Emergency Craniotomy

NeuroSurgery InfoNet

 

 

Imaging

 

Introduction

Imaging

Laboratory

           

 

Introduction


Imaging            

Following resuscitation, imaging of the head trauma victim’s head and neck should be done as quickly as possible. Intracranial imaging is essential not only for identification of surgical lesions but also for choosing, planning, and setting up the optimal craniotomy. 

Imaging provides information complementary to that obtained from clinical history and physical examination.  Without imaging, the physical examination and Glasgow Coma Scale (GCS) score may have to be the sole basis for surgical decision making (with or without placement of exploratory burr holes). 

Computerized tomography of the head (“head CT”) is the current standard for intracranial imaging following head trauma, although alternatives (such as MR or angiography), should be available in the event that CT scanning is not.

Computerized tomorgraphy ("CT")

Once the surgeon understands the principles underlying the acquisition, processing, display, and reproduction of head CT scans, he can learn how to read them.  CT scan reading requires not only detection of pathology but also interpretation and diagnosis of abnormalities based on knowledge of the normal anatomic disposition of intracranial structures as well as imaging characteristics of injured brain tissue and traumatic hematomas.

Principles

The principles of intracranial CT scanning important to reading and interpretation of these studies in head trauma victims are technical and related to the image characteristics of injured intracranial contents.

(i)                 Technical aspects

Technical aspects of head CT scanning include those of image acquisition, processing, display, and reproduction.  The technical features that must be understood to accurately interpret a CT scan are: 1.  Greyscale, 2. Windows, 3.Contrast enhancement, 4. Axis of acquisition, 5. Interval between slices, 5. Three dimensional reconstruction, and 6. Artifact. 

Sections

“Tomo”- comes from the Latin for “to cut” (craniotome, osteotome, …).  Xray information on a single plane across an object is translated by computer calculation into an image of the transected object along the plane.  Every object that lies in the plane of x-ray beams at that level appears on this tomogram.

One of the limitations of CT scans is that they are images in only the axial plane.  The resolution of so-called reformats-coronal or sagittal images created from serial axial image data-is only as good as the distance between axial cuts.  The computer can reconstruct Sagittal and coronal views but they are far less accurate and far less clear than the original axial images.  Even at 5 mm between cuts, the resulting reformatted image resolution is less than that of the axial images from which it was derived, which attain 1 to 2 mm resolution.  The axial cuts of the CT scan are taken 20 degrees off the orbitomeatal line to avoid catching too many air-filled sinuses in one cut.  Also, structures or lesions running parallel to but at a distance from the axis of the x-rays may be missed by a CT image.  In addition, linear fractures without displacement of bone may be thinner than the resolution of a given scanner.

(b)               Greyscale

The display of a CT scan is simple: two dimensional, black and white.  CT scan looks inside intracranial tissues with a display “greyscale” (degrees of black and white) based not on optical properties of a tissue surface but rather on the density characteristics of the tissue volume.

Plain x-rays and CT scans are images resulting from differential penetration of x-rays through tissues of different densities.   As on a plain x-ray, denser objects on CT are lighter, and less dense objects darker.  On a cranial CT, bone is white, air is black, and brain tissue is intermediate shades of gray. 

The Hounsfield unit (named after one of the inventors of CT) quantitates the gradations from black to white on a density gray scale. Individual tissue types, such as gray matter, white matter, and fat, each have their characteristic Hounsfield value, which can be used to determine density (and, therefore, type) of tissue.

Fat is very low density, even less than CSF: on CT it is closest in density of all tissues to air.  Within the parenchyma, the darker areas are less dense, thus the “gray” matter of gross specimens appear lighter on CT than the corresponding gross “white” matter.

FIGURE:  Hounsfield units

(c)               Windows

Windows are portions of the grayscale selected for attention and analysis.  Windows vary contrast characteristics within the image to highlight or obscure tissues within certain selected density ranges.

The brightness and contrast of the image can be adjusted to highlight tissue of any selected density while de-emphasizing surrounding tissue of other densities.  There are three brightness-contrast tissue settings, or “windows,” which are programmed into the computer for most head scans:

The “bone window” focuses on tissue of the uppermost limits of the Hounsfield scale-the densest structures- with virtual dropout of all others.  This setting enables visualization of most fractures of cranial vault and base.

