I m a g i n g o f Tr a u m a t i c B r a i n In j u r y Uttam K. Bodanapally, MBBS, Chandler Sours, PhD, Jiachen Zhuo, PhD, Kathirkamanathan Shanmuganathan, MD* KEYWORDS Trauma Brain injury Computed tomography MR imaging
KEY POINTS Most patients with traumatic brain injury (TBI) have mild TBI (mTBI) and typically have no abnormalities on computed tomography (CT) and conventional MR imaging. Advanced MR imaging techniques including diffusion weighting, functional imaging, and spectroscopy are potential biomarkers for mTBI. CT is the initial diagnostic test in TBI. Conventional MR imaging in the acute phase is used as a problem solver when CT does not explain the neurologic deficit.
Traumatic brain injury (TBI) is a major health and socioeconomic concern throughout the world1,2 and is the leading cause of mortality and morbidity among young people.3 The incidence of TBI is increasing in the developing countries, because of the increase in the number of motor vehicle accidents.3 In advanced nations, the incidence of TBI caused by falls among the aging population is increasing4 and is changing the occurrence of different forms of TBI, specifically increasing the incidence of focal brain injuries in the form of contusions. Meanwhile the incidence of diffuse axonal injury (DAI) caused by high-velocity traffic accidents is decreasing in developed nations.4 In the United States, an estimated 1.1 million emergency department visits and 235,000 hospital admissions occur yearly because of TBI.5 Although most of these injuries are categorized as mild TBI (mTBI), a considerable number of these patients nevertheless experience permanent deficits.6 Approximately 52,000 deaths are attributed to TBI per year in the United States.7–9 TBI principally
affects young men, resulting in lost productivity because of disability and lost years because of death. The financial burden to society is estimated to be more than $60 billion per year in the United States alone.10 This article discusses the role of imaging in diagnosis and the spectrum of findings seen in patients with mild, moderate, and severe TBI.
CLASSIFICATION TBI is usually classified by clinical severity using the Glasgow Coma Scale (GCS).11 The mortality from TBI is related to the severity of injury as determined by GCS score.12,13 GCS (range, 3–15) consists of the sum of the 3 component scores (eye, motor, and verbal scales): mTBI, greater than 12 to less than or equal to 15; moderate TBI, greater than 8 to less than or equal to 12; severe TBI, less than or equal to 8. TBI has also been classified according to the severity of structural damage based on neuroimaging.14 However, the classification systems based on neuroimaging have limitations, because of severe underestimation of the extent of DAI by imaging modalities.
Funding support: None. Disclosures/Conflicts of interest: None. Department of Diagnostic Radiology & Nuclear Medicine, University of Maryland Medical Center, 22 South Greene Street, Baltimore, MD 21201, USA * Corresponding author. E-mail address: [email protected] Radiol Clin N Am 53 (2015) 695–715 http://dx.doi.org/10.1016/j.rcl.2015.02.011 0033-8389/15/$ – see front matter Ó 2015 Elsevier Inc. All rights reserved.
radiologic.theclinics.com
INTRODUCTION
696
Bodanapally et al Also, the systems broadly place injuries into diffuse and focal categories and fail to account for the specific type of mass lesions (eg, epidural vs subdural). The lack of specification fails to correctly classify patients with combined DAI and focal injuries.4,15
TYPES OF BRAIN INJURY TBI is divided into primary and secondary injuries.4,5,16 Primary injuries occur as a direct result of traumatic impact. Secondary injuries result from a complex biochemical cascade of events that exacerbates the primary injury by resulting in cerebral edema and herniation.17 The primary parenchymal lesions contain an epicenter with axons, glial cells, and vascular structures that sustain irreversible damage.16 Primary injury can result in either focal or diffuse lesions. Primary injuries at the macroscopic level are calvarial fractures, extra-axial hemorrhage (epidural hematoma [EDH], subdural hematoma [SDH], subarachnoid hemorrhage [SAH], and intraventricular hemorrhage [IVH]), and intra-axial injuries (contusions, DAI, and brain stem injury).16 At the cellular level, the initial events include microporation of membranes, leaky ion channels, and changes in the intracellular proteins that occur minutes to hours after initial injury.4 Surrounding the epicenter is traumatic penumbra with cells that have sustained reversible damage.16 Penumbra is the site where most of the deleterious secondary biochemical changes occur. The extent of various physiologic changes that occur during the early or late posttraumatic period, such as hypoxia, hypotension, pyrexia, and coagulopathy, may exacerbate the secondary events and determine the evolution of penumbra either into irreversible lesions or complete resolution.4 The evolutionary changes in penumbra explain the appearance of new lesions not apparent on initial scans.18 Secondary injury is initiated by various pathophysiologic cascades of events that follow the initial injury at both cellular and macroscopic levels to manifest as secondary lesions on neuroimaging. The cellular reactions that develop over hours and days include neurotransmitter release, free-radical generation, calcium-mediated damage, gene activation, mitochondrial dysfunction, and inflammatory responses.4 The release of neurotransmitters exacerbates the already leaky ion channels that cause primary injury and results in astrocytic swelling and cerebral edema. Cell necrosis is mainly caused by free-radical generation and calcium-mediated injury. Gene activation and expression of proapoptotic protein factors cause
