Emerg Radiol

DOI 10.1007/s10140-016-1440-z

ORIGINAL ARTICLE

Reformatted images improve the detection rate of acute

traumatic subdural hematomas on brain CT compared with axial
images alone
Timothy J. Amrhein 1,2 & William Mostertz 2 & Maria Gisele Matheus 2 &
Genevieve Maass-Bolles 2 & Komal Sharma 2,3 & Heather R. Collins 4 & Peter G. Kranz 1

Received: 10 August 2016 / Accepted: 1 September 2016

# American Society of Emergency Radiology 2016

Abstract Subdural hematomas (SDHs) comprise a signifi- sagittal reformatted images to an axial brain CT increases
cant percentage of missed intracranial hemorrhage on axial the sensitivity and specificity for detection of acute traumatic
brain CT. SDH detection rates could be improved with the SDH. We retrospectively reviewed consecutive brain CTs ac-
addition of reformatted images. Though performed at some quired for acute trauma that contained new SDHs. An equiv-
centers, the potential additional diagnostic sensitivity of alent number of normal brain CTs served as control. Paired
reformatted images has not yet been investigated. The purpose sets of images were created for each case: (1) axial images
of our study is to determine if the addition of coronal and only (Baxial only^) and (2) axial, coronal, sagittal images
(Breformat added^). Three readers interpreted both the axial
only and companion reformat added for each case, separated
* Timothy J. Amrhein
by 1 month. Reading times and SDH detection rates were
timothy.amrhein@duke.edu compared. One hundred SDH and 100 negative examinations
were collected. Sensitivity and specificity for the axial-only
William Mostertz scans were 75.7 and 94.3 %, respectively, compared with 88.3
wmostert@wakehealth.edu and 98.3 % for reformat added. There was a 24.3 % false
Maria Gisele Matheus negative (missed SDH) rate with axial-only scans versus
matheus@musc.edu 11.7 % with reformat added (p = <0.001). Median reader
Genevieve Maass-Bolles interpretation times were longer with the addition of
maass@musc.edu reformatted images (125 versus 89 s), but this difference was
not significant (p = 0.23). The addition of coronal and sagittal
Komal Sharma
drkomal_sharma@yahoo.in images in trauma brain CT resulted in improved sensitivity
and specificity as well as a reduction in SDH false negatives
Heather R. Collins
by greater than 50 %. Reformatted images substantially re-
collinshr@gmail.com
duce the number of missed SDHs compared with axial images
Peter G. Kranz alone.
peter.kranz@duke.edu

1
Keywords Subdural hematoma . Brain CT . Sensitivity .
Department of Radiology, Duke University Medical Center|, Box
3808, Durham, NC 27710, UK
Specificity . Reformat
2
Department of Radiology and Radiological Science, Medical
University of South Carolina, MSC 322 169 Ashley Avenue,
Charleston, SC 29425, USA Introduction
3
Department of Radiology, St. Vincents Medical Center, Quinnipiac
School of Medicine, 2800 Main Street, Bridgeport, CT 06606, USA There are more than 1.5 million annual emergency department
4
Center for Biomedical Imaging, Department of Radiology and
visits in the USA for acute head trauma [1]. Among these
Radiological Science, Medical University of South Carolina, 96 patients, subdural hematomas (SDHs) are the reason for ap-
Jonathan Lucas St., MSC 323, Charleston, SC 29425-0323, USA proximately 50 % of hospital admissions and are the most

