Handbook of Clinical Neurology, Vol. 127 (3rd series) Traumatic Brain Injury, Part I J. Grafman and A.M. Salazar, Editors © 2015 Elsevier B.V. All rights reserved
Chapter 17
Current and future diagnostic tools for traumatic brain injury: CT, conventional MRI, and diffusion tensor imaging DAVID L. BRODY1*, CHRISTINE L. MAC DONALD1, AND JOSHUA S. SHIMONY2 Department of Neurology, Washington University School of Medicine, St. Louis, MO, USA
1
2
Department of Radiology, Washington University School of Medicine, St. Louis, MO, USA
INTRODUCTION
ASSESSMENT GOAL
The advent of the modern neuroimaging era with the development of the computed tomography (CT) scanner has fundamentally changed the way we think of traumatic brain injury (TBI). Instead of saying “a 24-year-old with a severe head injury due to a car crash” we started saying “a 24-year-old with a 2 cm right subdural hematoma, 15 mm midline shift, and bifrontal contusions due to a car crash.” This had immediate surgical implications. Then, with magnetic resonance imaging (MRI), instead of “she fell and hit her head on an icy sidewalk, but her CT was OK”, it has become “she fell and hit her head on an icy sidewalk and she has small bilateral temporal and right frontal hemorrhagic contusions on MRI.” While this does not have the same surgical implications, it certainly does provide objective evidence for a brain injury, which can assist with triage and rehabilitative planning. In the near future, instead of saying “concussion during a football game Friday night, clean MRI” we may say “concussion during a football game Friday night, reduced anisotropy in left anterior corona radiata and uncinate fasciculus on DTI.” This may help inform return-to-work and return-to-play decisions, along with many other implications. This review is written from two perspectives: that of a practitioner ordering the scans (DLB) and that of a radiologist reading and interpreting them (JSS). The review is divided up into assessments based on specific goals, rather than based on scan modalities. It is not meant to supplant but rather to supplement the authoritative previous works on this topic (Gean, 1994; Zee and Go, 2002). We also suggest a reading of the recommendations of the Common Data Elements Neuroimaging Working Group (Duhaime et al., 2010; Haacke et al., 2010).
Determine whether there is an immediately life-threatening intracranial lesion Clearly, a noncontrast head CT scan is the test of choice to address this issue, and such scans are virtually ubiquitously performed on all but the most trivially headinjured patients. Noncontrast head CT scans accurately resolve subdural, epidural, intraparenchymal, and intraventricular hemorrhages (Fig. 17.1). They reveal most types of clinically important herniation, most skull fractures, hydrocephalus, pneumocephalus, and most foreign bodies. Only in rare cases such as brainstem infarction due to traumatic arterial dissection is an immediately life-threatening traumatic lesion not revealed by noncontrast head CT. In our view, there is no need to replace head CT with any other modality at present. Only in rare cases, such as young children or pregnant women, where radiation exposure is a concern could MRI be considered to address the question of whether there is an immediately life-threatening intracranial lesion.
Assess for delayed progression of an intracranial lesion in stabilized patients Many intracranial lesions are not immediately lifethreatening, but have the potential to progress over time. Examples include delayed worsening of contusions, hemorrhages, cerebral edema, and hydrocephalus (Fig. 17.2). Again, noncontrast head CT scans are accurate methods for assessing progression of most of these lesions. Delayed vasospasm may occur after traumatic brain injury, perhaps especially following severe blast-related injuries (Armonda et al., 2006). CT angiography, MR
*Correspondence to: David L. Brody, Department of Neurology, Washington University School of Medicine, 660 South Euclid, Campus Box 8111, St. Louis, MO 63110, USA. Tel: +1-314-362-1381, E-mail: [email protected]
268
D.L. BRODY ET AL.
Fig. 17.1. Acute traumatic brain injury-related lesions, visualized on noncontrast head computed tomography (CT): (A) subdural hematoma; (B) epidural hematoma; (C) hemorrhagic contusion; (D) subarachnoid hemorrhage; (E) skull fracture and foreign body (bullet fragment); (F) pneumocephalus; (G) herniation (due to hemorrhage); (H) contusion and intraventricular hemorrhage with hydrocephalus.
Fig. 17.2. Subacute evolution of traumatic brain injury-related lesions, visualized on noncontrast head CT: (A) early trace hemorrhage; (B) later contusions “blossoming” in the same patient as in panel (A) rescanned 24 hours later; (C) early dilation of temporal horns heralding hydrocephalus; (D) diffuse edema with effacement of sulci.
angiography, and conventional digital subtraction catheter angiography (Fig. 17.3) are sensitive to vasospasm to varying degrees. Early intervention could potentially prevent the development of irreversible cerebral infarction, though this has not been definitively established.
Documentation of the presence of intracranial abnormalities for medicolegal reasons After a motor vehicle accident, assault, or work-related injury, the burden of proof is often on the injured person to document that there was in fact a brain injury. Signs and symptoms especially of mild or concussive injuries are sufficiently nonspecific that they often cannot be used in isolation to make a definitive diagnosis. In this setting, MRI is clearly superior to noncontrast head CT. Conventional MRI sequences (T1, T2, fluidattenuated inversion recovery (FLAIR)) without contrast reveal small contusions and hemorrhages that are missed
by CT (Fig. 17.4). In part this is due to the superior signalto-noise and spatial resolution, and in part due to fewer artifacts at interfaces between bone and brain. MRI is not immune from these artifacts, and even conventional MRI sequences are continuously improving. Modern blood-sensitive sequences such as high field gradientrecalled echo (GRE) and susceptibility-weighted imaging (SWI) have even greater sensitivity to small hemorrhages (Tong et al., 2004, 2008; Lee et al., 2008). A concern arises in that MRI may show “abnormalities” that are not related to traumatic brain injury. Examples include periventricular or small subcortical white matter T2 and FLAIR signal abnormalities which are common in the general population, and especially prevalent with increasing age and in those with cerebrovascular risk factors such as hypertension and diabetes. Also, cerebral amyloid angiopathy is a common cause of cerebral microhemorrhages unrelated to trauma (Fig. 17.4C). This vascular pathology is present in approximately 90% of subjects with Alzheimer’s
CURRENT AND FUTURE DIAGNOSTIC TOOLS FOR TRAUMATIC BRAIN INJURY
269
Fig. 17.3. Vasospasm following blast-related traumatic brain injury, visualized with digital subtraction catheter angiography. Narrowing of the right distal internal carotid artery (black arrows in a and b) and normalization of the arterial diameter after balloon angioplasty (white arrows in c and d). (Reproduced from Armonda et al., 2006, Figure 7B.)
disease pathology and common in others with cerebrovascular risk factors (Jellinger, 2002; Viswanathan and Chabriat, 2006). Ultimately, some subjective judgment is required based on the nature of the injury and the most common distributions of traumatic versus nontraumatic abnormalities in order to make a decision regarding the etiology of the signal changes. For example, in an extensive autopsy series, contusions were found most commonly in the frontal and temporal poles (Gurdjian and Gurdjian, 1976). These regions may be most vulnerable to mechanical forces due to rotational acceleration/ deceleration of the brain (Holbourn, 1943; Gennarelli et al., 1982). Limitations of MRI include longer scan times, reduced availability in the acute setting, contraindications due to metal, and claustrophobia.