Images focused on low-to-intermediate Hounsfield numbers highlight the soft tissues of the brain parenchyma as well as extracranial soft tissue (muscle, skin, and subgaleal space) and make up the “soft tissue window”.

A separate “subdural window” focuses on recently coagulated blood that might otherwise be difficult to distinguish from underlying brain tissue, using the wider “soft tissue” settings.

FIGURE: Bone, soft tissue, and subdural windows

(d)               Contrast

Contrast is used to demonstrate intracranial structures that have an absent or deficient blood brain barrier.  Head trauma does not diffusely disrupt the blood brain barrier and is therefore not an indication for the use of contrast enhancement.

FIGURE: With and without contract

(e)               Axis

The axis of the plane of two dimensional cross sections on standard CT scans throughout the world is 20 degrees elevated with respect to the inion-glabella line.

The axis of cross section in a head trauma CT scan is usually set to 20 degrees ____. Because the plane of the axial cuts is 20 degrees off parallel to the plane of the line from familiar external landmarks, such as the glabella and inion, the structures at each level are not those of axial cuts orthogonal to the rostrocaudal axis of the head.  With each successively higher view, the structures visualized are more posterior than those that would have been seen on slice, parallel to the glabella-inion line.

FIGURE: Angulation of axial CT scans (CT axis angle.tif)

The uppermost cut on the CT scan does not correspond to the uppermost part of the patient’s head (i.e., the vertex).   Brain structures that appear far anterior on successive CT scan slices are actually more posterior as the image series progresses from bottom to top.  The scout view is useful in remembering the axis orientation with respect to the cranium and extra-and intracranial structures.

(f)                Interval

The usual distance of 5 mm between respective axial sections provides a resolution that is usually greater than that required evaluating an extra-axial hematoma.  When time is limited, greater slice width, for example 10 mm or greater gives fewer scans more quickly.

(g)               3D

The computer can generate reconstructed images of the sagittal and coronal planes but these take a long time and provide no information that the surgeon does not already have from the axial images. 

(h)               Artifact

Technical artifacts can result in display image abnormalities that can erroneously be interpreted as pathologic and requiring intervention.

Image artifacts are distortions of the CT image due to the technique of processing the image rather than any factor intrinsic to the patient’s anatomy.  Computer-processing artifact increases directly and dramatically at areas where plain x-ray beams incompletely penetrate tissues such as thick bone (e.g., that anterior in the posterior fossa).  For this reason, CT is not a good technique for visualizing the brainstem, fourth ventricle, and cerebellum.

Artifacts are spots, lines, shapes, densities generated by the computer and displayed on the CT image that have no physical reality.  Computer-processing artifact increases directly and dramatically at areas where plain x-ray beams incompletely penetrate tissues such as thick bone (e.g., that anterior in the posterior fossa).  For this reason, CT is not a good technique for visualizing the brainstem, fourth ventricle, and cerebellum.

FIGURE:  Posterior fossa artifact

(ii)                Tissue imaging characteristics

Brain, CSF, and blood are the tissue types whose imaging characteristics must be understood in order to accurately make pathologic and monitor lesion evolution.

(a)               Brain tissue

The low cellularity and high fat (myelin) content of the “white matter” as seen in gross sections corresponds to darker areas on CT.   Heavily cellular “gray” matter appears lighter on CT than the “white” matter. 

Gray and white matter can be differentiated on CT scans because gray matter is more cellular  (and thus denser) than myelin-rich white matter.  Normally there is a distinct border between the gray and white matter on the CT soft tissue windows (latter appears lighter than former).  Loss of the gray-white boundary is suggestive of diffuse cerebral edema.  Cerebral edema is associated with an increase in the water content of the gray matter, which reduces its density (and Hounsfield value), making it darker (closer in appearance to white matter).

Increased intercellular water in the grey matter, as with post-traumatic edema decreases the overall density of the tissue towards that of the adjacent white matter.   As their densities become closer, the contrast between grey and white matter on CT decreases – an important finding suggestive of cerebral swelling.

FIGURE: Grey-white demarcation

Evaluation of the brain surface includes looking for subarachnoid blood, which is frequently present in quantities that are insufficient to layer out enough for detection of a CT scan.