common indication for intensive care unit admission [2].
Given the high prevalence of SDH in acute head trauma as
well as its potential severity, it is particularly concerning that
SDHs have been shown to comprise a significant percentage
(and often the majority) of missed intracranial hemorrhage on
brain CT [3, 4].
The American College of Radiology recommends that
brain CT for the evaluation of acute head trauma be per-
formed with contiguous or overlapping axial slices of a
thickness no greater than 5 mm [5]. There is no recommen-
dation for the inclusion of coronal or sagittal reformatted
images. However, the orientation of axial images may re-
duce their effectiveness for identifying pathology in the
transverse plane [4]. Further, reformatted images have been
shown to increase reader diagnostic confidence and sensitiv-
ity in the detection of pathology in many areas of radiology
including musculoskeletal trauma, chest trauma, appendici-
tis, small bowel obstruction, hepatocellular carcinoma, and
evaluation of the urinary tract [612]. A limited number of
prior investigations have supported the concept that
reformatted images will improve pathology detection rates
in acute traumatic brain injury. However, none have specif-
ically investigated the detection of SDHs and all have in-
cluded only a relatively small cohort of patients [4, 13, 14].
The purpose of this study is to determine if the addition of
coronal and sagittal reformatted images to an axial brain CT
increases the sensitivity and specificity for the detection of
SDHs.
Methods
Study group
This is a retrospective study of the diagnostic accuracy of axial
images (Baxial-only^ group) for the detection of acute SDH
compared to the combination of axial images with coronal
and sagittal reformats (Breformat-added^ group). Based on pre-
viously reported rates of SDH detection, we calculated that a
sample size of 100 patients in each arm (i.e., cases positive for
SDH and negative controls) would result in a greater than 80 %
power to identify a difference between SDH detection rates
[14]. Final diagnostic imaging reports from consecutive brain
CTs acquired for an indication of acute trauma were screened to
identify possible cases of SDH. An equivalent number of nor-
mal brain CTs were identified to serve as a control.
Examinations were excluded if they contained severe patient
motion artifact precluding accurate interpretation, did not con-
tain a full set of axial, coronal, and sagittal images, had pene-
trating head injuries (bullet shrapnel, etc.), or had surgical inter-
vention for intracranial hemorrhage prior to scan acquisition.
In order to establish the reference standard, an expert reader
(a board-certified radiologist with 8 years of experience
Emerg Radiol
interpreting brain CT who also holds a Certificate of Added
Qualification (CAQ) in neuroradiology) who was unblinded
to the intent of the investigation, reviewed both the axial and
reformatted images of all studies in order to determine the
presence or absence of SDH. Cases in which the neuroradiol-
ogists interpretation disagreed with the final diagnostic report
were adjudicated by a second expert reader (a board-certified
radiologist with a CAQ in neuroradiology and 11 years of
experience interpreting brain CT).
This study was approved by our institutional review board
with a waiver of the need to obtain informed consent and was
compliant with the Health Insurance Portability and
Accountability Act.
Imaging procedure
All brain CTs had been previously acquired using the same
protocol on a single 128-MDCT scanner (SOMATOM
Definition, Siemens). Images were acquired using helical
technique from the base of skull through the vertex with the
following scan parameters: 120 kVp, 420 mAs, pitch 0.55,
40 0.625 mm collimation, field of view 20 cm, and matrix
512 512. Axial images were 5 mm in thickness, and coronal
and sagittal reformations were generated at 2-mm thickness,
all previously constructed as part of routine clinical practice
using the standard software algorithm on the CT scanner.
Reading sessions
Three readers of varying skill levels were recruited to test the
diagnostic accuracy of the axial images (axial-only group) com-
pared to the entire data set including the reformatted images
(reformat-added group). The readers included a CAQ-certified
attending neuroradiologist, a neuroradiology fellow, and a post-
graduate year 4 radiology resident. The readers were blinded to
the purpose of the study, told only that the indication for the
exam was Btrauma,^ and were asked to provide interpretations.
Using three-dimensional reformatting software
(AquariusNet Viewer, Version 4.4; TeraRecon, Inc., Foster
City, CA), all examinations were anonymized and a paired
set of images was created for each case: one containing axial
images only (axial-only group) and one containing axial im-
ages with coronal and sagittal reformats (reformat-added
group). There were 10 reading sessions per reader. Each read-
ing session consisted of 40 cases (20 normal and 20 positive
for SDH) in random order. A single reading session contained
only either axial-only images or reformat-added images. At
the end of the 10 reading sessions, all readers had interpreted
the axial-only exam and the companion reformat-added exam
for every case. Paired axial-only and reformat-added exams
for each individual case were read at least 1 month apart to
reduce the potential for recall bias.
Emerg Radiol

The interpretations of the readers were recorded for all Table 2 False negatives (missed subdural hematomas)
sessions by the same independent observer who specifically Reader Axial alone Axial + reformats p values
noted whether SDHs were identified. Total reading times for
each session were recorded. Percentage Proportion Percentage Proportion