apoptotic astrocytic and oligodendrocytic cell death. Mitochondrial dysfunction can decrease adenosine triphosphate production and oxygen consumption, which can further lead to axonal necrosis and apoptosis.19 The differentiation of TBI into primary and secondary injuries is important, because secondary injuries are often preventable, whereas primary injuries are not.16
MILD TRAUMATIC BRAIN INJURY Most patients with TBI (up to 75%) are considered to have mTBI.20,21 Computed tomography (CT) and conventional MR imaging examinations are typically normal and advanced neuroimaging techniques are required. Diffusion-weighted imaging (DWI), functional MR (fMR), arterial spin labeling (ASL), and spectroscopy show structural and functional abnormalities (Fig. 1).22–29 Studies also indicate that these techniques can be used as biomarkers to diagnose and monitor recovery. Blunt force from contact sports and blast injuries is a common mechanism for mTBI. The angular or rotational force that produces this injury results in widespread, diffuse effect on the entire brain parenchyma. Early diagnosis of mTBI and careful follow-up imaging to monitor healing holds the potential to prevent long-term neurodegenerative processes, such as chronic traumatic encephalopathy, that occur as a long-term complication of repetitive mTBI.22,30 The criteria to diagnose the 2 types of mTBI are shown in Table 1.31 Unlike mTBI, patients with complicated mTBI have minor structural abnormality on CT or conventional MR imaging.
POSTCONCUSSIVE SYMPTOMS Although most patients with mTBI are discharged quickly from the trauma center, a significant portion (w40%) of patients with mTBI remain impaired for at least 3 months, and a substantial number of these patients show deficits up to 1 year after injury.20,21,32 One year following injury, 82% of patients with mTBI reported the presence of at least 1 postconcussive symptom.33 The resulting lost productivity has socioeconomic consequences. Postconcussive symptoms include neuropsychological (difficulty with socializing, depression, anxiety), cognitive (attention, executive function, working memory, reduced information processing speed), and somatic symptoms (headaches, chronic pain, sensory perception disorders, language difficulty).34–38 These symptoms may also be present in patients with mTBI who lack evidence of intracranial injury on conventional
Imaging of Traumatic Brain Injury
Fig. 1. Diffusion tensor imaging (DTI) fiber tracks and cognitive performance in a patient with mTBI admitted following a fall with no finding on CT or conventional MR imaging examinations. (A) Normal subject, (B) acute (7 days), and (C) subacute (1 month) stages show that disrupted fiber tracks from the corpus callosum partially recover in the subacute phase. (D) The decline in cognitive performance also shows corresponding recovery longitudinally.
CT or MR imaging, suggesting that more advanced MR imaging techniques are needed to fully characterize this supposedly mild injury.
MILD TRAUMATIC BRAIN INJURY Diffusion-Weighted Imaging (Diffusion Tensor Imaging and Diffusion Kurtosis Imaging) Diffusion MR imaging is an effective tool to study various neurologic disorders because it provides Table 1 Definition of mTBI Uncomplicated mTBI
Complicated mTBI
GCS 13–15 Loss of consciousness 0–30 min Alteration in consciousness or mental state for a moment up to 24 h Posttraumatic amnesia 0–1 d Normal structural imaging (CT or conventional MR imaging)
GCS 13–15 Loss of consciousness 0–30 min Alteration in consciousness or mental state for a moment up to 24 h Posttraumatic amnesia 0–1 d Minor structural changes on CT or conventional MR imaging
an in vivo measurement of changes in tissue microstructure. Diffusion tensor imaging (DTI) measures water diffusion in at least 6 directions to obtain an appropriate representation of a diffusion tensor, describing the preferential diffusion direction and an ellipsoidal diffusivity profile. Measurements such as mean diffusivity (MD), often referred to as apparent diffusion coefficient (ADC), and fractional anisotropy (FA) can be measured from such an acquisition; MD or ADC measures the overall diffusivity in the tissue, and FA measures the degree of diffusion anisotropy. Intact axons have high FA, because diffusion is greater along the axons (axial diffusivity) than perpendicular to it (radial diffusivity). Damaged axons have reduced FA, because of either reduced axial diffusivity (typically a result of axonal injury) or increased radial diffusivity (typically a result of myelin damage). Damaged axons may be visible when reconstructing white matter axons using diffusion tractography, a technique that provides axonal structure by measuring the principal diffusion direction within each voxel (see Fig. 1).39 The disrupted fiber tracks observed at the acute stage tend to recover at the subacute stage, corresponding with the patient’s cognitive function decline and symptoms followed by recovery. As a result of secondary injuries, patients experience varied recovery paths, some enduring to
