Prognostic Value of CT in Head Injuries Akira YAMAURA, Takao NAKAMURA, Junichi ONO ...

12 Pages·2006·1.02 MB·English

Prognostic Value of CT in Head Injuries
Akira YAMAURA, Takao NAKAMURA, Junichi ONO*,
Naobumi ARIGA**, Takeshi HORIE, Yoshikata SHINOHARA*,
Katsumi ISOBE*, Yoshihiko KOKUBO** and Hiroyasu MAKINO
Department of Neurological Surgery, Chiba
University School of Medicine, Chiba 280
*Department of Neurological Surgery
, Kimitsu Central Hospital
**Department of Neurological Surgery
, Narita Red Cross Hospital
Summary
Five hundred cases of acute head injury were analyzed on the basis of clinical severity, CT
findings and outcome. Parenchymal lesions, which were the most important information
in predicting the outcome, were classified into six categories: Isodensity WITHOUT Mass
Effect, Isodensity WITH Mass Effect, High Density, High and Low Density Complex,
Low Density, and Acute Diffuse Cerebral Swelling. Isodensity WITH Mass Effect was
related to the worst outcome (59% mortality and 24% functional recovery). This finding
was obtained 2 and a half hours following trauma and High and Low Density Complex
eventually appeared in repeated CT.
Epidural hematoma was associated with parenchymal lesions in only six (40%) out of 15
cases and acute subdural hematoma was accompanied by parenchymal lesions in 21 (84%)
out of 25 cases. The remarkable differences in the outcome of those two extracerebral
hematomas were entirely dependent on the associated parenchymal lesions.
In the minor head injury group, parenchymal lesions were not rare (8.5%) in patients
with a loss of consciousness even for a brief period.
To follow the rapid dynamic changes in parenchymal lesions, the importance of
repeated CT was emphasized.
Key words: Computerized tomography, head injury, cerebral contusion, epidural hem
atoma, subdural hematoma
period of 18 months up to August, 1979.
Introduction Surgical policy and adjunctive therapy of head
injuries were uniform.") To decompress in
In predicting the outcome of head injuries, creased intracranial pressure, a large de
the most important information undoubtedly compressive craniectomy was the method of
concerns the parenchymal condition. No other choice. Mannitol was used only in a single dose
diagnostic measure was able to visualize paren while waiting for surgery. Steroid was used
chymal lesions in life (in-vivo clinical-pathologi pre and postoperatively in the usual doses.
cal correlation") until the introduction of The three hospitals were equipped with two
computerized tomography (CT). Analyzed here EMI 1010 and one ACTA 0200FS.
are 500 cases of acute head injury, observed CT was carried out as soon as possible on
by CT, including minor and severe head injury arrival of the patients and most of the initial
groups. Stress is placed on analysis of the CT was carried out within 48 hours of the head
prognostic value of CT findings. injury. To determine the dynamic changes in
parenchymal lesions, CT was deliberately
Materials and Methods repeated in the severe head injury group.
Severity of the head injury was classified
This analysis was conducted on 500 cases of into four grades according to Araki 1,18) and was
head injury treated in the Chiba University also shown by the Glasgow Coma Scale.27) The
Hospital and two affiliated hospitals, for a former is a very practical classification based
on the clinical course and is widely used in is entirely dependent on the injury of the
Japan: (1) `simple type' (no loss of con brain itself. Parenchymal lesions were classified
sciousness, no neurological deficit), (2) `tran into six categories (Table 1).
sient type' (loss of consciousness for less than 1) Isodensity WITHOUT Mass Effect: Den
6 hours, no neurological deficit), (3) `pro sity is normal and there is no evidence of
longed type' (loss of consciousness for more swelling of the brain as far as can be seen from
than 6 hours), and (4) `worsening type' (de the shape and the size of the ventricular system
teriorating). The latter two types are referred or the subarachnoid spaces.