1 (Attending) 19 19/100 8 8/100 <0.001

Statistical analysis 2 (Fellow) 34 34/100 18 18/100 <0.001
3 (Resident) 20 20/100 9 9/100 <0.001
Total (mean) 24.3 73/300 11.7 35/300 <0.001
The rates of false negatives or false positives for the axial-only
and reformat-added groups were assessed using either chi-
square tests or Fischers exact tests, where appropriate. The
proportion of false negatives and false positives in both groups neurosurgical intervention. A total of 200 examinations were
for each reader was assessed using an exact McNemars test. collected (100 negative, 100 containing SDHs) within a peri-
The patient ages as well as the width of missed SDHs (false od of 12 months. The mean age of the study group was
negatives) were compared between the axial-only and 39.6 years (SD, 17.6) for negative examinations and 50.3
reformat-added groups using t tests. Association between false (SD, 22.4) for positive examinations (p < 0.001). There was
negative SDH location and each group was assessed using a no difference in gender distribution between the negative
binomial test. Interrater agreement was calculated using group (42 females) and the subdural group (40 females)
Fleisss Kappa. Median reading times for the axial-only and (p = 0.89).
reformat-added groups were compared using a Wilcoxon
signed-rank test. All statistical analyses were conducted using Diagnostic performance
SPSS 22 (IBM; Armonk, NY). Statistical significance was
considered at a threshold of p < 0.05. Overall sensitivity and specificity for the detection of SDH for
the axial-only scans were 75.7 and 94.3 %, respectively, com-
pared with 88.3 and 98.3 % for reformat-added scans
Results (Table 1). Overall, there was a 24.3 % false negative (missed
SDH) rate with axial-only scans versus 11.7 % with reformat
Study group added (p = <0.001). There were 17 false positives with axial-
only scans versus 5 with reformat added (p = <0.001).
During the screening process, six SDH cases were excluded False negative rates and number of false positives for each
due to penetrating head injuries or prior emergent reader are displayed in Tables 2 and 3. There was moderate

Table 1 Diagnostic Performance for the Detection of Subdural Hematomas

Axial
Reader Sensitivity Specificity Diagnostic Accuracy
Percentage Proportion Percentage Proportion Percentage Proportion
1 81 (71.787.9) 81/100 93 (85.696.9) 93/100 87 174/200
(Attending)
2 66 (55.875.0) 66/100 95 (88.298.1) 95/100 80.5 161/200
(Fellow)
3 80 (70.687.1) 80/100 95 (88.298.1) 95/100 87.5 175/200
(Resident)
Overall 75.7 (65.983.5) 227/300 94.3 (87.397.7) 283/300 85 510/600

Axial + Reformat
Reader Sensitivity Specificity Diagnostic Accuracy
Percentage Proportion Percentage Proportion Percentage Proportion
1 92 (84.896.5) 92/100 98 (92.399.7) 98/100 95 190/200
(Attending)
2 82 (72.888.7) 82/100 98 (92.399.7) 98/100 90 180/200
(Fellow)
3 91 (83.295.5) 91/100 99 (93.899.9) 99/100 95 190/200
(Resident)
Overall 88.3 (80.093.6) 265/300 98.3 (92.799.8) 295/300 93.3 560/600