697
698
Bodanapally et al chronic stages. In the chronic stages of TBI, commonly damaged regions as indicated by DTI include the corpus callosum,32,40–42 internal capsule,32,43 and cingulum bundles.23 Furthermore, whole-brain analysis reveals more widely spread white matter abnormalities, including the superior and inferior longitudinal fasciculus, corona radiate, and frontal and temporal lobes, which are consistent with the diffuse nature of the injury.44,45 Although reduced FA and increased MD are consistent findings in the chronic stages of injury thought to represent cell death and axonal damage, results in the acute and subacute stages of TBI are more heterogeneous. TBI is a dynamically evolving injury in the weeks following the initial impact, with multiple primary and secondary injury mechanisms occurring simultaneously. Nevertheless, in mTBI populations at very early stages postinjury, researchers have noted an increased FA and reduced MD. This finding has been hypothesized to indicate an inflammatory response including cytotoxic edema and axonal swelling.32,46,47
Diffusion Kurtosis Imaging DTI is a sensitive tool to assess damage in white matter regions because of the sensitivity of FA to diffusion anisotropy changes; however, it has limited ability to assess alterations in gray matter regions because diffusion there is largely isotropic. Diffusion kurtosis imaging (DKI) has become the vanguard technique to probe the heterogeneity of the microenvironment of both gray and white matter. This technique overcomes a limitation of DTI by introducing a nongaussian term to the diffusion estimation model.48,49 In addition to the diffusion tensor parameters measurable by the DKI model, the extra kurtosis parameter, the mean kurtosis (MK), captures the nongaussian diffusion property arising from the tissue microstructure heterogeneity. In a recent animal model of TBI, DKI measurements were sensitive to reactive astrogliosis caused by mild inflammation.29 An increased MK was found in regions with reduced cerebral blood flow (CBF) measured by ASL (discussed later), without accompanying MD and fluid-attenuated inversion recovery (FLAIR) signal changes (Fig. 2). Regions may be affected by ischemia and inflammation following injury. In human studies of chronic mTBI, reduced MK was found in the thalamus and internal capsule, indicating a reduction of diffusion heterogeneity, which may suggest degenerative processes leading to neuronal shrinkage and changes in axonal and myelin density.50 Furthermore, in the mTBI population, longitudinal changes in MK from the subacute to
chronic stages of injury have been shown to correlate with changes in cognitive functioning.27
Functional MR Imaging fMR imaging is a valuable tool that can be used to noninvasively identify alterations in the communication within and between various neural networks. fMR imaging indirectly measures which regions of the brain are recruited to perform certain cognitive tasks or processes and interprets sensory inputs. This method is based on the MR signal differences between deoxygenated blood and oxygenated blood. When specific neurons for a given task are recruited or activated, there is an increase in freshly oxygenated blood to the local tissue to keep up with the increased neuronal activity. This change from deoxygenated blood to oxygenated blood in the activated region causes a change in the tissue signal as the local tissue changes from a predominantly paramagnetic state to diamagnetic state. This measured change is called the blood oxygen level–dependent (BOLD) signal. At present, fMR imaging data are acquired in one of 2 ways: using either a task-based fMR imaging paradigm or resting state fMR imaging paradigm.
Task-Based Functional MR Imaging Using a task-based paradigm, participants are instructed to perform a specific task while in the scanner. Researchers are able to indirectly measure which regions of the brain are recruited to perform that task based on the regions that show increased BOLD signal compared with resting conditions. Because of the susceptibility of the frontal lobe to the mechanical forces of TBI, task-based fMR imaging studies have focused primarily on attention and executive functioning deficits. Studies using task-based fMR imaging to investigate alterations in neural recruitment in patients with TBI have shown altered BOLD responses to tasks involving working memory,51–54 executive function,55 sustained attention,56 and inhibitory control.57 However, there is disparity with regard to increased52,53,58 or reduced59–61 task-induced activations in TBI populations because research groups have reported both. This discrepancy is likely related to differences in task and study design as well as inconsistencies in time since injury and the wide range in difficulty of tasks chosen in various studies.62 Many of these inconsistencies can be addressed through the use of resting state fMR imaging, which can be administered at all stages of injury and is not influenced by task difficulty, study design, or patient participation.