Presurgical epilepsy planning Traumatic brain injury, especially severe closed skull injury and penetrating injury, has been well documented to cause epilepsy (Annegers et al., 1998). The posttraumatic seizures can often be controlled with standard antiepileptic medications, but when they are refractory,
surgical resections are frequently considered (Mathern et al., 1994; Jiang et al., 2004). In this setting, MRI is a key part of a multidisciplinary assessment of lesion identification and localization for presurgical planning (Fig. 17.5). Other methods such as ictal single-photon emission computed tomography (SPECT) or positron emission tomography (PET) scans are sometimes used in this context as well (Gilliam et al., 1997; Hogan et al., 1999; Henry and Roman, 2011). As stated above, retained metal objects (e.g., shrapnel) in the brain are a contraindication to MRI. A unilateral hippocampal abnormality apparent on MRI on the same side as interictal electroencephalogram (EEG) abnormalities is a strong predictor of seizure freedom after surgical resection (Gilliam et al., 1997). Again, conventional MRI sequences are rapidly improving in resolution, sensitivity, and signal-to-noise, allowing more and more subtle intracranial abnormalities to be localized (see Ch. 33).
Early prognostic decision making In an acute setting, prognostic decision making regarding withdrawal of care, organ donation, etc. following
270
D.L. BRODY ET AL.
Fig. 17.4. Acute traumatic brain injury-related lesions, visualized on structural MRI: (A) fluid-attenuated inversion recovery (FLAIR) image showing a small subdural hemorrhage and posterior corpus callosum abnormality (neither of which were apparent on CT); (B) susceptibility-weighted imaging (SWI) demonstrating multiple microhemorrhages (none of which were apparent on CT); (C) SWI in a patient with amyloid angiopathy with scattered microhemorrhages not related to traumatic brain injury; (D) FLAIR image demonstrating frontal contusions and small subdural hemorrhages; (E) FLAIR imagine demonstrating a small cortical contusion and subdural hemorrhage.
Fig. 17.5. Mesial temporal sclerosis following traumatic brain injury in a patient with refractory post-traumatic epilepsy. (A) Fluid-attenuated inversion recovery (FLAIR) approximately 1 year after a work-related traumatic brain injury, initially read as normal but in retrospect showing mild left hippocampal atrophy (red arrow). The patient’s “spells” were considered to be psychiatric in origin, or possibly motivated by secondary gain. (B) FLAIR repeated 4 years after injury, demonstrating progression of hippocampal atrophy and elevated signal consistent with mesial temporal sclerosis. At that time, video EEG demonstrated left temporal onset complex partial seizures. (C) T1-weighted MRI performed for presurgical planning. (D) Computed tomography (CT) scan after anterior amygdalohippocampectomy. This resulted in a near complete elimination of seizures, but some worsening of verbal memory function.
traumatic brain injuries is usually based on clinical criteria. CT scans showing devastating intracranial pathology such as severe bilateral cortical or thalamic damage can support these decisions. In rare cases where the clinical presentation is far worse than can be explained based on the CT scan, MRI may be helpful. For example, conventional MRI may reveal a brainstem or thalamic
infarction using diffusion-weighted imaging. MR angiography, CT angiography, or digital subtraction catheter angiography may reveal the source of the infarction, such as a traumatic vertebral dissection or basilar occlusion (Fig. 17.6). All three angiographic methods may be used confirm the absence of blood flow to the brain in the case of massively elevated intracranial pressure.
CURRENT AND FUTURE DIAGNOSTIC TOOLS FOR TRAUMATIC BRAIN INJURY
271
Fig. 17.6. Initially unexplained causes of impaired arousal following traumatic brain injury: (A) brainstem ischemia, apparent as a subtle pontine hypodensity on computed tomography (CT); (B) thalamic infarction, apparent as elevated signal on fluid-attenuated inversion recovery (FLAIR); (C) carotid dissection, apparent as intravascular low signal on magnetic resonance imaging (MRI); (D) carotid dissection following traumatic brain injury on digital subtraction angiography. Lesions indicated by red arrows.
Conversely, a profoundly comatose TBI patient with little apparent pathology on CT and conventional MRI may be considered a reasonable candidate for aggressive medical intervention, despite poor clinical status. The absence of major irreversible intracranial injury on a conventional MRI including diffusion-weighted imaging can indicate a reasonable chance for a good outcome. This can be reassuring, as early clinical signs and symptoms alone have limited prognostic power in the period soon after injury. This is commonly performed based mainly on clinical judgment; to our knowledge there have not been large prospective trials of the role of brain imaging for early prognostic decision making after severe TBI.
Prognostic assessment for rehabilitative planning Here we move out of the realm of what is currently standard practice, and into the domain of future applications. Nonetheless, a few motivating examples may help. It is very likely that a TBI patient with substantial bilateral medial temporal lobe contusions will have clinically important declarative memory deficits. Therefore, compensatory rehabilitation focused on automatic skill learning would be advisable. If the thalamus is completely ablated, it is unlikely that the patient will emerge from coma spontaneously and aggressive interventions may be warranted. Family, friends, and medical practitioners should be prepared for emotional dysregulation and personality changes in patients with large bifrontal injuries. It has not been established whether imaging-assisted rehabilitation planning is superior to rehabilitation based purely on observed clinical deficits and symptoms. Clearly, MRI is superior to CT for resolution of the exact extent and localization of injuries in the subacute setting. Newer modalities such as diffusion tensor imaging (see below) have greater predictive and correlative prognostic power for specific outcomes than other imaging methods. However, these methods
are not yet available for routine clinical use at present (see Chs 27 and 40).