to as severe head injuries in this analysis and 2) Isodensity WITH Mass Effect : Density
the former two types as minor head injuries. of the parenchyma is not different from that
CT findings were recorded in our own pro of the normal brain, but the deformity and
tocol with other clinical information. The CT shift of the ventricular system and the oblitera
was read for each structure of the head: a) tion of cerebrospinal fluid (CSF) spaces are
extracranial, b) cranial, c) epidural, d) sub evidence of local or hemisperic cerebral
dural, e) subarachnoid, f) ventricular, and swelling (Fig. 1).
g) parenchymal. The classification of the 3) High Density: Discrete spherical high
parenchymal lesion is the most crucial point density which may be surrounded by a low
in this study because the gravity of head injury
Table 1 Classification of parenchymal lesions
Fig. 1 Acute subdural hematoma and Isodensity Fig. 2 High Density. Discrete High Density
WITH Mass Effect on the left side. 41 year old lesion is surrounded by a low dense halo at the
male (dead). CT 5 hours and 15 minutes after top (12 hours after trauma) of the figure. 48 year
trauma. old male (good recovery).
with acute head injuries of various severities.
Two hundred and fifty nine were in the category
of the `simple type', 130 of the `transient type',
65 of the `prolonged type' and 46 of the
`worsening type' according to Araki
. The
former two types, the minor head injury group,
and the latter two types, the severe head
injury group, are discussed separately. In the
severe head injury group, 108 cases were
adequate for analysis of parenchymal lesions.
A) Extracranial structure: Subgaleal hema
toma is not only evidence of trauma, but also
points to the site of impact. Epidural hematoma
Fig. 3 High and Low Density Complex. Acute
subdural hematoma and admixture of high and just beneath the subgaleal hematoma and
low density in two locations of the parenchyma parenchymal lesions with acute subdural hema
on the left side. 42 year old male (dead). CT toma on the opposite side of the impact were
14 hours ofter trauma. common findings, suggesting coup and contre
coup injuries.
dense halo with a time elapse from impact B) Cranium: Depressed fractures at the
(Fig. 2). convexity and free fragments in the parenchyma
4) High and Low Density Complex: Small could be observed in ordinary CT films.
and large high density and low density areas Window level and width should be arranged to
are intermingled in a single lesion (Fig. 3). look for a minor bony injury, but CT was often
"High Density" and "High and Low Density
not as helpful as plain skull films. Also at
Complex" are based on what the observer sees tention was paid to the base of the skull,
in CT films and not related to the pathogenesis including the orbit, the paranasal sinuses and
of such phenomena. Such terminology as the petrous bone. Air from the basal skull
`contusion' or 'hemorrhagic contusion' is fracture was quite easily recognized in each
intentionally avoided. The ratio of the size of intracranial compartment even if the amount
high and low density areas is variable and all of air was very small, but the entrance of air
varieties of high and low densities are included or exit of CSF leakage was usually quite dif
in this category. ficult to determine. The cortical injury beneath
5) Low Density : Low density area with no the depressed bone fragments was also difficult
high density components. to identify until the fragment was removed.
6) Acute Diffuse Cerebral Swelling: As C) Epidural abnormality : Epidural hema
defined by Zimmerman et al.,38) the brain is toma was observed in five cases (1.3 %) in the
swollen bilaterally and equally, and the ven minor head injury group and 15 cases (13.5 %)
tricular system and the CSF spaces have in the severe head injury group. The other
collapsed The parenchymal density is within four cases of the latter group possessed both
the normal range in ordinary CT films. of epidural hematoma and acute subdural
The patients were followed-up for periods hematoma. Concomitant parenchymal lesions
of 12 months to 26 months in this series. The were observed in only six out of 15 cases. This
outcome was indicated by the Glasgow Out was reflected in their outcome which was quite
come Scale") and the prognostic value of CT favorable since useful recovery was achieved
in the acute stage of head injury was evaluated. in 11 cases (73 %) and only two cases (13 %)
`Functional Recovery' is defined as good
died.