Imaging of Traumatic Brain Injury
Fig. 2. A 22-year-old woman admitted following a motor vehicle collision, with a GCS of 6. Axial (A) FLAIR, (B) FA, (C) MD, (D) MK, and (E) CBF maps show decreased CBF (arrow in E) and increased MK (arrow in D) without accompanying MD or FLAIR abnormality. These findings may be caused by ischemia and inflammation caused by reduced blood flow and tissue heterogeneity.
Resting State Functional MR Imaging Using a resting state paradigm, researchers are able to examine functional brain networks by measuring the interaction between brain regions. In this method, participants are not required to perform a task during the scan. Referred to as intrinsic functional connectivity or resting state functional connectivity (rs-FC), this method determines the strength of functional interactions between brain regions by measuring the correlations between small fluctuations in the BOLD signal across the brain.63–65 Biswal and colleagues63 discovered that regions recruited to perform a specific task display similar temporal patterns of fluctuations during resting conditions. At present, analysis can be conducted using a data-driven independent component analysis66
or a hypothesis-driven seed-based method. Resting state networks that are consistently replicated across both methods include networks that are associated with sensory systems (auditory, visual, somatosensory, and motor) and networks associated with cognitive processes.67 Through understanding the difference in neural network communication among patients with TBI, researchers will gain a greater insight into the pathophysiology of postconcussive symptoms. Both severe TBI and mTBI populations have shown reduced interhemispheric rs-FC,25,60,68 which is likely caused by the compromised integrity of the corpus callosum. TBI also has been shown to reduce rs-FC in neural networks that are not directly linked through the corpus callosum, including the motor network,69,70 thalamic network,24,71 task-positive network,25,72,73 and
699
700
Bodanapally et al the default mode network.29,72,74 It is hypothesized that the loss of structural integrity of the axonal tracts caused by both primary and secondary injury mechanisms contributes to the reduced rs-FC noted in many of these studies. However, some groups have provided evidence for increased rs-FC in the left frontoparietal network in patents with mTBI 75 and default mode network in patients with severe TBI.75 Furthermore, increased rs-FC has been found in mTBI populations between 2 networks that are generally found to be anticorrelated: the default mode network and task-positive network.72,74 An increased functional connectivity was observed between the thalamus and the motor cortex in patients with mTBI as opposed to controls, indicating a disturbed functional network (Fig. 3). Longitudinal studies examining resting state fMR imaging over the time course of injury are still needed for a more complete understanding of these alterations and the subsequent effects on recovery from injury. Altered rs-FC within these networks as a result of TBI is an active area of research that has the potential for providing valuable information on the cognitive condition of patients, especially during the acute stage when they are unable to perform tasks or if they are in a vegetative state.
Arterial Spin Labeling The MR imaging technique ASL is able to measure CBF using endogenous blood contrast by MR imaging. ASL has only recently been applied to TBI in the chronic stages. Using ASL during resting state conditions, patients with severe TBI show a global hypoperfusion,76 whereas patients with mTBI show reduced CBF specific to the thalamus. Recently, investigators took the novel approach to collect perfusion measures during task-based
conditions in order to tease apart the contribution of altered perfusion patterns to changes noted in task-based fMR imaging studies. For example, a severe TBI population showed hypoperfusion of the frontal lobe during a working memory task, suggesting cerebrovascular dysregulation of the cortical regions that modulate cognitive processes.76 As shown in Fig. 2, ASL measured reduced CBF in a patient with severe TBI with otherwise normal FLAIR signal. It is likely that ASL will become widely used in the near future because it has great promise as a prognostic tool in determining patient outcomes.
Spectroscopy MR spectroscopy (MRS) is unique in its ability to noninvasively measure cellular metabolites in vivo. Metabolites such as N-acetylaspartate (NAA), a neuronal and axonal marker; choline, a product of myelin breakdown products; lactate, a marker of anaerobic metabolism; and creatine (Cre) can be measured to evaluate metabolic changes in the brain following injury. A common finding following TBI is reduced NAA, which is suggestive of neuronal injury through DAI.77,78 Recent work in mild and moderate TBI found widespread reductions in NAA and increased choline, which correlated with neuropsychological assessments (Fig. 4).79 In addition, strong correlations have been observed between metabolite concentrations and neuropsychological performance, suggesting that MRS provides novel information that can be used to determine the long-term outcome of patients with TBI (see Fig. 4).26,80–83 MRS is able to determine biochemical changes within brain tissue that appears healthy based on conventional structural imaging techniques. MRS may be able to detect abnormalities long before
Fig. 3. Resting state fMR imaging from 78 patients with mTBI in the acute stage and 34 controls. It shows increased resting state functional connectivity between the thalamus and the primary motor cortex in the mTBI group compared with the control group. The contrast is shown at a voxelwise threshold of P
KEY POINTS Most patients with traumatic brain injury (TBI) have mild TBI (mTBI) and typically have no abnormalities on computed tomography (CT) and conventional MR imaging. Advanced MR imaging techniques including diffusion weighting, functional imaging, and spectroscopy are potential biomarkers for mTBI. CT is the initial diagnostic test in TBI. Conventional MR imaging in the acute phase is used as a problem solver when CT does not explain the neurologic deficit.