Stratification of subjects for pathophysiologically targeted therapies A logical approach to the development of new therapeutics for traumatic brain injury would be to test them in patients with clear evidence for the pathophysiologic processes targeted by the candidate therapeutics (Saatman et al., 2008): medications designed to counteract coagulopathy should be tested in patients with clinically significant hemorrhage; edema reduction should be tested in patients with edema; axonal protectants should be tested in patients with axonal injury; oligodendroglial apoptosis inhibitors should be tested in patients with myelin injury, etc. This approach has not been taken historically, and this may account in part for the lack of success of the >30 major clinical trials that have been performed in TBI to date. A major challenge will be to develop imaging methods capable of resolving the relevant pathophysiologic processes with a clinically meaningful time window. Many processes will require intervention within a few hours, whereas MRI, PET, and other advanced imaging methods are typically not available late at night and in the weekend hours when TBIs often occur. Infrastructure changes, e.g., putting MRI scanners in emergency departments, or having at least one MRI scanner always available may be required to overcome this challenge. Significant advances along these lines in ischemic stroke research have been made recently (Lee et al., 2012). Of particular interest, traumatic axonal injury appears to be a nearly universal pathologic sequela of TBI. In one of the few published autopsy series of mild/concussive TBI cases (with deaths from other causes) traumatic axonal injury was present in 6/6 cases (Blumbergs et al., 1994). In contrast, hypoxic/ischemic damage, contusion, intraparenchymal hematoma, subdural hemorrhage, signs of elevated intracranial pressure, and skull fracture were absent. Likewise,
272
D.L. BRODY ET AL.
rotation injury in monkeys (Gennarelli et al., 1982) and pigs (Smith et al., 1997) resulted in axonal injury with little other pathology. This may indicate that the axons are the most mechanically vulnerable part of the brain. However, CT and conventional MRI sequences appear to substantially underestimate the extent of axonal injury. Stated another way, there can be extensive axonal injury even when CT and conventional MRI are normal. For this reason, the development of newer imaging methods sensitive to axonal injury has been a top priority for traumatic brain injury researchers. Other methods are discussed elsewhere in this volume, but diffusion tensor imaging (DTI) (Fig. 17.7) may be the most well established of these newer imaging methods (Niogi and Mukherjee, 2010; Mac Donald et al., 2011). DTI involves measurement of water diffusion in multiple directions and can be performed quickly on most clinical scanners (Pierpaoli et al., 1996). In brain white matter, water diffuses faster along the predominant fiber direction and more slowly in perpendicular directions. The resulting anisotropy (directional asymmetry) of water diffusion is high in intact axons and reduced after axonal injury (Arfanakis et al., 2002; Mac Donald et al., 2007a, b). Reduced anisotropy on DTI as a marker of traumatic axonal injury has been directly validated versus immunohistochemistry in experimental TBI, even when
conventional MRI is normal (Mac Donald et al., 2007a, b). Open questions include the extent to which injury to myelinated versus unmyelinated axons can be detected with DTI. We used DTI to test the hypothesis that blast-related TBI causes traumatic axonal injury in specific brain regions distinct from those typically injured in other forms of TBI (Mac Donald et al., 2011). Participants were 63 US military personnel evacuated to Landstuhl Regional Medical Center, clinically diagnosed with mild uncomplicated TBI, and scanned 1–90 days after injury. All had primary blast exposure plus another blast-related mechanism of injury (e.g., struck by a blunt object, fall, motor vehicle crash). Controls were 21 similar personnel with blast exposure and other injuries but no clinical diagnosis of TBI. DTI revealed abnormalities consistent with traumatic axonal injury in many TBI subjects. None had detectable intracranial injury on CT and only one had abnormalities on conventional MRI. By contrast, in 18/63 individual TBI subjects, there were significantly more DTI abnormalities than expected by chance ( p < 0.001). DTI was most markedly abnormal in the middle cerebellar peduncles (p ¼ < 0.001), cingulum bundles ( p ¼ 0.002), and right orbitofrontal white matter ( p ¼ 0.007). The middle cerebellar peduncle is rarely affected in conventional mild TBI, but predicted to
Diffusion Tensor Imaging
λ1
2. Calculate diffusion tensor for each voxel
λ2 λ3
A
Brain white matter: organized, myelinated axons
3. Separate parallel(λ1, axial) from perpendicular (λ2, λ3, radial) diffusion. Calculate anisotropy.
1. Collect diffusion weighted images in six or more directions
Traumatic axonal injury: simplified model
λ1
λ2
λ1
λ2
λ3
B
Axonal Disruption: reduced λ1 (axial), reduced anisotropy
λ3 Myelin Injury: incr. λ2, λ3 (radial), reduced anisotropy
λ1
λ2 λ3
Mixed Injury: greatly reduced anisotrophy
Fig. 17.7. Diffusion tensor imaging: theoretical basis for sensitivity to axonal injury. (A) Diffusion tensor imaging in normal white matter. (B) Predicted effects of axonal and myelin injury. (Adapted from M Budde.)
CURRENT AND FUTURE DIAGNOSTIC TOOLS FOR TRAUMATIC BRAIN INJURY
Providing pharmacodynamic biomarkers for candidate therapeutics If the challenges discussed above can be overcome, the same imaging methods could be used to assess the pharmacodynamic effects of the candidate therapeutics. For example, a relatively small number of patients could be sufficient to demonstrate that a candidate axonal protectant improves DTI signal abnormalities. Then, the question of whether prevention of progression of apparent axonal injury as indicated by DTI has a clinically meaningful benefit could be addressed in larger phase III clinical trial. If, on the other hand, a candidate therapeutic does not “hit the target” at the clinically achievable doses or administration times following injury, it may not be a good use of resources to proceed to a full-scale clinical trial. This strategy has not yet been implemented to our knowledge.