(b)               CSF

Cerebrospinal fluid (CSF), like water, is less dense than brain parenchyma and bone, so it appears darker on CT than these.  CSF is denser, and therefore slightly less black, than air.  When blood mixes with CSF it changes its imaging characteristics making it brighter (increased density) on CT.  Blood cells and heavier debris are bright and layer out (“hematocrit” effect) in dependent (in the patient supine during CT scanning) CSF-filled spaces such as the posterior (occipital) horns of the lateral ventricles.

Replacement of CSF density cisterns by brain parenchymal density tissue (“obliteration of the cistern”) indicates that brain tissue, either globally or focally in the region adjacent to that cistern, is swollen.  If the cistern is next to or surrounds the brainstem, its obliteration is a sign that brain tissue from another compartment (i.e., herniated) is pressing on the brainstem at that site.

(c)               Hematoma

Blood density depends on the concentration of cellular elements and the balance between clot formation and lysis.  Clotted blood has numerous fibrin and other protein coagulation substances that form a tight meshwork to stop the flow of blood.  Blood cells are more densely packed in this meshwork than freely floating in a liquid.  Blood can appear very dense when it first clots, but over time the protein meshwork breaks down, is dissociated by the activity of fibrinolytic enzymes, and the blood gradually returns to a more liquefied state. 

Concomitant with the dissolution of the protein coagulant meshwork is lysis, or breakup, of the cellular components of the blood that have died.  As these cellular components break down and are resorbed, the fluid that remains in less dense than fresh blood. 

The consistency of the hematoma correlates to density and is therefore well visualized on CT.  Low density hematomas are fluid in consistency.  High density = solid.  Thus, blood at different stages in different conditions can appear more dense than brain parenchyma, although always less dense than bone.  It can appear of approximately the same density as the parenchyma and thus be difficult to distinguish from it.  After a long period of time (possibly 2 weeks or longer), blood can appear much less dense, sometimes even comparable in density to that of CSF.

However, blood, when mixed with less dense CSF, tends to sink down, leaving the CSF above it.  When the patient’s head is laid flat in the gurney, the blood in the ventricles can be seen to layer out in a straight line, forming a meniscus – the so-called “hematocrit effect.”

Layering out of blood in a low- density fluid collection is indicative of a subdural hematoma.  This finding is important in planning treatment because it indicates that the likelihood of complete clot evacuation through one or two burr holes over the lesion will be sufficient for its complete evacuation and that a full craniotomy will be required.

FIGURE: Hematocrit effect in hematoma (hematocrit effect.tif)

Blood vessels can be, but are not reliably, visualized by head CT.  Although best visualized following administration of intravenous (i.e., intravascular) contrast, major vascular structures can be visualized by their density and round shape.

Active bleeding into the subarachnoid or epidural space may be detected on CT scan because of the different density of blood collections of differing ages.  Fresh bleeding under pressure has a lower-density appearance than that of coagulated blood into which it forces itself, thus, an actively bleeding epidural hematoma may appear as a high-density “biconvex lens” with low-density streaming bands.  Active or recent acute bleeding from the bridging veins into a preexisting low-density chronic subdural hematoma (the so-called subacute subdural hematoma) appears as a lower-density area juxtaposed onto a higher-density layer.

FIGURE: Active bleeding into a hematoma (active bleeding.tif)

(b)               Reading

The reading of a CT scan involves four distinct stages:  1. Organization, 2. Orientation, 3. Diagnosis, 4. Decision making. 

(i)                 Organization

Organized reading of a CT scan begins with arranging the films, verifying the identity of the patient, looking at the scout and, finally, at the axial views.

Head CT reading is an assessment of normal and normal anatomy at each layer on the three window views in the order of their display: bone, soft tissue, and subdural hematoma.  Frequently, a striking finding is accompanied by other findings that although subtle, are important for management of the patient.  For example, if detected preoperatively, the skull fracture above an epidural hematoma can be deliberately incorporated into the craniotomy bone flap.  The ipsilateral temporal bone hairline fracture probably explains the isolated facial weakness better than the 1-cm contralateral subdural hematoma.

(a)               Films

If a separate film has been made for each of multiple windows, the films should be lined up for viewing -- left to right, top to bottom.   First (from left) in the display sequence are bone windows, followed by soft tissue, and subdural windows.

CT reading is organized to evaluate each layer on the three window views in the order of their display: bone, soft tissue, and subdural hematoma. 