Traumatic brain injury (TBI) is a major health and socioeconomic concern throughout the world1,2 and is the leading cause of mortality and morbidity among young people.3 The incidence of TBI is increasing in the developing countries, because of the increase in the number of motor vehicle accidents.3 In advanced nations, the incidence of TBI caused by falls among the aging population is increasing4 and is changing the occurrence of different forms of TBI, specifically increasing the incidence of focal brain injuries in the form of contusions. Meanwhile the incidence of diffuse axonal injury (DAI) caused by high-velocity traffic accidents is decreasing in developed nations.4 In the United States, an estimated 1.1 million emergency department visits and 235,000 hospital admissions occur yearly because of TBI.5 Although most of these injuries are categorized as mild TBI (mTBI), a considerable number of these patients nevertheless experience permanent deficits.6 Approximately 52,000 deaths are attributed to TBI per year in the United States.7–9 TBI principally
affects young men, resulting in lost productivity because of disability and lost years because of death. The financial burden to society is estimated to be more than $60 billion per year in the United States alone.10 This article discusses the role of imaging in diagnosis and the spectrum of findings seen in patients with mild, moderate, and severe TBI.
CLASSIFICATION TBI is usually classified by clinical severity using the Glasgow Coma Scale (GCS).11 The mortality from TBI is related to the severity of injury as determined by GCS score.12,13 GCS (range, 3–15) consists of the sum of the 3 component scores (eye, motor, and verbal scales): mTBI, greater than 12 to less than or equal to 15; moderate TBI, greater than 8 to less than or equal to 12; severe TBI, less than or equal to 8. TBI has also been classified according to the severity of structural damage based on neuroimaging.14 However, the classification systems based on neuroimaging have limitations, because of severe underestimation of the extent of DAI by imaging modalities.
Funding support: None. Disclosures/Conflicts of interest: None. Department of Diagnostic Radiology & Nuclear Medicine, University of Maryland Medical Center, 22 South Greene Street, Baltimore, MD 21201, USA * Corresponding author. E-mail address: [email protected] Radiol Clin N Am 53 (2015) 695–715 http://dx.doi.org/10.1016/j.rcl.2015.02.011 0033-8389/15/$ – see front matter Ó 2015 Elsevier Inc. All rights reserved.
radiologic.theclinics.com
INTRODUCTION
696
Bodanapally et al Also, the systems broadly place injuries into diffuse and focal categories and fail to account for the specific type of mass lesions (eg, epidural vs subdural). The lack of specification fails to correctly classify patients with combined DAI and focal injuries.4,15
TYPES OF BRAIN INJURY TBI is divided into primary and secondary injuries.4,5,16 Primary injuries occur as a direct result of traumatic impact. Secondary injuries result from a complex biochemical cascade of events that exacerbates the primary injury by resulting in cerebral edema and herniation.17 The primary parenchymal lesions contain an epicenter with axons, glial cells, and vascular structures that sustain irreversible damage.16 Primary injury can result in either focal or diffuse lesions. Primary injuries at the macroscopic level are calvarial fractures, extra-axial hemorrhage (epidural hematoma [EDH], subdural hematoma [SDH], subarachnoid hemorrhage [SAH], and intraventricular hemorrhage [IVH]), and intra-axial injuries (contusions, DAI, and brain stem injury).16 At the cellular level, the initial events include microporation of membranes, leaky ion channels, and changes in the intracellular proteins that occur minutes to hours after initial injury.4 Surrounding the epicenter is traumatic penumbra with cells that have sustained reversible damage.16 Penumbra is the site where most of the deleterious secondary biochemical changes occur. The extent of various physiologic changes that occur during the early or late posttraumatic period, such as hypoxia, hypotension, pyrexia, and coagulopathy, may exacerbate the secondary events and determine the evolution of penumbra either into irreversible lesions or complete resolution.4 The evolutionary changes in penumbra explain the appearance of new lesions not apparent on initial scans.18 Secondary injury is initiated by various pathophysiologic cascades of events that follow the initial injury at both cellular and macroscopic levels to manifest as secondary lesions on neuroimaging. The cellular reactions that develop over hours and days include neurotransmitter release, free-radical generation, calcium-mediated damage, gene activation, mitochondrial dysfunction, and inflammatory responses.4 The release of neurotransmitters exacerbates the already leaky ion channels that cause primary injury and results in astrocytic swelling and cerebral edema. Cell necrosis is mainly caused by free-radical generation and calcium-mediated injury. Gene activation and expression of proapoptotic protein factors cause