FUTURE DIRECTIONS From skull X-rays, to CT, to conventional MRI, to newer MRI methods such as DTI and GRE and SWI, each generation of imaging approaches demonstrates more abnormalities following traumatic brain injury than the previous one. There is no reason to believe that this progression has reached an asymptote. Even our most sensitive methods may still underestimate the extent and severity of brain injury following head trauma. Ultimately, postmortem pathologic analysis remains the “gold standard” for the extent and severity of injury. However, even traditional pathologic methods are rapidly evolving in sensitivity. For example, immunohistochemical methods (e.g., amyloid
1.0
273
r2 = 0.8133 p < 0.0001
0.8
00 25
00
00 20
00
00 15
00
00 00 10
50
0
0.4
0
Control 1.0 mm 1.5 mm 2.5 mm
0.6
00
Normalized relative anisotropy
be especially vulnerable to blast (Taylor and Ford, 2009). Follow-up scans performed 6–12 months later in 47 TBI subjects demonstrated persistent DTI abnormalities consistent with evolving injuries. Thus, DTI findings in US military personnel supported the hypothesis that blast-related mild TBI can involve axonal injury. However, many TBI subjects did not have DTI abnormalities, though all had a clinical diagnosis of TBI. Subsequently, we enrolled four US military subjects with isolated primary blast TBI and no other mechanism of injury, and without any lifetime exposure to other TBI. We found that three of these subjects had reduced anisotropy in the middle cerebellar peduncle, but there were no other abnormalities detected despite an extensive evaluation (Mac Donald et al., 2013). These findings add to the evidence supporting the hypothesis that primary blast exposure contributes to brain injury in the absence of head impact and that the cerebellum may be particularly vulnerable.
APP stained axonal varicosities / mm3
Fig. 17.8. Diffusion tensor imaging in a mouse model of pericontusional traumatic axonal injury. Mice were injured with controlled cortical impact at three different severities (1.0 mm, 1.5 mm, or 2.5 mm impact depth), scanned with DTI 24 hours later, and then sacrificed for quantitative histologic assessment of axonal injury using stereological counting of amyloid precursor protein (APP)-stained axonal varicosities. (CL Mac Donald and DL Brody from unpublished data.) Methods were otherwise as previously described (Mac Donald et al., 2007a, b).
precursor protein (APP), neurofilament) have supplanted traditional histology in the assessment of axonal injury, and there is some indication from experimental animals that newer methods of silver staining may be more sensitive still than immunohistochemical methods. Electron microscopy may be the most sensitive method, but clearly it is too labor intensive for most applications. In our view, an important guiding principle for future development of imaging methods with greater and greater sensitivity will be that validation of the imaging results should based on direct comparisons with a pathologic “gold standard.” We have taken this approach in a mouse model of pericontusional traumatic axonal injury (Mac Donald et al., 2007a, b). We found that the extent of reduction in the diffusion tensor imaging parameter relative anisotropy directly correlated with the number of APPstained axonal varicosities (Fig. 17.8). Interestingly, and in line with the discussion above, the least severe contusional injuries (1.0 mm impact depth) did not result in APP-stained axonal varicosities, but did cause reduction in relative anisotropy. Likewise, repetitive concussive injury in a mouse model causes extensive white matter silver staining abnormalities, electron microscopic evidence for axonal injury, and behavioral impairments, but no APP-stained axonal varicosities (Shitaka et al., 2011). Interestingly, DTI was abnormal in these white matter regions (Bennett et al., 2012).
Table 17.1 Role of specific modalities of brain imaging in the diagnosis and management of traumatic brain injury Advantages
Disadvantages
Typical use in TBI
Potential future use in TBI
Readily available 24/7 in the ER fast Detect most immediately lifethreatening traumatic lesions
Insensitive to axonal injury and early infarction Poor visualization of brainstem
Initial assessment Acute surgical decision making Evolution of intracranial lesion and hydrocephalus Assessment of vascular injury
Automated lesion volume measurements for prognostic evaluation
Good gray-white tissue contrast Good visualization of brainstem
Anatomical visualization of brainstem, especially cisternal effacement Visualization of arterial dissection
Automated volumetric analysis of regional and global atrophy for prognostic and pharmacodynamic evaluations
T2-weighted MRI
Sensitive to edema
FLAIR MRI
Sensitive to edema
Diffusion-weighted MRI T2* blood-sensitive MRI Susceptibility-weighted MRI Magnetic resonance spectroscopy
Sensitive to early infarction Sensitive to microhemorrhage Fast Even more sensitive to microhemorrhage Sensitive to selected aspects of brain chemistry
Detection of small contusions and some axonal injury Detection of small contusions and some axonal injury Detection of early infarction Assessment of microhemorrhages for medicolegal purposes Assessment of microhemorrhages for medicolegal purposes Research only
Diffusion tensor MRI
Sensitive to axonal injury
High angular resolution diffusion MRI
May be even more sensitive to axonal injury, especially at crossing fibers and close to gray-white junction Potential use in tractography
Noncontrast CT
CT with contrast, including CT angiography T1-weighted MRI
Anatomical distortions Slower to acquire Slow to acquire Poor spatial resolution Extensive analysis required Relatively poor spatial resolution Insensitive to injury at crossing fibers and close to gray-white junction Extensive analysis required Relatively poor spatial resolution Slow acquisition Extensive analysis required
Research only
Research only
TBI, traumatic brain injury; CT, computed tomography; FLAIR, fluid-attenuated inversion recovery; ER, emergency room; MRI, magnetic resonance imaging.
Prognostic planning, especially in children Assessment of injured large white matter tracts for prognostic and pharmacodynamic evaluations Assessment of large and moderate sized injured white matter for prognostic and pharmacodynamic evaluations Assessment of injured white matter for prognostic and pharmacodynamic evaluations
CURRENT AND FUTURE DIAGNOSTIC TOOLS FOR TRAUMATIC BRAIN INJURY
CONCLUSIONS Clearly, brain imaging plays a major role in the diagnosis and management of traumatic brain injury (Table 17.1). This role is evolving rapidly as new imaging methods are developed and the focus turns towards the less severe injuries which make up the vast majority of cases. In our view, rigorous preclinical and clinical validation of each imaging method should be a top priority.
