale, S.D., Baxter, L., Roundy, N., and Johnson, S.C. (2005).
Traumatic brain injury and grey matter concentration: a pre-
liminary voxel based morphometry study. J. Neurol. Neuro-
surg. Psychiatry 76, 984–988.
Himanen, L., Portin, R., Isoniemi, H., Helenius, H., Kurki, T., and
Tenovuo, O. (2005). Cognitive functions in relation to MRI find-
ings 30 years after traumatic brain injury. Brain Inj. 19, 93–100.
Hofman, P.A., Stapert, S.Z., Van Kroonenburgh, M.J., Jolles, J.,
De Kruijk, J., and Wilmink, J.T. (2001). MR imaging, single-
photon emission CT, and neurocognitive performance after
mild traumatic brain injury. Am. J. Neuroradiol. 22, 441–449.
Hudak, A.M., Caesar, R.R., Frol, A.B., Krueger, K., Harper, C.R.,
Temkin, N.R., Dikmen, S.S., Carlile, M.C., Madden, C.,
and Diaz-Arrastia, R. (2005). Functional outcome scales in
traumatic brain injury: a comparison of the Glasgow Outcome
Scale (Extended) and the Functional Status Examination.
J. Neurotrauma 22, 1319–1326.
Housman, T.A., Sorensen, A.G., Hergan, K., Gonzalez, R.G., and
Schaefer, P.W. (2003). Diffusion-weighted imaging for the
evaluation of diffuse axonal injury in closed head injury.
J. Comput. Assist. Tomogr. 27, 5–11.
Huisman, T.A.G.M. (2002). Diffusion-weighted imaging: basic
concepts and application in cerebral stroke and head trauma.
Eur. Radiol. 13, 2283–2297.
Jasperse, B., Valsasina, P., Neacsu, V., Knol, D.L., De Stefano, N.,
Enzinger, C., Smith, S.M., Ropele, S., Korteweg, T., Giorgio,
A., Anderson, V., Polman, C.H., Flippi, M., Miller, D.H.,
Rovaris, M., Barkhof, F., and Vrenken, H. (2007). Intercenter
agreement of brain atrophy measurement in multiple sclerosis
patients using manually-edited SIENA and SIENAX. J. Magn.
Reson. Imaging 26, 881–885.
Jenkinson, M., Bannister, P., Brady, M., and Smith, S. (2002).
Improved optimization for the robust and accurate linear reg-
istration and motion correction of brain images. Neuroimage.
Jenkinson, M., and Smith, S. (2001). A global optimisation
method for robust affine registration of brain images. Med.
Image Anal. 5, 143–156.
Katz, D., and Alexander, M.P. (1994). Predicting course of re-
covery and outcome for patients admitted to rehabilitation.
Arch. Neurol. 51, 661–670.
Kim, J., Avants, B., Patel, S., Whyte, J., Coslett, B.H., Pluta, J.,
Detre, J.A., and Gee, J.C. (2008). Structural consequences of
diffuse traumatic brain injury: a large deformation tensor-
based morphometry study. Neuroimage 39, 1014–1026.
Kinoshita, T., Moritani, T., Hiwatashi, A., Wang, H.Z., Shrier,
D.A., Numaguchi, Y., and Westesson, P.L. (2005). Conspicuity
of diffuse axonal injury lesions on diffusion-weighted MR
imaging. Eur. J. Radiol. 56, 5–11.
Langlois, J.A., Keller, M., Butler, J., Gotsch, K.E., Johnson, R.L.,
Reichard, A.A., Webb, K.W., Coronado, V.G., Selassie, A.W.,
and Thurman, D.J. (2003). Traumatic brain injury-related
hospital discharges. Results from a 14-state surveillance sys-
tem, 1997. MMWR Surveill. Summ. 52, 1–20.
Langlois, J.A., Rutland-Brown, W., and Thomas, K. Traumatic
Brain Injury in the United States: Emergency Department visits,
Hospitalizations, and Deaths. National Center for Injury Pre-
vention and Control: Atlanta.
Levine, B., Fujiwara, E., O’Connor, C., Ricard, N., Kovacevic, N.,
Mandic, M., Restagno, A., Easdon, C., Robertson, I.H., Gra-
ham, S.J., Cheung, G., Gao, F., Schwartz, M.L., and Black, S.E.
(2006). In vivo characterization of traumatic brain injury neu-
ropathology with structural and functional neuroimaging.