apoptotic astrocytic and oligodendrocytic cell death. Mitochondrial dysfunction can decrease adenosine triphosphate production and oxygen consumption, which can further lead to axonal necrosis and apoptosis.19 The differentiation of TBI into primary and secondary injuries is important, because secondary injuries are often preventable, whereas primary injuries are not.16
MILD TRAUMATIC BRAIN INJURY Most patients with TBI (up to 75%) are considered to have mTBI.20,21 Computed tomography (CT) and conventional MR imaging examinations are typically normal and advanced neuroimaging techniques are required. Diffusion-weighted imaging (DWI), functional MR (fMR), arterial spin labeling (ASL), and spectroscopy show structural and functional abnormalities (Fig. 1).22–29 Studies also indicate that these techniques can be used as biomarkers to diagnose and monitor recovery. Blunt force from contact sports and blast injuries is a common mechanism for mTBI. The angular or rotational force that produces this injury results in widespread, diffuse effect on the entire brain parenchyma. Early diagnosis of mTBI and careful follow-up imaging to monitor healing holds the potential to prevent long-term neurodegenerative processes, such as chronic traumatic encephalopathy, that occur as a long-term complication of repetitive mTBI.22,30 The criteria to diagnose the 2 types of mTBI are shown in Table 1.31 Unlike mTBI, patients with complicated mTBI have minor structural abnormality on CT or conventional MR imaging.
POSTCONCUSSIVE SYMPTOMS Although most patients with mTBI are discharged quickly from the trauma center, a significant portion (w40%) of patients with mTBI remain impaired for at least 3 months, and a substantial number of these patients show deficits up to 1 year after injury.20,21,32 One year following injury, 82% of patients with mTBI reported the presence of at least 1 postconcussive symptom.33 The resulting lost productivity has socioeconomic consequences. Postconcussive symptoms include neuropsychological (difficulty with socializing, depression, anxiety), cognitive (attention, executive function, working memory, reduced information processing speed), and somatic symptoms (headaches, chronic pain, sensory perception disorders, language difficulty).34–38 These symptoms may also be present in patients with mTBI who lack evidence of intracranial injury on conventional
Imaging of Traumatic Brain Injury
Fig. 1. Diffusion tensor imaging (DTI) fiber tracks and cognitive performance in a patient with mTBI admitted following a fall with no finding on CT or conventional MR imaging examinations. (A) Normal subject, (B) acute (7 days), and (C) subacute (1 month) stages show that disrupted fiber tracks from the corpus callosum partially recover in the subacute phase. (D) The decline in cognitive performance also shows corresponding recovery longitudinally.
CT or MR imaging, suggesting that more advanced MR imaging techniques are needed to fully characterize this supposedly mild injury.
MILD TRAUMATIC BRAIN INJURY Diffusion-Weighted Imaging (Diffusion Tensor Imaging and Diffusion Kurtosis Imaging) Diffusion MR imaging is an effective tool to study various neurologic disorders because it provides Table 1 Definition of mTBI Uncomplicated mTBI
Complicated mTBI
GCS 13–15 Loss of consciousness 0–30 min Alteration in consciousness or mental state for a moment up to 24 h Posttraumatic amnesia 0–1 d Normal structural imaging (CT or conventional MR imaging)
GCS 13–15 Loss of consciousness 0–30 min Alteration in consciousness or mental state for a moment up to 24 h Posttraumatic amnesia 0–1 d Minor structural changes on CT or conventional MR imaging
an in vivo measurement of changes in tissue microstructure. Diffusion tensor imaging (DTI) measures water diffusion in at least 6 directions to obtain an appropriate representation of a diffusion tensor, describing the preferential diffusion direction and an ellipsoidal diffusivity profile. Measurements such as mean diffusivity (MD), often referred to as apparent diffusion coefficient (ADC), and fractional anisotropy (FA) can be measured from such an acquisition; MD or ADC measures the overall diffusivity in the tissue, and FA measures the degree of diffusion anisotropy. Intact axons have high FA, because diffusion is greater along the axons (axial diffusivity) than perpendicular to it (radial diffusivity). Damaged axons have reduced FA, because of either reduced axial diffusivity (typically a result of axonal injury) or increased radial diffusivity (typically a result of myelin damage). Damaged axons may be visible when reconstructing white matter axons using diffusion tractography, a technique that provides axonal structure by measuring the principal diffusion direction within each voxel (see Fig. 1).39 The disrupted fiber tracks observed at the acute stage tend to recover at the subacute stage, corresponding with the patient’s cognitive function decline and symptoms followed by recovery. As a result of secondary injuries, patients experience varied recovery paths, some enduring to