REFERENCES Annegers JF, Hauser WA, Coan SP et al. (1998). A populationbased study of seizures after traumatic brain injuries. N Engl J Med 338: 20–24. Arfanakis K, Haughton VM, Carew JD et al. (2002). Diffusion tensor MR imaging in diffuse axonal injury. AJNR Am J Neuroradiol 23: 794–802. Armonda RA, Bell RS, Vo AH et al. (2006). Wartime traumatic cerebral vasospasm: recent review of combat casualties. Neurosurgery 59: 1215–1225, discussion 1225. Bennett RE, Mac Donald CL, Brody DL (2012). Diffusion tensor imaging detects axonal injury in a mouse model of repetitive closed-skull traumatic brain injury. Neurosci Lett 513: 160–165. Blumbergs PC, Scott G, Manavis J et al. (1994). Staining of amyloid precursor protein to study axonal damage in mild head injury. Lancet 344: 1055–1056. Duhaime AC, Gean AD, Haacke EM et al. (2010). Common data elements in radiologic imaging of traumatic brain injury. Arch Phys Med Rehabil 91: 1661–1666. Gean AD (1994). Imaging of Head Trauma. Lippincott Williams and Wilkins, Philadelphia. Gennarelli TA, Thibault LE, Adams JH et al. (1982). Diffuse axonal injury and traumatic coma in the primate. Ann Neurol 12: 564–574. Gilliam F, Bowling S, Bilir E et al. (1997). Association of combined MRI, interictal EEG, and ictal EEG results with outcome and pathology after temporal lobectomy. Epilepsia 38: 1315–1320. Gurdjian ES, Gurdjian ES (1976). Cerebral contusions: re-evaluation of the mechanism of their development. J Trauma 16: 35–51. Haacke EM, Duhaime AC, Gean AD et al. (2010). Common data elements in radiologic imaging of traumatic brain injury. J Magn Reson Imaging 32: 516–543. Henry TR, Roman DD (2011). Presurgical epilepsy localization with interictal cerebral dysfunction. Epilepsy Behav 20: 194–208. Hogan RE, Lowe VJ, Bucholz RD (1999). Triple-technique (MR imaging, single-photon emission CT, and CT) coregistration for image-guided surgical evaluation of patients with intractable epilepsy. AJNR Am J Neuroradiol 20: 1054–1058. Holbourn A (1943). Mechanics of head injuries. Lancet 2: 438–441. Jellinger KA (2002). Alzheimer disease and cerebrovascular pathology: an update. J Neural Transm 109: 813–836.
275
Jiang T, Gao Y, Fu Y (2004). Clinical research and surgical treatment of posttraumatic epilepsy. J Huazhong Univ Sci Technolog Med Sci 24: 392–395. Lee H, Wintermark M, Gean AD et al. (2008). Focal lesions in acute mild traumatic brain injury and neurocognitive outcome: CT versus 3T MRI. J Neurotrauma 25: 1049–1056. Lee M, Saver JL, Alger JR et al. (2012). Blood–brain barrier permeability derangements in posterior circulation ischemic stroke: frequency and relation to hemorrhagic transformation. J Neurol Sci 313: 142–146. Mac Donald CL, Dikranian K, Bayly P et al. (2007a). Diffusion tensor imaging reliably detects experimental traumatic axonal injury and indicates approximate time of injury. J Neurosci 27: 11869–11876. Mac Donald CL, Dikranian K, Song SK et al. (2007b). Detection of traumatic axonal injury with diffusion tensor imaging in a mouse model of traumatic brain injury. Exp Neurol 205: 116–131. Mac Donald CL, Johnson AM, Cooper D et al. (2011). Detection of blast-related traumatic brain injury in US military personnel. N Engl J Med 364: 2091–2100. Mac Donald C, Johnson A, Cooper D et al. (2013). Cerebellar white matter abnormalities following primary blast injury in US military personnel. PLoS One 8: e55823. Mathern GW, Babb TL, Vickrey BG et al. (1994). Traumatic compared to nontraumatic clinical-pathologic associations in temporal lobe epilepsy. Epilepsy Res 19: 129–139. Niogi SN, Mukherjee P (2010). Diffusion tensor imaging of mild traumatic brain injury. J Head Trauma Rehabil 25: 241–255. Pierpaoli C, Jezzard P, Basser PJ et al. (1996). Diffusion tensor MR imaging of the human brain. Radiology 201: 637–648. Saatman KE, Duhaime AC, Bullock R et al. (2008). Classification of traumatic brain injury for targeted therapies. J Neurotrauma 25: 719–738. Shitaka Y, Tran HT, Bennett RE et al. (2011). Repetitive closed-skull traumatic brain injury in mice causes persistent multifocal axonal injury and microglial reactivity. J Neuropathol Exp Neurol 70: 551–567. Smith DH, Chen XH, Xu BN et al. (1997). Characterization of diffuse axonal pathology and selective hippocampal damage following inertial brain trauma in the pig. J Neuropathol Exp Neurol 56: 822–834. Taylor PA, Ford CC (2009). Simulation of blast-induced earlytime intracranial wave physics leading to traumatic brain injury. J Biomech Eng 131: 061007. Tong KA, Ashwal S, Holshouser BA et al. (2004). Diffuse axonal injury in children: clinical correlation with hemorrhagic lesions. Ann Neurol 56: 36–50. Tong KA, Ashwal S, Obenaus A et al. (2008). Susceptibilityweighted MR imaging: a review of clinical applications in children. AJNR Am J Neuroradiol 29: 9–17. Viswanathan A, Chabriat H (2006). Cerebral microhemorrhage. Stroke 37: 550–555. Zee C-S, Go JL (2002). Imaging of Head Trauma, WB Saunders, Philadelphia.
Chapter 17
Current and future diagnostic tools for traumatic brain injury: CT, conventional MRI, and diffusion tensor imaging DAVID L. BRODY1*, CHRISTINE L. MAC DONALD1, AND JOSHUA S. SHIMONY2 Department of Neurology, Washington University School of Medicine, St. Louis, MO, USA
1
2
Department of Radiology, Washington University School of Medicine, St. Louis, MO, USA
INTRODUCTION
ASSESSMENT GOAL
The advent of the modern neuroimaging era with the development of the computed tomography (CT) scanner has fundamentally changed the way we think of traumatic brain injury (TBI). Instead of saying “a 24-year-old with a severe head injury due to a car crash” we started saying “a 24-year-old with a 2 cm right subdural hematoma, 15 mm midline shift, and bifrontal contusions due to a car crash.” This had immediate surgical implications. Then, with magnetic resonance imaging (MRI), instead of “she fell and hit her head on an icy sidewalk, but her CT was OK”, it has become “she fell and hit her head on an icy sidewalk and she has small bilateral temporal and right frontal hemorrhagic contusions on MRI.” While this does not have the same surgical implications, it certainly does provide objective evidence for a brain injury, which can assist with triage and rehabilitative planning. In the near future, instead of saying “concussion during a football game Friday night, clean MRI” we may say “concussion during a football game Friday night, reduced anisotropy in left anterior corona radiata and uncinate fasciculus on DTI.” This may help inform return-to-work and return-to-play decisions, along with many other implications. This review is written from two perspectives: that of a practitioner ordering the scans (DLB) and that of a radiologist reading and interpreting them (JSS). The review is divided up into assessments based on specific goals, rather than based on scan modalities. It is not meant to supplant but rather to supplement the authoritative previous works on this topic (Gean, 1994; Zee and Go, 2002). We also suggest a reading of the recommendations of the Common Data Elements Neuroimaging Working Group (Duhaime et al., 2010; Haacke et al., 2010).