recovery and moderately disabled in the D) Subdural space: Acute subdural hema
Glasgow Outcome Scale. toma was observed in only three cases (0.8 %)
in the minor head injury group and in 25 cases
Results (23 %) in the severe head injury group. Paren
chymal lesions were associated with 21 (84 %)
The present series is composed of 500 patients out of 25 such cases. Among the parenchymal
chronic subdural hematoma was observed in
repeated CT (Fig. 4). The initial CT showed
bilateral subdural collections of low dense
fluid in the acute stage of the head injury. Such
collections were not homogenous in density
but high and low density components were
mixed as if a clot and CSF were present in
layers. After several weeks, this fluid collection
changed into a typical isodense chronic sub
dural hematoma which required surgical evac
uation. All three patients were in their seventies.
E) Subarachnoid space: Observations of
the cortical sulci, Sylvian fissure, interhemi
spheric fissure and basal cisterns offer very
important information in predicting the out
come. Abnormalities are clots of high density
and deformity or obliteration of subarachnoid
spaces. A subarachnoid clot was rarely (0.5 %)
observed in the minor head injury group but
was seen in as many as 25 cases (23 %) in the
severe head injury group. Their Glasgow Coma
Scale was 5.4 ± 2.5 (3 11) and parenchymal
lesions, Isodensity WITH Mass Effect and
High and Low Density Complex were seen in
all but three cases (12%). The outcome was
poor with functional recovery in only six cases
(24%) and death in 15 cases (60%). Five cases
were associated with ventricular hemorrhage,
resulting in death in four cases and a persistent
vegetative state in one case.
Compression or obliteration of the basal
cistern is a sign of poor outcome. The basal
cistern was observed in two compartments,
one anterior to the brain stem and the other
Fig. 4 Development of isodense chronic subdural
posterior to the brain stem. Obliteration of the
hematoma on the left side (62 days after trauma,
anterior basal cistern is associated with a
the bottom of the figure) from subdural hygroma
mortality of 59%, and the posterior basal
(2 days after trauma, the top of the figure). 74
cistern is associated with a mortality of 72%.
year old male (good recovery).
As long as the basal cistern was clearly visual
lesions, Isodensity WITH Mass Effect was the ized in CT, their chance of functinal recovery
most common finding, observed in 15 cases, was more than 70%. When the Sylvian fissure
followed by High Density or High and Low was obliterated bilaterally in adults, 54% died
Density Complex in six cases. Acute subdural and 33% achieved functional recovery. When
hematoma in the severe head injury group the Sylvian fissure was opened bilaterally, 15
had a poor outcome; only 11 cases (44 %) died and 73 % achieved functional recovery.
achieved functional recovery and 13 cases F) Ventricular system: CT shows the in
(52%) died. Mortality of acute subdural fluence of mass from deformities, obliteration,
hematoma has been rather high since before shifts of the ventricular system and the presence
CT appeared.',' 3,20) The authors have reported of blood in various compartments of the
that the reason for such a high mortality is ventricular system.
the presence of the parenchymal lesions.") Ventricular hemorrhages were seen in eight
On three occasions, the development of cases (7.4%) in the severe head injury group
and they were all observed within 48 hours of period of unconsciousness. None of them had
the head injury. Seven cases were associated neurological deficits initially. However, 15
with parenchymal lesions of High Density and (3.9%) of them possessed unexpected paren
High and Low Density Complex. Their Glasgow chymal lesions (Table 2, Fig. 5). Once a patient
Coma Scale was 6.2±2.2 (3' 9) in the presence loses consciousness by a trauma, even if it
of ventricular hemorrhages and the outcome is a very brief episode, the chance of paren
was death in seven and a persistent vegetative chymal lesions is 8.5%, which can not be
state in one case. neglected in practice. Parenchymal lesions were
G) Parenchyma High Density or High and Low Density Com
G Minor Head Injury Group: Included plex, and on no occasion was there Isodensity
here are 389 patients who did not lose con WITH Mass Effect which was the most dismal
sciousness or became conscious after a brief sign for the severe head injury group.