(b)               Demographic

The demographic panel with the patient’s name, sex, and date of birth should be compared with the information obtained from both the patient’s chart and an attached name band.

(c)               Scout

The scout view is a useful reference for the axis of orientation of the cuts with respect to extra-and intracranial structures.  Because the plane of the axial cuts is 20 degrees off parallel to the plane of the line from familiar external landmarks, such as the glabella and inion, the structures at each level are not those of axial cuts orthogonal to the rostrocaudal axis of the head

With each successively higher view, the structures visualized are more posterior than those that would have been seen on slice, parallel to the glabella-inion line.

FIGURE: Scout views CT.

The scout image is analyzed before the axial sections. Parallel lines from the base of the skull to the vertex are displayed and numbered on the scout view.  Finding the corresponding number on any given slice makes it possible to determine the level of that slice with respect to external head landmarks. External landmarks visible grossly and on CT provide the surgeon with a stereotactic reference for surgical target  localization.

The uppermost cut on the CT scan does not correspond to the uppermost part of the patient’s head (i.e., the vertex).  Instead, it corresponds to a few centimeters posterior.  Brain structures that appear far anterior on successive CT scan slices are actually more posterior as the image series progresses from bottom to top.  The scout view is useful in remembering the axis orientation with respect to the cranium.

(d)               Axial sections 

Read the sequence of axial cuts from the base of skull to vertex (i.e., upper to lower; left to right).  Read each axial cut starting from the outside and moving inward, identify asymmetries of structure and tissue density, and inspect the midline for left or right shifts.

FIGURE: CT reading sequence

CT axial cuts should be read left to right, top to bottom.  Each slice is read starting from the outside and moving inward, identify asymmetries of structure and tissue density.  Structures in the midline are identified and assessed for left or right shifts.

(ii)                Orientation

Orientation to an individual scan requires identification of certain basic anatomic landmarks (usually bone) present on all, and relies heavily on the symmetry of structures to the right and left of midline.

(a)               Geometry

The midline is the most important geometric feature of head CT scans.  It results from the right-left bilaterally of many of the intracranial structures. Midline shift refers to displacement of normally midline structures (e.g., the pineal body, the falx, the third ventricle, the fourth ventricle, a large prepontine vertebral artery) to the right or left.

Even to those without knowledge of intracranial anatomy the arrangement of the middle cuts of a CT scan is reminiscent of a face: The anterior horns of the lateral ventricles are the eyes, the midline third ventricle the nose, and the quadrigeminal cistern, the smiling mouth: the face results from the symmetry of its constituent structures. 

Symmetry of paired structures such as the globes, lateral ventricles (frontal, occipital, and temporal horns), choroid plexus (often calcified), cerebral peduncles, colliculi, Sylvian fissures, and basal cisterns is the norm against which purported asymmetry of any of these structures should be evaluated.

“Mass effect” refers to alterations in symmetry due to unilateral compression of solid structures, effacement is a decrease in the volume of those that are fluid-filled. 

The pineal body is the most easily recognized midline structure (that can be displaced by pathologic swelling of mass effect).  The pineal is not only in the midline but is usually calcified (thus visible without contrast).  The falx or portions of it can be identified if calcified or overlain by a (subdural) hematoma.

TABLE: CT geometry

FINDING

GEOMETRIC EFFECT

Mass effect

Structure shape on CT altered

Shift

Structure location changed with respect to x-y axis

Effacement

Decrease in size

Compression

Structure shape on CT altered

 

FIGURE: Degrees of shift and mass effect

SHIFT GRADE

Structures, distance of shift, changes in structure shape

1

One structure

2

 

3

 

(b)               Anatomy

Anatomic analysis of axial views requires familiarity with the appearance of the major intra-and extracranial structures on CT.

(i)                 Extracranial

Anatomic analysis of CT scan layers begins with the extracranial soft tissues. Although the CT resolution is not fine enough to show the individual layers of the scalp, certain extracranial pathology will show up. 

Disruption of the normally smooth contour of extracranial soft tissue layers is an important CT finding.   Although the CT resolution is not fine enough to show the individual layers of the scalp, certain extracranial pathology will show up.  Frequently, based on this evaluation when the intracranial compartment is normal, the diagnosis of head trauma is conclusively made. 