Note: 95% confidence intervals within parentheses


Table 3 Number of false positives
Reader Axial Axial +
alone (n) reformats (n)
1 (Attending) 7 2
2 (Fellow) 5 2
3 (Resident) 5 1
Total (mean) 5.7 1.7
inter-reader agreement for false negatives in the axial-only
group (k = 0.45) and fair agreement for false negatives in the
reformat-added group (k = 0.33).
There was no significant difference in median interpre-
tation times per scan between the reformat-added sessions
(125 s) and the axial-only sessions (89 s, p = 0.23). On
average, there was a 21.6 % increase in mean interpreta-
tion time with the reformat-added images compared with
the axial-only images.
Characteristics of missed subdural hematomas
While there were a wide variety of SDH locations in false
negative cases, those overlying the cerebral convexities
accounted for the majority in both imaging data sets
(Table 4). There was no significant difference in the preva-
lence of particular locations of missed SDHs between false
negative cases in the axial-only and reformat-added groups
(p = 0.67). However, there was a trend toward a slightly larger
width of SDH with false negative cases in the axial-only scan
group than in the reformat-added group, 2.7 mm (SD, 1.3) and
2.3 mm (SD, 1.3), respectively (p = 0.14).
Discussion
The widespread adoption of multidetector CT has resulted in
the routine availability of multiplanar reformatted images.
These reformatted images are now a standard component of
many CT imaging protocols including within the abdomen,
chest, and spine. Although some centers have adopted their
routine use in head trauma, acceptance has not been universal,
and the benefits have not been fully investigated. Given the
significant annual morbidity and mortality associated with
Table 4 Proportion of false
negative (missed) subdural Convexity
hematomas by location
Axial alone 61/84 (72.6 %)
Axial + reformats 28/45 (62.2 %)
Parenthesis indicates percentage of the total missed subdural hematomas in either the axial alone or axial +
reformats group
Emerg Radiol
acute traumatic brain injury (TBI) and intracranial hemorrhage,
p values optimization of trauma brain CT protocols is essential. This
study compared SDH detection rates for brain CTs containing
only axial images to detection rates for scans that included
<0.001 axial, coronal, and sagittal images. We found a greater than
<0.001 50 % decrease in false negatives (missed SDHs) with the ad-
<0.001 dition of reformatted images, a surprising and impressive mag-
<0.001 nitude of improved detection. Further, there was considerably
improved sensitivity and specificity for the detection of SDH
with the reformat-added scans (88.3 and 98.3 %, respectively)
compared with the axial-only scans (75.7 and 94.3 %, respec-
tively). This was the case regardless of reader experience.
No prior studies have directly compared the rate of SDH
detection using brain CT with axial images alone to that with
both axial and reformatted images. The few prior investiga-
tions that have assessed the utility of reformatted brain CT
images in acute head trauma have included only very small
numbers of SDHs limiting generalizable conclusions in this
patient cohort. Wei et al. assessed the addition of coronal
reformatted images in the detection of all types of intracranial
hemorrhage (ICH) after acute TBI and found that 14 % of
ICHs were only identifiable on the coronal images [4].
However, their study only included 32 patients with ICH of
which only 7 were SDHs. Additionally, readers were not
blinded to the purpose of the study, which could have biased
the results. Zacharia et al. evaluated reformatted images in
patients with acute TBI and reported additional findings in
18.2 % of cases with the addition of coronal and sagittal
reformats [14]. However, this investigation did not specifical-
ly report on SDH and instead included a wide variety of dif-
ferent types of posttraumatic pathology including fractures
and all types of ICH. Further, their study evaluated a relatively
small sample size that contained only 14 cases of SDH.
Axial images may be less sensitive for SDH detection giv-
en the common location of this pathologic entity adjacent to
the calvarium, which can result in obscuration due to volume
averaging (Figs. 1 and 2). While the false negative cases in the
axial-only and reformat-added groups exhibited a similar dis-
tribution of missed SDH across different locations, there was a
greater percentage along the convexity with the axial-only
scans (73 versus 62 %). Further, SDH with a long axis orient-
ed perpendicular to the axial plane (e.g., parafalcine or lateral
convexity) will have a larger volume of hemorrhage present
on a coronal or sagittal image than on an axial image, which
may increase its conspicuity and aid in detection. We found a
Parafalcine Tentorial Skull base Posterior fossa
11/84 (13.1 %) 8/84 (9.5 %) 1/84 (1.2 %) 3/84 (3.6 %)
9/45 (20 %) 6/45 (13.3 %) 0/45 2/45 (4.4 %)
Emerg Radiol

Fig. 1 Example of false negative

on axial images. Noncontrast
brain CT in a 39-year-old female
with a left frontal convexity
subdural hematoma (SDH)
secondary to acute trauma that
was missed with axial images but
identified with reformatted
images. a Coronal reformatted
image demonstrates a 2-mm acute
SDH overlying the left cerebral
convexity (white arrows). b Axial
source image. The left convexity
SDH is not visualized