20 Severe Head Injury Group: Dynamic
Table 2 changes of parenchymal lesions. The brain
shows very dramatic and dynamic changes in
CT observations (Fig. 6). Immediately after
impact, when there is no chance of CT from a
practical point of view, the density of the
brain should be within the normal range. Then
the brain shows `swelling', which may be focal,
hemispheric, or diffuse, but the density remains
normal for various periods. In this swollen
brain, High Density or High and Low Density
Complex appears in several hours or days,
which turns into Low Density in approximately
3 weeks. This observation is a prototype of
brain injury. Some cases may be lost at some
point in this natural course after rapid deteriora
tion and also others may be observed only in
the latter part of the course.
<Isodensity without Mass Effect> Although
the unconsciousness continued for more than
6 hours or even deepened, the brain did not
differ from the normal state in 46 (43 %) out of
108 cases. Extracerebral hematomas were
I: Simple type, II: Transient type. present in 28 %. There were nine epidural
Fig. 5 Unexpected parenchymal lesion (High and Low Density Complex) in the
right Sylvian area following a minor head injury. 69 year old female (good
recovery). CT 1 hour after trauma.
Repeated CT was carried out in 12 cases and
High and Low Density Complex appeared in
seven cases. High Density was not observed
in any case. Only one of them returned to
Isodensity WITHOUT Mass Effect. The age
was distributed between 3 and 79 years (mean :
34.2 ± 20.3 years).
<High Density and High and Low Density
Complex> High Density and High and Low
Density Complex were quite similar phenomena
in that the incidence was 11 % and 18 %, the
mortality 42 % and 37 %, the chance of func
Fig. 6 Dynamic change of parenchymal lesion tional recovery 50 % and 53 %, the Glasgow
A prototype of brain injury. Coma Scale 7.6 ± 3.5 and 8.5 ± 3.9, and the
mean age 45 years and 43 years, respec
Table 3 Isodensity with mass effect tively. They differed only in the time interval
after head injury. High Density was observed
within 6 hours of impact and High and Low
Density Complex between 6 hours and 24
hours after head injury. High Density was sur
rounded by a low density halo in the following
course, but this was not read as High and Low
Density Complex.
Complicated classification of parenchymal
lesions with high dense components was
avoided here, and our classification into two
unequivocal categories is quite simple to
facilitate wide practical use and minimize
observer variabilities. Recording the lesion as
`single' or `multiple' was also an unequivocal
matter. There were remarkable differences in
hematomas and four acute subdural hema the outcome between single and multiple
tomas. After an initial CT of Isodensity lesions. Multiple lesions, no matter where the
WITHOUT Mass Effect, repeated CT was lesions were located in the unilateral hemi
carried out in 18 cases. Parenchymal lesions of sphere or in the bilateral hemisphere, showed
High Density or High and Low Density a much higher mortality (Table 4).
Complex appeared in three of the 10 non In two children, there was a single lesion of
operative cases and in four of the eight opera
tive cases. Their prognosis was favorable; Table 4 High density and high-low density
functional recovery was achieved in 37 cases
(80 %) and only seven patients (15 %) died.
<Isodensity with Mass Effect> This finding
was observed in the initial CT of 17 cases (16 %)
and their outcome was very dismal with a
mortality of 59%. Only four achieved func
tional recovery and the other three remained in
a persistent vegetative state (Table 3). The
initial CT was obtained within 6 hours in 88
with an average time of 2 hours and 37 minutes
(mean) and a minimum of 30 minutes after
trauma. Their Glasgow Coma Scale was 5.2+
2.0. Operations were required in 14 cases.
High Density and High and Low Dentsiy of such patients was clearly divided into two
Complex at the basal ganglion in each case. categories, favorable (good recovery in six
These children showed hemiparesis which and moderately disabled in one) and quite
gradually improved. This isolated lesion in the dismal (death in three). All deaths were within
basal ganglia seems to have a unique clinical 3 hours to 15 hours of head injury (Fig. 8).
picture (Fig. 7). Angiography failed to show At the present time, there are no recognizable
any abnormality. differences in CT findings between the above
<Low Density> In acute stages of head two groups. In the benign group, the ventricular
injury, Low Density lesion with no high dense system returned to the normal size in 5 to 15
components was rare and seen in only four days (mean : 9 days) in repeated CT.