Irregularities of the scalp include high-density swellings consistent with acute blood, so-called cephalohematoma, between the galea and periosteum.  Radio-opaque foreign bodies, such as bullet and bone fragments, can be seen in cephalohematomas. The next layer in is low-density subcutaneous tissue (fat).  The ears stick out from the scalp in the midlevel cuts.

Radio-opaque foreign bodies, such as bullet and bone fragments, can be seen in cephalohematomas. In comatose patients “found down” (coma etiology unknown) with intracranial CT negative for pathology, cephalohematoma on extracranial CT is the frequently the basis for the diagnosis of “head trauma”.

The next layer in from the skin surface is low-density subcutaneous tissue (fat). 

The ears stick out from the scalp in the midlevel cuts.  The external auditory canal and pinnae are readily visualized and are useful reference landmarks for vertical orientation.

(ii)                Skull

The next extracranial to intracranial layer is the calvarium.  The internal and external contours of the skull are smooth.  Discontinuities and step-offs involving dense cortical bone are readily identified.

At the skull base, look for any lines out of place or the distortion of the normally circular or ovoid temporal and petrous bony foramina.

SKULL BASE

The status of the bony structures of the skull base can be assessed from careful review of the CT bone windows.  The petrous bone and its many foramina (holes), which are the sites for cranial entry and exit of major blood vessels and cranial nerves, can be seen on CT bone windows (FIGURE: 2-10).  The foramina rotundum (passageway for second division of CN5). The ovale (the third division of CN5), and the spinosum (which admits the middle meningeal artery) are all distinguished.

The internal auditory canal is seen if the axial cut intersects it.  Other easily identified skull base structures include the orbits (which are filled with fat) and the aerated bony sinuses, including the frontal, maxillary, ethmoid, and sphenoid sinuses as well as the mastoid air cells.  Bleeding into the air-filled sinuses can appear as a meniscus of fluid within the normally air-filled (black) space or as its complete opacification and suggests a fracture of the basal denomination frequently associated with Battle’s sign.  Maxillary sinuses are unaerated (gray) when filled with blood, but are black when filled with air.  Maxillary sinus opacification is the CT correlate of “raccoon’s eyes” seen on physical examination.  The middle ear is normally aerated; therefore, opacification at this site may suggest an accompanying hemotympanum.

Look for fractures of maxillary, frontal, and ethmoid bones.  Next, look at the intraorbital contents: The contents of the bony orbits are black (fat).  The optic nerves originate at the back of the globe and extend back towards the optic canal.  Disruption of the optic nerve is sometimes discernible on a cut through the orbit.  However, disruptions in the bone and fragment placement are also suggestive of optic nerve compression, distortion, or transection.

Although the foramen magnum is well visualized on CT, herniation at this level is difficult to visualize because of bone artifact in the posterior fossa.

FIGURE Petrous foramina

The integrity of the bony structures of the skull base can be assessed from careful review of the CT bone windows that show the petrous bone and its many foramina (entry and exit holes) for cranial blood vessels and nerves.  The foramina rotundum (passageways right and left for second division of CN5), the ovale (for third division of CN5), and the spinosum (admits middle meningeal artery) can all be seen on most contemporary CT scans.

The internal auditory canal is seen if the axial cut intersects it and is a guide to the location of temporal bone fractures. The middle ear is normally aerated; therefore, opacification at this site may suggest an accompanying hemotympanum.

Other easily identified skull base structures include the orbits (which are filled with fat) and the aerated bony sinuses, including the frontal, maxillary, ethmoid, and sphenoid sinuses as well as the mastoid air cells. 

The optic nerves originate at the back of the globe and extend back towards the optic canal.  Disruption of the optic nerve is sometimes discernible on a cut through the orbit.  However, disruptions in the bone and fragment placement are also suggestive of optic nerve compression, distortion, or transection.

Bleeding into the air-filled sinuses can appear as a meniscus of fluid within the normally air-filled (black) space or as its complete opacification and suggests a fracture of the basal denomination frequently associated with Battle’s sign.  Maxillary sinuses are unaerated (gray) when filled with blood, but are black when filled with air.  Maxillary sinus opacification is the CT correlate of “raccoon’s eyes” seen on physical examination. 

Although the foramen magnum is well visualized on CT, herniation at this level is difficult to visualize because of bone artifact in the posterior fossa.