trend toward an increased width of SDHs in false negative
cases from the axial-only group compared with those from
the reformat-added group. This may suggest that smaller
SDHs can be detected with the inclusion of reformatted
images.
The early identification of SDH in the setting of acute head
trauma is important in order to allow for appropriate monitor-
ing or neurosurgical intervention [2]. Several prior studies
have demonstrated that a considerable number of patients with
acute SDHs undergo delayed expansion of hemorrhage requir-
ing surgical intervention remote from the time of injury. Son
et al. prospectively analyzed 177 SDH cases and found that
16.4 % required surgical intervention greater than 7 days after
the causative event [15]. Similarly, Kim et al. found that 35 %
of patients with SDHs after mild head trauma required delayed
surgical evacuation (defined as greater than 1 week) [16]. This
suggests that a substantial number of SDHs will expand in a
subacute or delayed time course. Further, there are extensive
reports in the neurosurgical literature of delayed acute subdur-
al hematoma in which an initial posttraumatic axial brain CT is
negative followed by development of SDH on a subsequent
scan [1720]. Identifying a greater percentage of SDHs on the
initial scan by including coronal and sagittal reformatted im-
ages could allow for an earlier diagnosis, resulting in closer
monitoring and potentially improving patient outcomes.
An improved SDH detection rate with the addition of
reformatted images occurred regardless of experience level.
While baseline SDH detection rates varied with reader ability
and experience level, all readers appreciated improvement with
the addition of reformatted images and the magnitude of that
improvement was the same regardless of experience level.
One potential criticism of the routine inclusion of
reformatted images is the increased interpretation time associ-
ated with a larger number of images. In our study, we found a
21.6 % increase in mean interpretation time with the inclusion
of reformatted images. The potential benefits of improved
SDH detection from the addition of reformatted images will
need to be weighed against this increased interpretation time.
Further prospective study is warranted to investigate differ-
ences in patient outcomes with the inclusion of reformatted
images in trauma brain CT protocols.
There were several limitations with our investigation. First,
this was a retrospective study performed at a single institution
with only three readers. Different readers at a different

Fig. 2 Another example of false

negative on axial images.
Noncontrast brain CT in a 43-
year-old female with a right
parietal convexity subdural
hematoma (SDH) secondary to
motor vehicle collision that was
missed with axial images but
identified with reformatted
images. a Coronal reformatted
image demonstrates a 3-mm acute
SDH overlying the right parietal
lobe (white arrows). b Axial
source image. The right parietal
SDH is not visualized