cases. All parenchymal lesions, associated with The outcome in each parenchymal lesion is
considerable destruction of brain tissue, are summarized in Table 5.
destined to end in Low Density. Such an
end stage should not be confused with the Discussion
phenomena in the acute stage. Low Density
in the posterior cerebral artery distribution, as The role of CT in the study of acute head in
a consequence of infarction due to trans juries has been highly evaluated. There have
tentorial herniation, is a common finding in been many reports on CT observation of head
the severe head injury group, but is not included injury, but papers on the prognositic value of
in this category. CT are limited. Our attention was focused on
<Acute Diffuse Cerebral Swelling> Acute the relation between the parenchymal lesions
Diffuse Cerebral Swelling was observed in the and the prognosis. CT also disclosed un
initial CT of 10 patients (9 %). Six of them were expected lesions in the minor head injury
findings within 6 hours of impact. The outcome group which could not be detected in vivo
before CT.
To obtain as much information as possible
from any single CT film, a protocol was
Fig. 7 A single isolated parenchymal lesion
(High and Low Density Complex) at the left Fig. 8 Acute Diffuse Cerebral Swelling. 20 year
basal ganglion. 6 year old male (moderately old female (dead). CT 1 hour and 30 minutes
disabled). CT 7 days after trauma. after trauma.
Table 5 Incidence of parenchymal lesions and prognosis (initial CTS
prepared to record all abnormalities in each the depressed fracture or the massive epidural
structure of the head. The recording sequence or subdural hematoma, both of which are
was as follows: extracranial soft tissue, cranium, significantly high dense, might not appear in
epidural and subdural spaces, subarachnoid CT. Becker et al.2) stressed the importance of
spaces, ventricular system and parenchyma. direct inspection of the subdural space and
<Extracranial structure> CT is very sen over the cortex through a small dural opening
sitive to abnormal densities of the extracranial during surgery even if an epidural hematoma
soft tissue. The most common abnormality is definitely confirmed in CT. Isodense
at this site is a subgaleal hematoma which is epidural hematomas have been reported.21)
definite evidence of impact and which reveals Loss of red blood cells into the other injury
much about the possible mechanism of brain site is the cause of this unusual density of
injury. When the parenchymal lesion was hematoma.
unilateral, differentiation of a coup or contre <Subdural space> Subdural abnormalities
coup injury was apparent in a single CT film. are controversial even in CT. It is difficult to
<Cranium> CT disclosed the accurate loca know whether subcranial low density, par
tion of free bone fragments in the parenchyma ticularly in children, is a subdural space or a
or a significant injury of the brain beneath a subarachnoid space in a plain CT. Metrizamide
depressed fracture, but the use of CT for cisternography is indicated in such cases.