697
698
Bodanapally et al chronic stages. In the chronic stages of TBI, commonly damaged regions as indicated by DTI include the corpus callosum,32,40–42 internal capsule,32,43 and cingulum bundles.23 Furthermore, whole-brain analysis reveals more widely spread white matter abnormalities, including the superior and inferior longitudinal fasciculus, corona radiate, and frontal and temporal lobes, which are consistent with the diffuse nature of the injury.44,45 Although reduced FA and increased MD are consistent findings in the chronic stages of injury thought to represent cell death and axonal damage, results in the acute and subacute stages of TBI are more heterogeneous. TBI is a dynamically evolving injury in the weeks following the initial impact, with multiple primary and secondary injury mechanisms occurring simultaneously. Nevertheless, in mTBI populations at very early stages postinjury, researchers have noted an increased FA and reduced MD. This finding has been hypothesized to indicate an inflammatory response including cytotoxic edema and axonal swelling.32,46,47
Diffusion Kurtosis Imaging DTI is a sensitive tool to assess damage in white matter regions because of the sensitivity of FA to diffusion anisotropy changes; however, it has limited ability to assess alterations in gray matter regions because diffusion there is largely isotropic. Diffusion kurtosis imaging (DKI) has become the vanguard technique to probe the heterogeneity of the microenvironment of both gray and white matter. This technique overcomes a limitation of DTI by introducing a nongaussian term to the diffusion estimation model.48,49 In addition to the diffusion tensor parameters measurable by the DKI model, the extra kurtosis parameter, the mean kurtosis (MK), captures the nongaussian diffusion property arising from the tissue microstructure heterogeneity. In a recent animal model of TBI, DKI measurements were sensitive to reactive astrogliosis caused by mild inflammation.29 An increased MK was found in regions with reduced cerebral blood flow (CBF) measured by ASL (discussed later), without accompanying MD and fluid-attenuated inversion recovery (FLAIR) signal changes (Fig. 2). Regions may be affected by ischemia and inflammation following injury. In human studies of chronic mTBI, reduced MK was found in the thalamus and internal capsule, indicating a reduction of diffusion heterogeneity, which may suggest degenerative processes leading to neuronal shrinkage and changes in axonal and myelin density.50 Furthermore, in the mTBI population, longitudinal changes in MK from the subacute to
chronic stages of injury have been shown to correlate with changes in cognitive functioning.27
Functional MR Imaging fMR imaging is a valuable tool that can be used to noninvasively identify alterations in the communication within and between various neural networks. fMR imaging indirectly measures which regions of the brain are recruited to perform certain cognitive tasks or processes and interprets sensory inputs. This method is based on the MR signal differences between deoxygenated blood and oxygenated blood. When specific neurons for a given task are recruited or activated, there is an increase in freshly oxygenated blood to the local tissue to keep up with the increased neuronal activity. This change from deoxygenated blood to oxygenated blood in the activated region causes a change in the tissue signal as the local tissue changes from a predominantly paramagnetic state to diamagnetic state. This measured change is called the blood oxygen level–dependent (BOLD) signal. At present, fMR imaging data are acquired in one of 2 ways: using either a task-based fMR imaging paradigm or resting state fMR imaging paradigm.
Task-Based Functional MR Imaging Using a task-based paradigm, participants are instructed to perform a specific task while in the scanner. Researchers are able to indirectly measure which regions of the brain are recruited to perform that task based on the regions that show increased BOLD signal compared with resting conditions. Because of the susceptibility of the frontal lobe to the mechanical forces of TBI, task-based fMR imaging studies have focused primarily on attention and executive functioning deficits. Studies using task-based fMR imaging to investigate alterations in neural recruitment in patients with TBI have shown altered BOLD responses to tasks involving working memory,51–54 executive function,55 sustained attention,56 and inhibitory control.57 However, there is disparity with regard to increased52,53,58 or reduced59–61 task-induced activations in TBI populations because research groups have reported both. This discrepancy is likely related to differences in task and study design as well as inconsistencies in time since injury and the wide range in difficulty of tasks chosen in various studies.62 Many of these inconsistencies can be addressed through the use of resting state fMR imaging, which can be administered at all stages of injury and is not influenced by task difficulty, study design, or patient participation.