Determine whether there is an immediately life-threatening intracranial lesion Clearly, a noncontrast head CT scan is the test of choice to address this issue, and such scans are virtually ubiquitously performed on all but the most trivially headinjured patients. Noncontrast head CT scans accurately resolve subdural, epidural, intraparenchymal, and intraventricular hemorrhages (Fig. 17.1). They reveal most types of clinically important herniation, most skull fractures, hydrocephalus, pneumocephalus, and most foreign bodies. Only in rare cases such as brainstem infarction due to traumatic arterial dissection is an immediately life-threatening traumatic lesion not revealed by noncontrast head CT. In our view, there is no need to replace head CT with any other modality at present. Only in rare cases, such as young children or pregnant women, where radiation exposure is a concern could MRI be considered to address the question of whether there is an immediately life-threatening intracranial lesion.
Assess for delayed progression of an intracranial lesion in stabilized patients Many intracranial lesions are not immediately lifethreatening, but have the potential to progress over time. Examples include delayed worsening of contusions, hemorrhages, cerebral edema, and hydrocephalus (Fig. 17.2). Again, noncontrast head CT scans are accurate methods for assessing progression of most of these lesions. Delayed vasospasm may occur after traumatic brain injury, perhaps especially following severe blast-related injuries (Armonda et al., 2006). CT angiography, MR
*Correspondence to: David L. Brody, Department of Neurology, Washington University School of Medicine, 660 South Euclid, Campus Box 8111, St. Louis, MO 63110, USA. Tel: +1-314-362-1381, E-mail: [email protected]
268
D.L. BRODY ET AL.
Fig. 17.1. Acute traumatic brain injury-related lesions, visualized on noncontrast head computed tomography (CT): (A) subdural hematoma; (B) epidural hematoma; (C) hemorrhagic contusion; (D) subarachnoid hemorrhage; (E) skull fracture and foreign body (bullet fragment); (F) pneumocephalus; (G) herniation (due to hemorrhage); (H) contusion and intraventricular hemorrhage with hydrocephalus.
Fig. 17.2. Subacute evolution of traumatic brain injury-related lesions, visualized on noncontrast head CT: (A) early trace hemorrhage; (B) later contusions “blossoming” in the same patient as in panel (A) rescanned 24 hours later; (C) early dilation of temporal horns heralding hydrocephalus; (D) diffuse edema with effacement of sulci.
angiography, and conventional digital subtraction catheter angiography (Fig. 17.3) are sensitive to vasospasm to varying degrees. Early intervention could potentially prevent the development of irreversible cerebral infarction, though this has not been definitively established.
Documentation of the presence of intracranial abnormalities for medicolegal reasons After a motor vehicle accident, assault, or work-related injury, the burden of proof is often on the injured person to document that there was in fact a brain injury. Signs and symptoms especially of mild or concussive injuries are sufficiently nonspecific that they often cannot be used in isolation to make a definitive diagnosis. In this setting, MRI is clearly superior to noncontrast head CT. Conventional MRI sequences (T1, T2, fluidattenuated inversion recovery (FLAIR)) without contrast reveal small contusions and hemorrhages that are missed
by CT (Fig. 17.4). In part this is due to the superior signalto-noise and spatial resolution, and in part due to fewer artifacts at interfaces between bone and brain. MRI is not immune from these artifacts, and even conventional MRI sequences are continuously improving. Modern blood-sensitive sequences such as high field gradientrecalled echo (GRE) and susceptibility-weighted imaging (SWI) have even greater sensitivity to small hemorrhages (Tong et al., 2004, 2008; Lee et al., 2008). A concern arises in that MRI may show “abnormalities” that are not related to traumatic brain injury. Examples include periventricular or small subcortical white matter T2 and FLAIR signal abnormalities which are common in the general population, and especially prevalent with increasing age and in those with cerebrovascular risk factors such as hypertension and diabetes. Also, cerebral amyloid angiopathy is a common cause of cerebral microhemorrhages unrelated to trauma (Fig. 17.4C). This vascular pathology is present in approximately 90% of subjects with Alzheimer’s
CURRENT AND FUTURE DIAGNOSTIC TOOLS FOR TRAUMATIC BRAIN INJURY
269
Fig. 17.3. Vasospasm following blast-related traumatic brain injury, visualized with digital subtraction catheter angiography. Narrowing of the right distal internal carotid artery (black arrows in a and b) and normalization of the arterial diameter after balloon angioplasty (white arrows in c and d). (Reproduced from Armonda et al., 2006, Figure 7B.)
disease pathology and common in others with cerebrovascular risk factors (Jellinger, 2002; Viswanathan and Chabriat, 2006). Ultimately, some subjective judgment is required based on the nature of the injury and the most common distributions of traumatic versus nontraumatic abnormalities in order to make a decision regarding the etiology of the signal changes. For example, in an extensive autopsy series, contusions were found most commonly in the frontal and temporal poles (Gurdjian and Gurdjian, 1976). These regions may be most vulnerable to mechanical forces due to rotational acceleration/ deceleration of the brain (Holbourn, 1943; Gennarelli et al., 1982). Limitations of MRI include longer scan times, reduced availability in the acute setting, contraindications due to metal, and claustrophobia.
Presurgical epilepsy planning Traumatic brain injury, especially severe closed skull injury and penetrating injury, has been well documented to cause epilepsy (Annegers et al., 1998). The posttraumatic seizures can often be controlled with standard antiepileptic medications, but when they are refractory,
surgical resections are frequently considered (Mathern et al., 1994; Jiang et al., 2004). In this setting, MRI is a key part of a multidisciplinary assessment of lesion identification and localization for presurgical planning (Fig. 17.5). Other methods such as ictal single-photon emission computed tomography (SPECT) or positron emission tomography (PET) scans are sometimes used in this context as well (Gilliam et al., 1997; Hogan et al., 1999; Henry and Roman, 2011). As stated above, retained metal objects (e.g., shrapnel) in the brain are a contraindication to MRI. A unilateral hippocampal abnormality apparent on MRI on the same side as interictal electroencephalogram (EEG) abnormalities is a strong predictor of seizure freedom after surgical resection (Gilliam et al., 1997). Again, conventional MRI sequences are rapidly improving in resolution, sensitivity, and signal-to-noise, allowing more and more subtle intracranial abnormalities to be localized (see Ch. 33).