diagnosis of linear fractures of the convexity Subdural collection in children shows a benign
or basal skull fractures was of limited value. natural history and a spontaneous regression
<Epidural abnormality> The true incidence can be expected in most of the cases. 12)
of epidural hematoma was not known until Acute subdural hematoma was associated
the introduction of CT. In minor head injuries with significant parenchymal lesions. The most
with no neurological deficits, angiography common lesion is Isodensity WITH Mass
was rarely carried out and asymptomatic Effect with a high mortality. Diagnosis of acute
epidural hematoma was not noticed. CT in subdural hematoma is not difficult with CT
minor head injuries was reviewed for this because of its characteristic shape and wide
purpose, and epidural hematomas were found extension, but underlying brain injury is often
in only 1.3 % of the 389 cases. This incidence missed in CT. A normal looking brain with a
is fairly low compared with 4.6 % for sub hemispheric swelling is a sign of significantly
dural abnormalities. poor prognosis. The outcome is entirely
In the severe head injury group, the incidence dependent on this brain injury, not on the
of epidural hematoma was 13.5 % and paren presence of a subdural hematoma.33)
chymal lesions were present in only six out of Chronic subdural hematomas developed in
15 cases. This is the reason for favorable repeated CT based on the subdural hygroma in
recovery from epidural hematoma in 73 older patients. This observation was also
of the cases. It is important to know that confirmed by Yamada et al.32)
contusions of the cerebral cortex just beneath <Subarachnoid spaces> Blood clots in the
various subarachnoid spaces were rarely ob associated with minor head injury, using clinical
served in the minor head injury group but and experimental materials. In our series,
were seen it 23 % of the severe head injury 8.5 % of the patients with transient loss of
group. These findings were associated with a consciousness had parenchymal lesions of
very grave clinical state, concomitant with High Density or High and Low Density
parenchymal lesions and poor outcome. Complex. There was no case where Isodensity
Compression and obliteration of the basal WITH Mass Effect was demonstrated. Waga
cistern is known as one of the signs of descend et al.") reported a case of a 20 year old man who
ing transtentorial herniation.24) In our observa had not lost consciousness where CT revealed
tions, patency of the posterior perimesencephalic marked swelling of a unilateral cerebral
cistern was more important for a better outcome. hemisphere beneath a thin acute subdural
<Ventricular system> CT was the first hematoma. This lesion was exactly the same
noninvasive method which allowed observa as Isodensity WITH Mass Effect in our defini
tion of ventricular hemorrhages in vivo. Ven tion. Such parenchymal lesions in the minor
tricular hemorrhage was associated with head injury group should be carefully observed
parenchymal lesions and with a significantly for the possibility of foci of epilepsy. CT, at
poor prognosis in this series. In the literature, least at the present time, is unable to visualize
ventricular hemorrhage was seen in 3 to 9 a contusion which is limited to the cortex or
of head injuries',",' 6,22,29) and Tsai et al.") a small lesion in the brain stem. It should be
listed the corpus callosum, fornix, choroid stressed that CT is not omnipotent.
plexus and remote intracerebral hematoma as <Severe head injury group> Parenchymal
sources of ventricular hemorrhage. This was lesions in the severe head injury group are
also observed without concomitant paren never static phenomena, but show slow or
chymal lesions in three out of six patients by rapid dynamic changes as summarized in Fig. 6.
Oliff et al.,16) one out of three patients by <Isodensity without Mass Effect> In spite
Merino-de Villasante et al.,") and six of 17 of clinically severe head injuries, there were no
patients by Tsai et al .29) French et al.7) and definite parenchymal lesions in 43 % of the
Roberson et al.22) found parenchymal lesions severe head injury group, but seven out of 18
in high dense components in all of their cases. cases where there was opportunity to repeat
The outcome in the presence of ventricular CT showed the new appearance of High Den
hemorrhage was not so poor and development sity or High and Low Density Complex.
of hydrocephalus was not common in their Isodensity WITHOUT Mass Effect was fol
experience.' °22) lowed by a favorable outcome.
<Parenchyma> There had been no op <Isodensity with Mass Effect> Only one
portunity for neurosurgeons to observe paren fourth of such cases returned to a useful life
chymal injury in vivo until the introduction of in the presence of this parenchymal lesion.
CT. Inspection of the brain surface during This most important parenchymal lesion was
surgery or brain cutting was the only chance not well defined before. In initial CT in the
of studing the parenchyma. It is our opinion acute stage (about 2 and a half hours after
that CT should be repeated with adequate trauma), the density of the parenchyma is
timing to follow the dynamic changes in the within the normal range in print-out data.34)
parenchyma. Parenchymal findings were clas The swelling is remarkable, usually com
sified into six categories and all possible errors pressing or obliterating the ventricular system
and observer variablities were minimized in on one side. There are two possible mechanisms
this classification. which cause of this phenomena. One is that
<Minor head injury group> Minor head small low density areas (edema?) and high
injury, defined as the `simple type' without loss density areas (petechiae?) are equally mixed
of consciousness or the `transient type' with and the size of each area is smaller than the
brief unconsciousness, might cause paren size of CT resolution. If the areas of abnormal
chymal damage. Parenchymal changes in such densities were larger, this could be read as a
`salt and pepper' appearance
minor head injuries were also not known until .6) The second
CT. Sekino et al.23) mentioned brain injury possible mechanism is a marked increase of
the vascular bed, which could be an immediate artery due to transtentorial herniation, were
or very rapid response to trauma-" 35 ) The not considered under this title.