J. Neurotrauma 23, 1396–1411.
Levine, B., Kovacevic, N., Nica, E.I., Cheung, G., Gao, F.,
Schwartz, M.L., and Black, S.E. (2008). The Toronto traumatic
brain injury study: injury severity and quantified MRI. Neu-
rology 70, 771–778.
Livingston, D.H., Lavery, R.F., Mosenthal, A.C., Knudson, M.M.,
Lee, S., Morabito, D., Manley, G.T., Nathens, A., Jurkovich, G.,
Hoyt, D.B., and Coimbra, R. (2005). Recovery at one year
following isolated traumatic brain injury: a Western Trauma
Association prospective multicenter trial. J. Trauma 59, 1298–
Mackenzie, J.D., Siddiqi, F., Babb, J.S., Bagley, L.J., Mannon, L.J.,
Sinson, G.P., and Grossman, R.I. (2002). Brain atrophy in mild
or moderate traumatic brain injury: a longitudinal quantita-
tive analysis. AJNR Am. J. Neuroradiol. 23, 1509–1515.
CEREBRAL ATROPHY AFTER TRAUMATIC WHITE MATTER INJURY7
Marquez de la Plata, C.D., Hart, T., Hammond, F.M., Frol, A.B., Download full-text
Hudak, A., Harper, C.R., O’Neil-Pirozzi, T.M., Whyte, J.,
Carlile, M., and Diaz-Arrastia, R. (2008). Impact of age on
long-term recovery from traumatic brain injury. Arch. Phys.
Med. Rehabil. 89, 896–903.
Marquez de la Plata, C., Ardelean, A., Koovakkattu, D., Srini-
vasan, P., Miller, A., Phuong, V., Harper, C., Moore, C.,
Whittemore, A., Madden, C., Diaz-Arrastia, R., and Devous,
M. (2007). Magnetic resonance imaging of diffuse axonal in-
jury: quantitative assessment of white matter lesion volume. J.
Neurotrauma 24, 591–598.
Meythaler, J.M., Peduzzi, J.D., Eleftheriou, E., and Novack, T.A.
(2001). Current concepts: Diffuse axonal injury-associated
Pierallini, A., Pantano, P., Fantozzi, L.M., Bonamini, M., Vichi,
R., Zylberman, R, Pisarri, F., Colonnese, C., and Bozzao, L.
(2000). Correlation between MRI findings and long-term out-
come in patients with severe brain trauma. Neuroradiology
Povlishock, J.T., and Katz, D.I. (2005). Update of neuropathology
and neurological recovery after traumatic brain injury. J. Head
Trauma Rehabil. 20, 76–94
Scheid, R., Preul, C., Gurber, O., Wiggins, C., and Von Cramon,
D.Y. (2003). Diffuse axonal injury associated with chronic
traumatic brain injury: evidence from T2*-weighted gradient-
echo imaging at 3T. AJNR Am. J. Neuroradiol. 24, 1049–1056.
Sidaros, A., Engberg, A.W., Sidaros, K., Liptrot, M.G., Herning,
M., Petersen, P., Paulson, O.B., Jernigan, T.L., and Rostrup, E.
(2008). Diffusion tensor imaging during recovery from severe
traumatic brain injury and relation to clinical outcome: a
longitudinal study. Brain 131, 559–572.
Smith,D.H.,Meaney, D.F.,andShull,W.H. (2003).Diffuseaxonal
injury in head trauma. J. Head Trauma Rehabil. 18, 307–306.
Smith, S.M. (2002). Fast robust automated brain extraction.
Hum. Brain Mapp. 17, 143–155.
Smith, S.M., Jenkinson, M., Woolrich, M.W., Bechmann, C.F.,
Behrens, T.E., Johansen-Berg, H., Bannister, P.R., De Luca, M.,
Drobnjak, I., Flitney, D.E., Niazy, R.K., Saunders, J., Vickers, J.,
Zhang, Y., De Stefano, N., Brady, J.M., and Matthews, P.M.
(2004). Advances in functional and structural MR image
analysis and implementation as FSL. Neuroimage 23, Suppl 1,
Smith, S.M., Rao, A., De Stefano, N., Jenkinson, M., Schott, J.M.,
Matthews, P.M., and Fox, N.C. (2007). Longitudinal and
cross-sectional analysis of atrophy in Alzheimer’s disease:
cross-validation of BSI, SIENA and SIENAX. Neuroimage 36,
Smith, S.M., Zhang, Y., Jenkinson, M., Chen, J., Matthews, P.M.,
Federico, A., and De Stefano, N. (2002). Accurate, robust, and
automated longitudinal and cross-sectional brain change
analysis. Neuroimage 17, 479–489.