CRANIAL VAULT

Fractures of the cranial vault are identified by reference first to the cranial sutures, which are symmetric.  Fractures occur along cranial suture lines, but when they do, the width of the lucent lines is greater than that of the nonfractured suture line.  These suture lines include the coronal, lambdoid, and sagittal sutures, any of which can be confused with a fracture.

Disruptions of the smooth internal and external contours of the cranial bones (inner and outer tables) suggest fracture.  Low-density lines traversing the bone can represent fractures.  Facial fractures, particularly those involving the zygoma, maxilla, and orbits are visible on CT.  The air-filled sinuses provide good contrast for detection of fractures through their surrounding high-density, white bony walls.  Fractures of the facial bones on CT can appear at disruptions of cortical margins or lucent lines through bone.

Fractures of the cranial vault are identified by reference first to the cranial sutures, which are symmetric.  Fractures occur along cranial suture lines (“diastatic” (<???) fractures), but when they do, the width of the lucent lines is greater than that of the nonfractured suture line.  These suture lines include the coronal, lambdoid, and sagittal sutures, any of which can be confused with a fracture.

Disruptions of the smooth internal and external contours of the cranial bones (inner and outer tables) suggest fracture.  Low-density lines traversing the bone can represent fractures.  Facial fractures, particularly those involving the zygoma, maxilla, and orbits are visible on CT.  The air-filled sinuses provide good contrast for detection of fractures through their surrounding high-density, white bony walls.  Fractures of the facial bones on CT can appear at disruptions of cortical margins or lucent lines through bone.

(iii)              Dura

The dura, the next layer moving inward, can be visualized in some but not all locations on CT.  Dural structures that can be evaluated on the CT scan include the tentorium and the falx.  Sinus thrombosis appears as a high density in the region of the torcular Herophili.  The sagittal sinus is often seen as a lower-than-dural density triangular structure just inside a higher density triangle of dural leaves.  Thrombosis of the sinus can result in uniform density of the sinus triangle and its content.

Midline shift and mass effect can displace dural structures.  Because it is only a few millimeters thick and in a plane not parallel to the 20 degrees of the axial images, the tentorium is not usually well visualized on CT.  The falx, denser than the adjacent parenchyma and frequently calcified, can appear nearly as dense as bone.

Air in the subarachnoid space with low-density blackness in sulci is the best sign that the dura has been perforated at some point.   In the event that subarachnoid air is seen, its source may be found on review of the bone window film for facial sinus or mastoid fracture.   Subarachnoid air (frequently reported as “pneumocephalus”) is not always associated with CSF leak but frequently is when associated with mastoid or frontal fracture.  

An irregularity of the skull, with a portion of it placed interiorly, may have been the site of a laceration of the dura.  There may be a high-density mass suggestive of contusion or bleeding in brain substance just under the dura at the site of a depressed fracture fragment.

A hematoma formed by high-pressure arterial blood tends to displace the dura, dissecting it off of the overlying cranium and thereby forming a hematoma that is shaped like a biconvex lens.  Although not absolutely pathognomonic, the “lens” configuration of the hematoma, especially in the presence of as temporal fracture of the skull without prominent swelling of the underlying brain, is suggestive of a local impact epidural hematoma.

A subdural hematoma that is formed by transection of low-pressure veins, spanning the space from the surface of the brain to the sagittal or other dural sinuses, fills the space above the arachnoid and below the dura.  This subdural hematoma follows a path of low resistance, assuming a crescent shape.  Because the force required to tear the bridging veins is greater and involves more displacement of the brain with respect to the skull, small subdural hematomas can be associated with disproportionate amounts of cerebral swelling with midline shift.

Because the dura is so thin compared with the other structures and the resolution of the CT scanner is not good enough to detect structures so thin, the dura may be identified only by default.  A hematoma formed by high-pressure arterial blood tends to displace the dura, dissecting it off of the overlying cranium and thereby forming a hematoma that is shaped like a biconvex lens.  Although not absolutely pathognomonic, the “lens” configuration of the hematoma, especially in the presence of as temporal fracture of the skull without prominent swelling of the underlying brain, is suggestive of a local impact epidural hematoma.

(iv)              Brain surface

Evaluation of the brain surface on a CT scan includes looking for subarachnoid blood, which is frequently present in quantities that are insufficient to layer out enough for detection on head CT.