latter mechanism is considered in Acute Diffuse <Acute Diffuse Cerebral Swelling> There is
Cerebral Swelling, which, however, was re no question in the diagnosis of this phenomena
ported to have a rather favorable outcome in from its characteristic findings. Zimmerman
the pediatric age group. In our observations, et al.") found a favorable outcome in such
approximately 60% showed remarkable high patients, but in our series, few cases of the
dense components in later CT and in only one fulminating type were included. At the present
case, the swelling subsided without leaving any time, it is difficult to differentiate these two
organic lesion. This observation probably quite different conditions by CT. In Roberson's
supports the former mechanism of admixture series,") there were 10 cases (I1 %) of this
of small lesions below the level of resolution phenomena in 95 cases and eight of them fol
in CT. lowed a benign course, but two were dead or
<High Density, High and Low Density severely disabled. Pathogenesis of this lesion
Complex> Zimmerman et al.40) classified was considered as a remarkable increase of
high dense parenchymal lesions into hemor cerebral blood volume. Obrist et al.15 ) confirmed
rhagic contusions, intracerebral hematomas an increase in cerebral blood flow by intrave
and diffuse white matter shearing injuries.") nous injection of 133Xe.
Tsubokawa et al.") found three types; primary <Importance of repeated CT> For a full
intracerebral hematoma, hematoma within understanding of brain injury, CT should be
contusion and contusional hematoma. The carried out at the earliest possible time after
latter classification is based on a possible trauma and should be repeated according to
causative mechanism of high dense lesions. clinical conditions. New parenchymal lesions
The pathogenesis is not clearly understood at appeared in 16 cases (15 %) of the severe head
present. Tsai et al.") stated that differentiation injury group. Four of them were in nonopera
of contusion, hematoma or edema is not `clear tive cases and the other 12 cases were post
cut' even in CT observations. We have operative. The most common new findings
avoided a complicated categorization based were High and Low Density Complex, fol
on unknown pathogenesis. The differentiation lowed by High Density. Such postoperative
of single or multiple lesions was quite useful High Density appeared between 3 and 6 hours
concerning predictive values. Multiplicity of after trauma. This timing is agrees with that of
lesions has not been stressed with respect to initial CT observations of High Density.
the prognostic value. Sweet et al.26) found a Pathogenesis of High Density is probably not
high mortality (70%) in patients with bilateral related to the surgical procedure.
high dense lesions. Extracerebral hematomas The usefulness of repeated CT has been
were included in the high dense abnormalities reported by many authors .3,4,7,17,19,28) French
in their series. et al.') found new lesions resulting from
In 1978, Zimmerman et al.39) pointed out further progress of the previous lesion in 52
the presence of small high dense areas, scattered of their patients and Cooper et al. found them
in the swollen brain in shearing injury in 65 %.
(Strich25)). The hemorrhage was characteris CT has brought us a considerable amount of
tically located at the corpus callosum, cor information about intracranial lesions but CT
ticomedullary junction, and superior cerebellar is not `omnipotent'. It is difficult to visualize
peduncle. Shearing injury is included in multiple small lesions at the cerebral cortex and the
High Density areas with a very high mortality brain stem due to the partial volume phenomena
in our series and the limits of spatial resolution of the
<Low Density) Low Density lesions with present CT. The size of the ventricular system
out any high dense components were rare in might be seen as smaller than the true size.
the acute stage. Low dense areas, which were Ise et al.') reported the importance of delayed
seen as a halo around the High Density, as a blood flow observed in ang ography in relation
part of the High and Low Density Complex, to poor outcome. This information is not
or as an infarction of the posterior cerebral obtained by regular CT. French et al. listed