Strich, S.J. (1956). Diffuse degeneration of cerebral white matter
in severe dementia following head injury. J. Neurol. Neuro-
surg. Psychiatry 19, 163–185.
Stuss, D.T., Binns, M.A., Carruth, F.G., Levine, B., Brandys, C.F.,
Moulton, R.J., Snow, W.G., and Schwartz, M.L. (2000). Pre-
diction of recovery of continuous memory after traumatic
brain injury. Neurology 54, 1337–1344.
Takaoka, M., Tabuse, H., Kumura, E., Nakajima, S., Tsuzuki,
T., Nakamura, K., Okada, A., and Sugimoto, H. (2002).
Semiquantitative analysis of corpus callosum injury using
magnetic resonance imaging indicates clinical severity in pa-
tients with diffuse axonal injury. J. Neurol. Neurosurg. Psy-
chiatry 73, 289–293.
Tomaiuolo, F., Carlesimo, G.A., Di Paola, M., Petrides, M., Fera,
F., Bonanni, R., Formisano, R., Rasqualetti, P., and Caltagir-
one, C. (2004). Gross morphology and morphometric sequelae
in the hippocampus, fornix, and corpus callosum of patients
with severe non-missile traumatic brain injury without mac-
roscopically detectable lesions: a T1-weighted MRI study.
J. Neurol. Neurosurg. Psychiatry 75, 1314–1322.
Tomaiuolo, F., Worsley, K.J., Lerch, J., Di Paola, M., Carlesimo,
G.A., Bonanni, R., Caltagirone, C., and Paus, T. (2005).
Changes in white matter in long-term survivors of severe non-
missile traumatic brain injury: a computational analysis of
magnetic resonance images. J. Neurotrauma 22, 76–82.
Tong, K.A., Ashwal, S., Holshouser, B.A., Nickerson, J.P., Wall,
C.J., Shutter, L.A., Osterdock, R.J., Haacke, E.M., and Kido, D.
(2004). Diffuse axonal injury in children: clinical correlation
with hemorrhagic lesions. Ann. Neurol. 56, 36–50.
Tong, K.A., Ashwal, S., Holshouser, B.A., Shutter, L.A.,
Herigault, G., Haacke, E.M., and Kido, D. (2003). Hemorrhagic
shearing lesions in children and adolescents with post-trau-
matic diffuse axonal injury. Radiology 227, 332–339.
Tong, K.A., Ashwal, S., Obenaus, A., Nickerson, J.P., Kido, D.,
and Haacke, E.M. (2008). Susceptibility-weighted MR imag-
ing: a review of clinical applications in children. AJNR Am.
J. Neuroradiol. 29, 9–17.
Trivedi, M.A., Ward, M.A., Hess, T.M., Gale, S.D., Dempsey,
R.J., Rowley, H.A., and Johnson, S.C. (2007). Longitudinal
changes in global brain volume between 79 and 409 days after
traumatic brain injury: relationship with duration of coma.
J. Neurotrauma 24, 766–771.
Van Der, N.J., Hew, J.M., Van Zomeren, A.H., Sluiter, W.J., and
Minderhoud, J.M. (1999). Computed tomography and mag-
netic resonance imaging in mild to moderate head injury:
early and late imaging related to outcome. Ann. Neurol. 46,
Wang, J.Y., Bakhadirov, K., Devous, M., Abdi, H., McColl, R.,
Moore, C., Marquez de la Plata, C., Ding, K., Whittemore, A.,
Babcock, E., Rickbeil, T., Dobervich, J., Kroll, D., Dao, B.,
Mohindra, N., and Diaz-Arrastia, R. (2008). Diffusion tensor
tractography in traumatic diffuse axonal injury. Arch. Neurol.
Weiss, N., Galanaud, D., Carpenter, A., Naccache, L., and Puy-
basset, L. (2007). Clinical review: prognostic value of magnetic
resonance imaging in acute brain injury and coma. Crit. Care
Wilde, E.A., Bigler, E.D., Pedroza, C., and Ryser, D.K. (2006).
Post-traumatic amnesia predicts long-term cerebral atrophy in
traumatic brain injury. Brain Inj. 20, 695–699.
Address reprint requests to:
Ramon Diaz-Arrastia, M.D.
Department of Neurology
UT Southwestern Medical Center at Dallas
5323 Harry Hines Boulevard
Dallas, TX 75390-9036
8 DING ET AL.