The most common cause of death and permanent disability in the first few decades of life is trauma, and the neurologic components of trauma are responsible for most of these deaths and disabilities. In the United States, the annual incidence of TBI has been estimated at 180 to 250 per 100,000 population. As the leading cause of disability in people under 40, there are 15 to 20 per 100,000 TBI-related disabilities every year. The incidence of head injury peaks at 550 per 100,000 population for ages 15
to 24 years. The incidence then declines slightly until the age of 50 years, when it again starts to increase. About 500.000 cases of head injury can be expected to occur in the United States each year. However, these estimates almost certainly underrepresent the true incidence of TBI when accounting for all cases of mild TBI, which accounts for the vast majority of head injuries (5). For example, with sports-related concussion, a lack of systematic reporting and tendency to underreport when there has been no loss of consciousness (LOC) leads to a discrepancy in estimates and actual incidence (4,6). The yearly incidence of sports-related TBI alone in the United States has ranged from 300,000 to 3.8 million (4).

Within the severe spectrum of TBI, up to 10% of the new cases will be fatal. As many as 5% to 10% of those that survive the initial trauma will experience some degree of residual neurologic deficit (7). Head injury mortality rates are estimated at 25 per 100,000 population each year (7). Fatal head injuries are four times more common in males than females (7). Traffic-related injuries account for between 20% and 50% of head injury deaths (7). Gunshot wounds to the head are responsible for 20% to 40% of deaths. Falls and nonfirearm assaults account for most of the remaining deaths (7). Falls typically account for a higher percentage of injuries at both extremes of life. Seventy-five percent of head injuries in preschool children are secondary to falls (7). Falls are also responsible for the vast majority of head injuries in the elderly population. Two-thirds of all head injury deaths occur before hospitalization (7). Measures designed to reduce mortality from head injury, therefore, will be unproductive unless preventative actions are also incorporated (7).

Neurotrauma is unique among brain disorders with respect to the critical role that biomechanics contributes to the understanding of injury type and severity (8). Holbourn’s early pioneering work concerning the mechanisms of head injury in adults has served as the fundamental basis for understanding the means by which traumatic stresses produce cerebral injury (Fig. 13.1) (9,10). Using a gelatin model of the brain, Holbourn concluded that injury to the brain occurred through two major mechanisms: direct injuries due to skull distortion (contact phenomenon) and indirect or inertial injuries that arise irrespective of skull deformation. The former is produced by localized fracture or inbending of the skull with direct injury to adjacent brain parenchyma. Neural damage of this type is typically superficial and localized to the immediate vicinity of the calvarial injury and almost always isolated to moderate and severe injuries. Examples of lesions caused by this mechanism are cortical lacerations and contusions due to depressed fracture fragments, and EDHs secondary to lacerations of meningeal arteries.

The second mechanism of injury relates to inertial forces which operate irrespective of skull deformation and may produce extensive neural damage, even in the absence of a direct blow to the head. Two related strain metrics applied to model brain deformation after traumatic injury include (1) pressure-induced volumetric strain and (2) shear strain. Largely composed of water, the brain tissue is highly resistant to pressure-induced volumetric deformation and thus volumetric strain does not significantly contribute to TBI pathology and will not be considered further (8). In contrast, shear-strain deformation is characterized by a change
in shape without a change in volume and is responsible for most mechanically induced TBI lesions. Because of their inherently low rigidity, neurons are extremely susceptible to shear-strain deformations. These shear strains develop because of differential movements of one portion of the brain with respect to another. Shear–strain forces are greatest at the junction of tissues that differ in density and rigidity (e.g., CSF–brain, gray–white matter, brain–pia-arachnoid, pia-arachnoid–dura, skull–dura).

Early investigations on shear strain focused on linear accelerations measured during impact in animal models of TBI and correlations with pressure recordings within brain surrogates (11). A strong correlation between peak pressure, peak acceleration, and neurologic dysfunction was established (12). Similar studies examining embalmed human cadavers identified injury thresholds for skull fracture and intracranial pressure thresholds, resulting in a human injury tolerance curve, commonly referred to as the Wayne State Tolerance Curve (WSTC). The WSTC, which is based on linear acceleration thresholds, ultimately formed the basis for testing standards for protective equipment design and automotive safety systems (8).