Early prognostic decision making In an acute setting, prognostic decision making regarding withdrawal of care, organ donation, etc. following
270
D.L. BRODY ET AL.
Fig. 17.4. Acute traumatic brain injury-related lesions, visualized on structural MRI: (A) fluid-attenuated inversion recovery (FLAIR) image showing a small subdural hemorrhage and posterior corpus callosum abnormality (neither of which were apparent on CT); (B) susceptibility-weighted imaging (SWI) demonstrating multiple microhemorrhages (none of which were apparent on CT); (C) SWI in a patient with amyloid angiopathy with scattered microhemorrhages not related to traumatic brain injury; (D) FLAIR image demonstrating frontal contusions and small subdural hemorrhages; (E) FLAIR imagine demonstrating a small cortical contusion and subdural hemorrhage.
Fig. 17.5. Mesial temporal sclerosis following traumatic brain injury in a patient with refractory post-traumatic epilepsy. (A) Fluid-attenuated inversion recovery (FLAIR) approximately 1 year after a work-related traumatic brain injury, initially read as normal but in retrospect showing mild left hippocampal atrophy (red arrow). The patient’s “spells” were considered to be psychiatric in origin, or possibly motivated by secondary gain. (B) FLAIR repeated 4 years after injury, demonstrating progression of hippocampal atrophy and elevated signal consistent with mesial temporal sclerosis. At that time, video EEG demonstrated left temporal onset complex partial seizures. (C) T1-weighted MRI performed for presurgical planning. (D) Computed tomography (CT) scan after anterior amygdalohippocampectomy. This resulted in a near complete elimination of seizures, but some worsening of verbal memory function.
traumatic brain injuries is usually based on clinical criteria. CT scans showing devastating intracranial pathology such as severe bilateral cortical or thalamic damage can support these decisions. In rare cases where the clinical presentation is far worse than can be explained based on the CT scan, MRI may be helpful. For example, conventional MRI may reveal a brainstem or thalamic
infarction using diffusion-weighted imaging. MR angiography, CT angiography, or digital subtraction catheter angiography may reveal the source of the infarction, such as a traumatic vertebral dissection or basilar occlusion (Fig. 17.6). All three angiographic methods may be used confirm the absence of blood flow to the brain in the case of massively elevated intracranial pressure.
CURRENT AND FUTURE DIAGNOSTIC TOOLS FOR TRAUMATIC BRAIN INJURY
271
Fig. 17.6. Initially unexplained causes of impaired arousal following traumatic brain injury: (A) brainstem ischemia, apparent as a subtle pontine hypodensity on computed tomography (CT); (B) thalamic infarction, apparent as elevated signal on fluid-attenuated inversion recovery (FLAIR); (C) carotid dissection, apparent as intravascular low signal on magnetic resonance imaging (MRI); (D) carotid dissection following traumatic brain injury on digital subtraction angiography. Lesions indicated by red arrows.
Conversely, a profoundly comatose TBI patient with little apparent pathology on CT and conventional MRI may be considered a reasonable candidate for aggressive medical intervention, despite poor clinical status. The absence of major irreversible intracranial injury on a conventional MRI including diffusion-weighted imaging can indicate a reasonable chance for a good outcome. This can be reassuring, as early clinical signs and symptoms alone have limited prognostic power in the period soon after injury. This is commonly performed based mainly on clinical judgment; to our knowledge there have not been large prospective trials of the role of brain imaging for early prognostic decision making after severe TBI.
Prognostic assessment for rehabilitative planning Here we move out of the realm of what is currently standard practice, and into the domain of future applications. Nonetheless, a few motivating examples may help. It is very likely that a TBI patient with substantial bilateral medial temporal lobe contusions will have clinically important declarative memory deficits. Therefore, compensatory rehabilitation focused on automatic skill learning would be advisable. If the thalamus is completely ablated, it is unlikely that the patient will emerge from coma spontaneously and aggressive interventions may be warranted. Family, friends, and medical practitioners should be prepared for emotional dysregulation and personality changes in patients with large bifrontal injuries. It has not been established whether imaging-assisted rehabilitation planning is superior to rehabilitation based purely on observed clinical deficits and symptoms. Clearly, MRI is superior to CT for resolution of the exact extent and localization of injuries in the subacute setting. Newer modalities such as diffusion tensor imaging (see below) have greater predictive and correlative prognostic power for specific outcomes than other imaging methods. However, these methods
are not yet available for routine clinical use at present (see Chs 27 and 40).
Stratification of subjects for pathophysiologically targeted therapies A logical approach to the development of new therapeutics for traumatic brain injury would be to test them in patients with clear evidence for the pathophysiologic processes targeted by the candidate therapeutics (Saatman et al., 2008): medications designed to counteract coagulopathy should be tested in patients with clinically significant hemorrhage; edema reduction should be tested in patients with edema; axonal protectants should be tested in patients with axonal injury; oligodendroglial apoptosis inhibitors should be tested in patients with myelin injury, etc. This approach has not been taken historically, and this may account in part for the lack of success of the >30 major clinical trials that have been performed in TBI to date. A major challenge will be to develop imaging methods capable of resolving the relevant pathophysiologic processes with a clinically meaningful time window. Many processes will require intervention within a few hours, whereas MRI, PET, and other advanced imaging methods are typically not available late at night and in the weekend hours when TBIs often occur. Infrastructure changes, e.g., putting MRI scanners in emergency departments, or having at least one MRI scanner always available may be required to overcome this challenge. Significant advances along these lines in ischemic stroke research have been made recently (Lee et al., 2012). Of particular interest, traumatic axonal injury appears to be a nearly universal pathologic sequela of TBI. In one of the few published autopsy series of mild/concussive TBI cases (with deaths from other causes) traumatic axonal injury was present in 6/6 cases (Blumbergs et al., 1994). In contrast, hypoxic/ischemic damage, contusion, intraparenchymal hematoma, subdural hemorrhage, signs of elevated intracranial pressure, and skull fracture were absent. Likewise,
272
D.L. BRODY ET AL.