Imaging of Traumatic Brain Injury
Fig. 2. A 22-year-old woman admitted following a motor vehicle collision, with a GCS of 6. Axial (A) FLAIR, (B) FA, (C) MD, (D) MK, and (E) CBF maps show decreased CBF (arrow in E) and increased MK (arrow in D) without accompanying MD or FLAIR abnormality. These findings may be caused by ischemia and inflammation caused by reduced blood flow and tissue heterogeneity.
Resting State Functional MR Imaging Using a resting state paradigm, researchers are able to examine functional brain networks by measuring the interaction between brain regions. In this method, participants are not required to perform a task during the scan. Referred to as intrinsic functional connectivity or resting state functional connectivity (rs-FC), this method determines the strength of functional interactions between brain regions by measuring the correlations between small fluctuations in the BOLD signal across the brain.63–65 Biswal and colleagues63 discovered that regions recruited to perform a specific task display similar temporal patterns of fluctuations during resting conditions. At present, analysis can be conducted using a data-driven independent component analysis66
or a hypothesis-driven seed-based method. Resting state networks that are consistently replicated across both methods include networks that are associated with sensory systems (auditory, visual, somatosensory, and motor) and networks associated with cognitive processes.67 Through understanding the difference in neural network communication among patients with TBI, researchers will gain a greater insight into the pathophysiology of postconcussive symptoms. Both severe TBI and mTBI populations have shown reduced interhemispheric rs-FC,25,60,68 which is likely caused by the compromised integrity of the corpus callosum. TBI also has been shown to reduce rs-FC in neural networks that are not directly linked through the corpus callosum, including the motor network,69,70 thalamic network,24,71 task-positive network,25,72,73 and
699
700
Bodanapally et al the default mode network.29,72,74 It is hypothesized that the loss of structural integrity of the axonal tracts caused by both primary and secondary injury mechanisms contributes to the reduced rs-FC noted in many of these studies. However, some groups have provided evidence for increased rs-FC in the left frontoparietal network in patents with mTBI 75 and default mode network in patients with severe TBI.75 Furthermore, increased rs-FC has been found in mTBI populations between 2 networks that are generally found to be anticorrelated: the default mode network and task-positive network.72,74 An increased functional connectivity was observed between the thalamus and the motor cortex in patients with mTBI as opposed to controls, indicating a disturbed functional network (Fig. 3). Longitudinal studies examining resting state fMR imaging over the time course of injury are still needed for a more complete understanding of these alterations and the subsequent effects on recovery from injury. Altered rs-FC within these networks as a result of TBI is an active area of research that has the potential for providing valuable information on the cognitive condition of patients, especially during the acute stage when they are unable to perform tasks or if they are in a vegetative state.
Arterial Spin Labeling The MR imaging technique ASL is able to measure CBF using endogenous blood contrast by MR imaging. ASL has only recently been applied to TBI in the chronic stages. Using ASL during resting state conditions, patients with severe TBI show a global hypoperfusion,76 whereas patients with mTBI show reduced CBF specific to the thalamus. Recently, investigators took the novel approach to collect perfusion measures during task-based
conditions in order to tease apart the contribution of altered perfusion patterns to changes noted in task-based fMR imaging studies. For example, a severe TBI population showed hypoperfusion of the frontal lobe during a working memory task, suggesting cerebrovascular dysregulation of the cortical regions that modulate cognitive processes.76 As shown in Fig. 2, ASL measured reduced CBF in a patient with severe TBI with otherwise normal FLAIR signal. It is likely that ASL will become widely used in the near future because it has great promise as a prognostic tool in determining patient outcomes.
Spectroscopy MR spectroscopy (MRS) is unique in its ability to noninvasively measure cellular metabolites in vivo. Metabolites such as N-acetylaspartate (NAA), a neuronal and axonal marker; choline, a product of myelin breakdown products; lactate, a marker of anaerobic metabolism; and creatine (Cre) can be measured to evaluate metabolic changes in the brain following injury. A common finding following TBI is reduced NAA, which is suggestive of neuronal injury through DAI.77,78 Recent work in mild and moderate TBI found widespread reductions in NAA and increased choline, which correlated with neuropsychological assessments (Fig. 4).79 In addition, strong correlations have been observed between metabolite concentrations and neuropsychological performance, suggesting that MRS provides novel information that can be used to determine the long-term outcome of patients with TBI (see Fig. 4).26,80–83 MRS is able to determine biochemical changes within brain tissue that appears healthy based on conventional structural imaging techniques. MRS may be able to detect abnormalities long before
Fig. 3. Resting state fMR imaging from 78 patients with mTBI in the acute stage and 34 controls. It shows increased resting state functional connectivity between the thalamus and the primary motor cortex in the mTBI group compared with the control group. The contrast is shown at a voxelwise threshold of P