Rotational acceleration is the second form of inertial loading contributing to shear-strain injury experienced by the brain at the moment of impact. Animal studies, largely performed subsequent to the development of the WSTC, have validated the early pioneering work of Holbourn and led to the now commonly accepted wisdom that shear strain secondary to rotational acceleration is largely responsible for the majority of traumatic cerebral injury (13). In these studies, primates subjected to high-magnitude angular inertial forces developed widespread axonal injury and prolonged unconsciousness. The complex, interconnected, oblique, and orthogonal orientations of many organized fiber bundles in the brain make its structure highly susceptible to diffuse injury as a result of shear forces resulting from rotational acceleration. This phenomenon can explain the often diffuse distribution of TAI, even in the absence of direct contact injury. Although traumatic coma is very difficult to produce when the head is constrained to exclude rotational motion, allowing rotational acceleration markedly increases the risk of TAI and an unconscious episode (14). However, the precise relative contributions of linear and rotational accelerations as well as other impact factors to neurologic dysfunction remain under active debate.

Unlike traumatic lesions from contact phenomenon, rotationally induced shear-strain lesions are typically multiple, bilateral, and widespread. They can occur remote from the site of impact, and may be either superficially or deeply situated (Figs. 13.2–13.5) (9,10). The location of lesions produced by this mechanism, as would be expected, closely corresponds to the locations of maximum shear strains that develop during rotational acceleration of the head. The expected anatomic distribution was experimentally mapped by Holbourn (9,10) for the three orthogonal planes of rotation using a gelatin brain model (Fig. 13.1). Despite the limitations of this model, the distribution of traumatic lesions in fatal head injuries and animal trauma models has generally followed Holbourn’s theoretical predictions. Figure 13.1 illustrates the location of maximum shear strain for all three orthogonal planes of rotation (9,10). Rotationally induced shear-strain injury typically produces lesions at one of the four topographic levels: cortical surface of brain (contusions), TAI, brainstem, and penetrating blood vessels (arteries or veins). The predominate types of lesions observed in a particular patient are determined by the specific mechanical circumstances present during trauma. Although neurons are the most susceptible tissue to rotationally induced shear-strain deformations, non-neuronal tissues (glia, penetrating blood vessels, bridging veins, pia-arachnoid) may also be injured by this mechanism (9,10).

In addition to biomechanical mechanisms, an understanding of the molecular and cellular changes underlying TBI can
inform our radiologic evaluation of underlying pathology. The biomechanical forces outlined above produce a primary injury which directly disrupts the normal structure and function of neurons, glia, and blood vessels. Subsequent to this primary insult, a vast array of secondary injury processes are initiated resulting in complex molecular and cellular alterations which may persist for days, weeks, and months after injury (15). Mechanical disruption of both neurons and glial cells results in distortions of membrane-bound ion channels with loss of normal electrochemical gradients necessary for maintenance of cell function. Associated mitochondrial failure depletes ATP supply and prohibits energy-dependent restoration of ionic gradients. Simultaneously, disruption of the vasculature results in opening of the blood–brain barrier and vasogenic edema. Loss of vascular autoregulatory mechanisms contributes to hypoxic conditions with resultant ischemia and cytotoxic injury (7,15,16,17,18). Widespread neuronal depolarizations are also implicated in expression of immediate early genes including those for excitotoxic neurotransmission and neuroinflammatory genes which further exacerbate tissue damage (19). At the axonal level, varying degrees of shear strain influence the extent of TAI (15). With less than 5% strain, transient axonal depolarizations occur with full membrane and functional recovery. Greater than 20% strain leads to membrane fragmentation, cytoskeletal breakdown, and primary axotomy. Intermediate strain levels (5% to 20%) activate a complex cascade of ionic fluxes, cytoskeletal disruptions, and myelin breakdown precipitated by axolemmal damage (15). These processes can occur over hours to weeks. As Bigler and Maxwell point out in their recent review, a broad appreciation of TBI pathophysiology is important for interpretation of neuroimaging studies (15). For example, alterations in blood-oxygen level–dependent (BOLD) contrast are noted after TBI in a growing body of functional MRI (fMRI) literature. Such findings may reflect, in part, primary neural damage, but consideration of microvascular pathology and autoregulatory dysfunction must be integrated into the interpretation of these results (15). We are entering an exciting era in neuroimaging where advanced imaging tools are beginning to noninvasively tease out these molecular and cellular underpinnings of TBI.