Blood vessels can be, but are not reliably, visualized by head CT.  Although best visualized following administration of intravenous (i.e., intravascular) contrast, major vascular structures can be visualized by their density and round shape.

(v)               Gray and white matter

Gray and white matter can be differentiated on CT scans because gray matter is more cellular  (and thus denser) than myelin-rich white matter.  Normally there is a distinct border between the gray and white matter on the CT soft tissue windows (latter appears lighter than former).  Loss of the gray-white boundary is suggestive of diffuse cerebral edema.  Cerebral edema is associated with an increase in the water content of the gray matter, which reduces its density (and Hounsfield value), making it darker (closer in appearance to white matter).

(vi)              CSF spaces

Replacement of CSF density cisterns by brain parenchymal density tissue (“obliteration of the cistern”) indicates that brain tissue, either globally or focally in the region adjacent to that cistern, is swollen.  If the cistern is next to or surrounds the brainstem, its obliteration is a sign that brain tissue from another compartment (i.e., herniated) is pressing on the brainstem at that site.

Cisterns that can be asymmetrically obliterated (and that should be evaluated on every head CT scan) include the cisterna magna, mesencephalic cistern, prepontine cistern, and all sulci and gyri.  Loss of the contour of the CSF space over gyri and within sulci suggests tissue expansion and mass effect.

Posterior fossa cisterns that can be evaluated on CT and that are important in detection of phenomena such as herniation include the cisterna magna, which appears as a CSF density space posterior to the cerebellum.  The perimesencephalic cistern (surrounding the midbrain and site of exit from the brainstem of the third cranial nerve) is also important in assessing brainstem compromise.  Herniation of the mesial portion of the temporal lobe (the uncus) over the tentorium, into the perimesencephalic cistern, with compression of the third nerve shortly after its exit from the midbrain, results in the characteristic “blown” pupil.. Asymmetry of this posterior fossa cistern is most ominous in terms of impending or evolving brainstem compression.

The interhemispheric cistern and that around the pineal body may be asymmetric in the presence of hemispheric swelling.  The size, configuration, and location of the cerebral ventricles can also be altered by head trauma.  Effacement refers to a decrease in the size of CSF-filled spaces seen on CT and thus is applicable to ventricles as well as cisterns and sulci.

The appearance of the anterior and posterior portions of the lateral ventricles is mirror images on axial CT cuts.  Asymmetry of the ventricles usually means that asymmetric intrahemispheric forces are compressing one or the other.  The third ventricle appears on CT as a vertically oriented line between two lobes of the thalamus.  The fourth ventricle is located between the pons and the cerebellum.  Shift of either normally midline CSF-filled cavity suggests the presence of mass effect.

FIGURE:  Cisterns and cerebral edema

(c)               3D

The surgeon needs to be able to mentally construct a three-dimension representation of intracranial lesions and nearby structures from the series of axial CT images.

With other forms of x-ray, objects are localized in three-dimensional space by determining the object’s location in the intersection of two flat planes.  With multiple tomographic views, the three-dimensional plane is mentally reconstructed by stacking axial views vertically with intervening extrapolation, creating a virtual image of the head and brain.

The x and y axes define the plane of the CT axial images.  The x axis moves left to right; the y axis moves up and down.  The z axis is orthogonal to the xy plane.  The CT scan provides explicit visual information about the x and y axial locations of structures on different slices, with respect to a z (vertical) axis orthogonal to the plane of the axial cuts.  The surgeon must use anatomic clues, such as structures readily recognized on the axial views, to determine the location of a given slice with respect to the z axis of the patient’s head.

The pinnae of the ears, the glabella of the nose, and the coronal suture are readily seen extracranial landmarks for localization along the z axis. The pinnae are at the level of the floor of the temporal fossa; the glabella is at the floor of the frontal fossa.

Intracranial landmarks, such as the orbital floor, the mastoid air cells, pineal calcification, third ventricle, and cerebral peduncles, and Sylvian fissures are also useful for determining vertical lesion location.  The internal auditory canal is a divot in the petrous bone that is a useful intracranial landmark for localizing the pons at the level of the facial and vestibulocochlear nerve complex.