institution could have produced different results. Second, we
did not investigate other types of pathology beyond SDHs nor
did we evaluate the utility of other types of additional images
on SDH detection rates (e.g., thin section axial images). Third,
there was a difference in the thickness of the axial (5 mm) and
reformatted (2 mm) slices, which could have accounted, in
part, for some of the increased detection rate. However, this
was a study designed to assess SDH detection rates using the
standard axial-only imaging protocol to that of a modified
protocol adding coronal and sagittal images to this standard
axial protocol. It was not a study directly comparing axial
images with coronal or sagittal images; in such a study, equiv-
alent slice thicknesses would have been necessary. Rather, in
our study, we aimed to focus on real-world application, and it
is common that coronal and sagittal reformats are created at a
2-mm slice thickness. Finally, there was a significant differ-
ence in mean age between the normal and SDH cohorts.
However, this is not particularly surprising given the known
increased propensity for SDH with advancing age, in part due
to progressive brain atrophy, and it is unlikely that this differ-
ence would affect SDH detection rates [21].
Conclusions
In conclusion, the addition of coronal and sagittal images to
axial images in trauma brain CT resulted in considerable im-
provement in the sensitivity and specificity for the detection of
SDH as well as a reduction in SDH false negatives by greater
than 50 %. This finding warrants prospective assessment of its
effects on patient outcomes. Neuroradiologists should consid-
er modifying their standard trauma brain CT protocols to in-
clude coronal and sagittal images.
Acknowledgments This study was approved by our institutional re-
view board with a waiver of the need to obtain informed consent and
was compliant with the Health Insurance Portability and Accountability
Act.
Compliance with ethical standards
Conflict of interest Statement The authors declare that they have no
conflict of interest.
References
1. Taylor CA, Greenspan AI, Xu L, M-j K (2015) Comparability of
National Estimates for traumatic brain injury-related medical en-
counters. J Head Trauma Rehabil 30(3):150159. doi:10.1097
/HTR.0000000000000105
Emerg Radiol
2. Frattalone AR, Ling GS (2013) Moderate and severe traumatic
brain injury. Neurosurg Clin N Am 24(3):309319. doi:10.1016/j.
nec.2013.03.006
3. Strub WM, Leach JL, Tomsick T, Vagal A (2007) Overnight pre-
liminary head CT interpretations provided by residents: locations of
misidentified intracranial hemorrhage. Am J Neuroradiol 28(9):
16791682. doi:10.3174/ajnr.A0653
4. Wei SC, Ulmer S, Lev MH, Pomerantz SR, Gonzalez RG,
Henson JW (2010) Value of coronal reformations in the CT
evaluation of acute head trauma. Am J Neuroradiol 31(2):
334339. doi:10.3174/ajnr.A1824
5. ACRASNRSPR Practice Parameter for the Performance of
Computed Tomography (CT) of the Brain (2015).115
6. Bhullar, IS, Block E (2011) CT with coronal reconstruction iden-
tifies previously missed smaller diaphragmatic injuries after blunt
trauma. The American Surgeon
7. Caoili EM, Cohan RH, Korobkin M, Platt JF, Francis IR, Faerber
GJ, Montie JE, Ellis JH (2002) Urinary tract abnormalities: initial
experience with multidetector row CT Urography1. Radiology
222(2):353360. doi:10.1148/radiol.2222010667
8. Cho SH, Sung YM, Kim MS (2012) Missed rib fractures on eval-
uation of initial chest CT for trauma patients: pattern analysis and
diagnostic value of coronal multiplanar reconstruction images with
multidetector row CT. Br J Radiol 85(1018):e845e850.
doi:10.1259/bjr/28575455
9. Jaffe TA, Martin LC, Thomas J, Adamson AR, DeLong DM,
Paulson EK (2006) Small-bowel obstruction: coronal reformations
from isotropic voxels at 16-section multidetector row CT.
Radiology 238(1):135142. doi:10.1148/radiol.2381050489
10. Kung JW, JS W, Shetty SK, Khasgiwala VC, Appleton P, Hochman
MG (2014) Spectrum and detection of musculoskeletal findings on
trauma-related CT torso examinations. Emerg Radiol 21(4):359
365. doi:10.1007/s10140-014-1201-9
11. Marin D, Catalano C, De Filippis G, Di Martino M, Guerrisi A,
Rossi M, Passariello R (2009) Detection of hepatocellular carcino-
ma in patients with cirrhosis: added value of coronal reformations
from isotropic voxels with 64-MDCT. Am J Roentgenol 192(1):
180187. doi:10.2214/AJR.07.3652
12. Paulson EK, Jaffe TA, Thomas, J (2004) MDCT of patients with
acute abdominal pain: a new perspective using coronal reformations
from submillimeter isotropic voxels. American Journal of
Roentgenology
13. Kim ES, Yoon DY, Lee HY, YJ K, Han A, Yoon SJ, Kim HC (2014)
Comparison of emergency cranial CT interpretation between radi-
ology residents and neuroradiologists: transverse versus three-
dimensional images. Diagn Interv Radiol 20(3):277284.
doi:10.5152/dir.2014.13401
14. Zacharia TT, Nguyen DTD (2009) Subtle pathology detection with
multidetector row coronal and sagittal CT reformations in acute
head trauma. Emerg Radiol 17(2):97102. doi:10.1007/s10140-
009-0842-6
15. Son S, Yoo CJ, Lee SG, Kim EY, Park CW, Kim WK (2013)
Natural course of initially non-operated cases of acute subdural
hematoma: the risk factors of hematoma progression. J Kor
Neurosurg Soc 54(3). doi:10.3340/jkns.2013.54.3.211
16. Kim B, Park K, Park D, Lim D, Kwon T, Chung Y, Kang S
(2014) Risk factors of delayed surgical evacuation for ini-
tially nonoperative acute subdural hematomas following mild
head injury. Acta Neurochir 156(8):16051613. doi:10.1007
/s00701-014-2151-4
17. Bordes J, Goutorbe P, Lacroix G, Prunet B, Asencio Y, Kaiser E
(2012) A case of massive delayed acute subdural hematoma. J
Emerg Med 42(4):459460. doi:10.1016/j.jemermed.2010.05.084
18. Cooper PR (1992) Delayed traumatic intracerebral hemorrhage.
Neurosurg Clin N Am 3(3):659665
Emerg Radiol
19. Hadjigeorgiou GF, Anagnostopoulos C, Chamilos C, Petsanas A
(2014) Patients on anticoagulants after a head trauma: is a negative
initial CT scan enough? Report of a case of delayed subdural
Haematoma and review of the literature. J Kor Neurosurg Soc
55(1):51. doi:10.3340/jkns.2014.55.1.51
20. Matsuda W, Sugimoto K, Sato N, Watanabe T, Fujimoto A,
Matsumura A (2008) Delayed onset of posttraumatic acute
subdural hematoma after mild head injury with normal computed
tomography: a case report and brief review. J Trauma 65(2):461
463. doi:10.1097/01.ta.0000202465.13784.2a
21. Gerard C, Busl KM (2013) Treatment of acute subdural hema-
toma. Curr Treat Options Neurol 16(1):275. doi:10.1007
/s11940-013-0275-0