In the setting of acute head trauma, CT remains the imaging standard of care for evaluation of lesions requiring immediate neurosurgical intervention (i.e., acute EDH) (20). Advantages of noncontrast CT imaging for acute head trauma include rapid acquisition, widespread availability, and lack of any significant contraindications. As a diagnostic exam, CT provides excellent sensitivity for demonstrating mass effect, acute hemorrhage, bone injury, and ventricular size/displacement (21,22). For these reasons, CT remains the diagnostic study of choice for the initial evaluation of TBI patients (Table 13.2) (20). This is likely to be true for the immediate future, especially for the patient with

multiple-organ trauma. While CT imaging for acute moderate and severe TBI is clearly indicated, guidelines for screening CT in mild TBI (mTBI) are less well defined. Indications for head CT in patients with mTBI (defined as GCS 13 to 15) according to most major guidelines include LOC >30 seconds to 1 minute, prolonged altered or deteriorating level of consciousness, severe headache, focal neurologic deficit or seizure, or worsening symptoms (4).

The role of MRI in head trauma evaluation is evolving. Factors that limit the widespread use of MR as the primary diagnostic study for evaluation of trauma patients include its limited availability in the acute trauma setting, long imaging times relative to CT, sensitivity to patient motion, greater cost, greater difficulty of patient monitoring, lower sensitivity for detecting fractures, and physician unfamiliarity with the MR appearance of traumatic lesions (23,24,25,26). An additional major drawback to the use of MR in head trauma has been the logistic difficulty of safely imaging severely injured patients in the MR environment (27). These limitations have been greatly reduced, or eliminated, during the last few years and now offer little impediment to the use of MR for evaluating acute TBI patients (16,17). Notable advances include the development of self-shielded magnets, wider and more accessible scanning gantries, and a wide range of nonferromagnetic life-support and monitoring devices that are compatible with the MR environment. These developments have greatly facilitated the evaluation of these critically ill patients. Every physiologic parameter that might need to be monitored in these patients can be safely monitored (7). Sensors for monitoring respiratory rate and effort, invasive or noninvasive blood pressure, arterial oxygen saturation (pulse oximetry), heart rate and rhythm, end-expiratory CO₂ levels, electrocardiographic activity, temperature, and intracranial pressure are available. Multiple types of respirators, ventilators, and full anesthesia machines that are MR compatible are also available. The radiologist should be familiar with the many important concerns, principles, and techniques regarding patient monitoring in the MR environment.

With these considerations in mind, the decision to use MRI for diagnostic evaluation of TBI patients largely relies on the clinical scenario. While the most appropriate imaging test must be tailored to each individual patient, several general guidelines can be provided (Table 13.2). A significant factor to consider is the severity of injury as assessed by the GCS. As mentioned above, neurotrauma patients are routinely categorized into three clinical subgroups, depending on the degree of impairment of the admission GCS score: mild (GCS = 13 to 15), moderate (GCS = 9 to 12), and severe (GCS <9) (22).