FIGURE: Internal auditory canal on CT (IAC on CT.tif)

(iii)              Diagnosis

Detection of CT abnormalities and their correct pathologic attribution depends on knowledge of the normal appearance and anatomic anatomic arrangement of universally present structures.  Fractures, hematomas, edema, and mass effect are the pathologic abnormalities with whose CT appearance the head trauma surgeon must be familiar in order to intelligently and effectively diagnose and plan treatment.

(a)               Fracture

(i)                 Hematomas

The head CT provides information about an intracranial hematoma-location, size, and consistency-that is important in planning a craniotomy.

The location must be determined in relation to the skull, the dura, and the brain parenchyma, as well as to both intra- and extracranial lobes and surface landmarks.  Structures contiguous to a traumatic hematoma, potentially injured in evacuation of same, should be identified on the CT preoperatively so that they are approached expectantly rather than stumbled on or into at surgery.

The size of the hematoma is measured with reference to the vertical line with centimeter interval tick marks seen on each axial image.  The volume is calculated as the length and depth of the hematoma, multiplied by the number of slices, then divided by the centimeter interval between the slices.

Active bleeding can be identified by the tissue characteristics described above.

(ii)                Edema

(iii)              Mass effect

(iv)              Planning

(a)               Exposure

(b)               Craniotomy

(c)               Hazards

 

TABLE: Head CT abnormalities in the head injured patient

 

(2)               Angiography

Some trauma centers do not have CT scanning available.  Vascular displacements seen on angiogram were used pre-CT to determine if an extra-axial or intraparenchymal clot is present.

Angiography may also be helpful if a patient lost consciousness before head trauma, suggesting a possible bleed from an intracranial vascular malformation as the cause of a fall, motor vehicle accident, etc.  The unwary surgeon who approaches the clot as if it were nothing more than a traumatic intracerebral hematoma may, within seconds, be inundated with arterial blood on entry into arteriovenous malformation. 

An unusual contusion location (namely, above the floor of the anterior of middle cranial fossae) is a clue that a hematoma is caused by something other than head trauma, such as an anterior communicating artery rupture.  When hemorrhages occur in locations other than at the bases of the frontal or temporal lobes typical for traumatic contusions, a preoperative angiogram should be considered to rule out a vascular malformation.

 

Some trauma centers do not have CT scanning available.  These centers must rely on examination of cerebral vasculature through use of an angiogram to determine if an extra-axial or intraparenchymal clot is present.

Angiography may also be helpful if a patient lost consciousness before head trauma, suggesting a possible bleed from an intracranial vascular malformation.  The unwary surgeon who approaches the clot as if it were nothing more than a traumatic intracerebral hematoma may, within seconds, be inundated with arterial blood on entry into arteriovenous malformation.  An unusual contusion location (namely, above the floor of the anterior of middle cranial fossae) is a clue that a hematoma is caused by something other than head trauma, such as an anterior communicating artery rupture.  When hemorrhages occur in locations other than at the bases of the frontal or temporal lobes typical for traumatic contusions, a preoperative angiogram should be considered to rule out a vascular malformation.

c)                  Social

If historical, physical examination, and imaging data consistently indicate surgical intervention, social (including legal) considerations must be factored into the decision as to whether or not to operate to remove a traumatic intracranial hematoma.  Family members know about living wills or other advance directives which should be respected during the decision-making process. 

Very old patients, patients with terminal diseases (cancer, end stage HIV infection,… may not be surgical candidates regardless of their GCS or the size of an intracranial hematoma.  While the social and ethical controversies surrounding the concept and management of brain death are many, there is universal agreement that patients actually or iminently brain dead, there is no indication for surgical intervention. 

The neurotrauma surgical decision with the greatest social and ethical overlay is whether or not to operate on patients with GCS of 3 and 4.  Although the surgeon may have strong personal and religious reasons for opposing surgical intervention, the family’s are precedent.  Surgical informed consent conversations can drag on for hours while families agonize over the quality-versus-preservation of life decision.

TABLE Consent for and refusal of treatment

AUTHORITY

SIGNIFICANCE

Patient

Gold standard

Advanced directive

 

Conservator

 

Family

 

Friends

 

d)                  Grade

A Surgical Candidacy Grading system quantitates findings and can be the basis for surgical decision making.

TABLE Surgical Candidacy Grade

 

 

 

 

 

 

 

 

 

2.       Implementation<