Mild acute TBI (mTBI) is by far the largest subgroup of TBI patients, representing at least 80% of all traumatic head injuries, and likely much greater given the frequency of underreporting (6). Although the clinical definition for mTBI is variable, patients generally present after head trauma with transient altered consciousness of less than 30-minutes duration, posttraumatic amnesia not exceeding 24 hours, and GCS of 13 to 15 at presentation. The term “concussion” is often used in the sports literature synonymously with mTBI, although the absence of intracranial structural abnormalities on CT and conventional MRI sequences is implied. When traumatic intracranial lesions are identified on CT in patients with GCS 13 to 15, the term complex mTBI is often applied. Some clinicians maintain that mild head injury patients can be safely managed by clinical observation alone without initially obtaining any imaging study (22). Others, however, strongly advocate early diagnostic evaluation. As described above, according to most major clinical guidelines, indication for acute head CT include LOC for >30 seconds to 1 minute, prolonged altered consciousness, severe headache, seizure or other focal neurologic deficit, or worsening symptoms (4) (Table 13.2). This subgroup of patients has a 5% to 30% incidence of intracranial hemorrhage on initial head CT (22). Mild TBI patients are less likely than those with more severe degrees of injury, however, to have significant intraparenchymal lesions. For patients meeting clinical criteria for mTBI, CT is the most efficacious method for evaluating the lesions for many reasons including availability, speed, and high sensitivity and specificity for both fracture and intracranial hemorrhage (7). Conventional MRI sequences have been shown in numerous studies to have superior sensitivity compared with CT for detecting many types of traumatic injury in these patients, including TAI, small contusions, and small extra-axial hematomas (7,28). As a result, MRI is appropriate in the acute setting of mTBI when the clinical symptomology does not match the CT imaging findings (Table 13.2) (20). Moreover, due to its superiority for detection and characterization of nonsurgical injuries, MRI is recommended for evaluation of mTBI patients with persistent symptoms in the subacute and/or chronic stages of injury (20) (Table 13.2). Whereas these findings may not be known to alter the clinical management or outcome of mild head injury patients (22), there is increasing evidence that valuable prognostic information may be derived from subacute MRI (28,29,30). Although advanced MRI neuroimaging techniques, including diffusion tensor imaging (DTI), fMRI, susceptibility-weighted imaging (SWI), MR spectroscopic imaging (MRSI), and perfusion-weighted MRI (PW-MRI) may have utility in identifying and characterizing otherwise occult injury of brain trauma (discussed later in this chapter), these techniques are not yet recommended for guiding routine clinical management of mTBI patients (20).

The moderate and severe head injury categories share many clinical features and management concerns (7). These two groups, therefore, will be considered together. Neurologically unstable patients in these two categories should be initially studied with CT (7,20). These patients, by definition, already have significant impairment of consciousness and focal neurologic deficits that are deteriorating. The most critical issue in this situation is to rapidly detect potentially treatable hematomas or other surgically correctable lesions. Although CT may miss several important types of injuries, it is still the most efficient means of rapidly excluding surgical lesions (7). Moderate and severe head injury patients who are initially stable can be primarily evaluated by MR (7,16,17,27). It is advantageous to use MR in these patients, if possible, because it is considerably more sensitive for detecting most traumatic parenchymal and extraparenchymal lesions (7,17,27).

It is reasonable to suggest that all moderate to severe head injury patients should be evaluated with MR at some point in time during the first 2 weeks after injury (Table 13.2). The full extent of TBI will not be fully determined if only CT is used to evaluate this group of patient as MR is more valuable than CT for assessing the full magnitude of injury (7,17). It also provides more accurate information regarding the expected degree of final neurologic recovery (28,29). The MR exam can be done as the initial examination in stable patients (7,17,18,27). In unstable patients, however, it is probably the best to initially study the patients with CT and postpone the MR examination until they can be safely imaged. It is advisable to obtain the MR study within the first 2 weeks after injury, if possible, because most parenchymal lesions are more visible during this time period. Edema and axoplasmic leakage around areas of neuronal disruption are maximal during the first 2 weeks, rendering lesions more conspicuous. Smaller lesions are more difficult to detect over the ensuing weeks as intra- and extracellular edema gradually subsides (7). After the edema resolves, many traumatic intra-axial lesions may be quite indistinct on CT and standard T2-weighted MR scans (7).

CT remains the study of choice for the acute evaluation of neurologically or hemodynamically unstable patients who have significant impairment of consciousness (Table 13.2) (7).
The most critical issue in this situation is rapid detection of potentially treatable hematomas or other surgically correctable lesions. In unstable patients, it is unwise to spend even a few extra minutes to obtain an MR exam when CT can more quickly and safely answer the urgent questions. Similarly, CT can be used for evaluation of patients with rapid changes of neurologic status at any point in time after acute head injury, although emergency MR is increasingly common in this setting when available (Table 13.2).