rotation injury in monkeys (Gennarelli et al., 1982) and pigs (Smith et al., 1997) resulted in axonal injury with little other pathology. This may indicate that the axons are the most mechanically vulnerable part of the brain. However, CT and conventional MRI sequences appear to substantially underestimate the extent of axonal injury. Stated another way, there can be extensive axonal injury even when CT and conventional MRI are normal. For this reason, the development of newer imaging methods sensitive to axonal injury has been a top priority for traumatic brain injury researchers. Other methods are discussed elsewhere in this volume, but diffusion tensor imaging (DTI) (Fig. 17.7) may be the most well established of these newer imaging methods (Niogi and Mukherjee, 2010; Mac Donald et al., 2011). DTI involves measurement of water diffusion in multiple directions and can be performed quickly on most clinical scanners (Pierpaoli et al., 1996). In brain white matter, water diffuses faster along the predominant fiber direction and more slowly in perpendicular directions. The resulting anisotropy (directional asymmetry) of water diffusion is high in intact axons and reduced after axonal injury (Arfanakis et al., 2002; Mac Donald et al., 2007a, b). Reduced anisotropy on DTI as a marker of traumatic axonal injury has been directly validated versus immunohistochemistry in experimental TBI, even when
conventional MRI is normal (Mac Donald et al., 2007a, b). Open questions include the extent to which injury to myelinated versus unmyelinated axons can be detected with DTI. We used DTI to test the hypothesis that blast-related TBI causes traumatic axonal injury in specific brain regions distinct from those typically injured in other forms of TBI (Mac Donald et al., 2011). Participants were 63 US military personnel evacuated to Landstuhl Regional Medical Center, clinically diagnosed with mild uncomplicated TBI, and scanned 1–90 days after injury. All had primary blast exposure plus another blast-related mechanism of injury (e.g., struck by a blunt object, fall, motor vehicle crash). Controls were 21 similar personnel with blast exposure and other injuries but no clinical diagnosis of TBI. DTI revealed abnormalities consistent with traumatic axonal injury in many TBI subjects. None had detectable intracranial injury on CT and only one had abnormalities on conventional MRI. By contrast, in 18/63 individual TBI subjects, there were significantly more DTI abnormalities than expected by chance ( p < 0.001). DTI was most markedly abnormal in the middle cerebellar peduncles (p ¼ < 0.001), cingulum bundles ( p ¼ 0.002), and right orbitofrontal white matter ( p ¼ 0.007). The middle cerebellar peduncle is rarely affected in conventional mild TBI, but predicted to
Diffusion Tensor Imaging
λ1
2. Calculate diffusion tensor for each voxel
λ2 λ3
A
Brain white matter: organized, myelinated axons
3. Separate parallel(λ1, axial) from perpendicular (λ2, λ3, radial) diffusion. Calculate anisotropy.
1. Collect diffusion weighted images in six or more directions
Traumatic axonal injury: simplified model
λ1
λ2
λ1
λ2
λ3
B
Axonal Disruption: reduced λ1 (axial), reduced anisotropy
λ3 Myelin Injury: incr. λ2, λ3 (radial), reduced anisotropy
λ1
λ2 λ3
Mixed Injury: greatly reduced anisotrophy
Fig. 17.7. Diffusion tensor imaging: theoretical basis for sensitivity to axonal injury. (A) Diffusion tensor imaging in normal white matter. (B) Predicted effects of axonal and myelin injury. (Adapted from M Budde.)
CURRENT AND FUTURE DIAGNOSTIC TOOLS FOR TRAUMATIC BRAIN INJURY
Providing pharmacodynamic biomarkers for candidate therapeutics If the challenges discussed above can be overcome, the same imaging methods could be used to assess the pharmacodynamic effects of the candidate therapeutics. For example, a relatively small number of patients could be sufficient to demonstrate that a candidate axonal protectant improves DTI signal abnormalities. Then, the question of whether prevention of progression of apparent axonal injury as indicated by DTI has a clinically meaningful benefit could be addressed in larger phase III clinical trial. If, on the other hand, a candidate therapeutic does not “hit the target” at the clinically achievable doses or administration times following injury, it may not be a good use of resources to proceed to a full-scale clinical trial. This strategy has not yet been implemented to our knowledge.
FUTURE DIRECTIONS From skull X-rays, to CT, to conventional MRI, to newer MRI methods such as DTI and GRE and SWI, each generation of imaging approaches demonstrates more abnormalities following traumatic brain injury than the previous one. There is no reason to believe that this progression has reached an asymptote. Even our most sensitive methods may still underestimate the extent and severity of brain injury following head trauma. Ultimately, postmortem pathologic analysis remains the “gold standard” for the extent and severity of injury. However, even traditional pathologic methods are rapidly evolving in sensitivity. For example, immunohistochemical methods (e.g., amyloid
1.0
273
r2 = 0.8133 p < 0.0001
0.8
00 25
00
00 20
00
00 15
00
00 00 10
50
0
0.4
0
Control 1.0 mm 1.5 mm 2.5 mm
0.6
00
Normalized relative anisotropy
be especially vulnerable to blast (Taylor and Ford, 2009). Follow-up scans performed 6–12 months later in 47 TBI subjects demonstrated persistent DTI abnormalities consistent with evolving injuries. Thus, DTI findings in US military personnel supported the hypothesis that blast-related mild TBI can involve axonal injury. However, many TBI subjects did not have DTI abnormalities, though all had a clinical diagnosis of TBI. Subsequently, we enrolled four US military subjects with isolated primary blast TBI and no other mechanism of injury, and without any lifetime exposure to other TBI. We found that three of these subjects had reduced anisotropy in the middle cerebellar peduncle, but there were no other abnormalities detected despite an extensive evaluation (Mac Donald et al., 2013). These findings add to the evidence supporting the hypothesis that primary blast exposure contributes to brain injury in the absence of head impact and that the cerebellum may be particularly vulnerable.
APP stained axonal varicosities / mm3
Fig. 17.8. Diffusion tensor imaging in a mouse model of pericontusional traumatic axonal injury. Mice were injured with controlled cortical impact at three different severities (1.0 mm, 1.5 mm, or 2.5 mm impact depth), scanned with DTI 24 hours later, and then sacrificed for quantitative histologic assessment of axonal injury using stereological counting of amyloid precursor protein (APP)-stained axonal varicosities. (CL Mac Donald and DL Brody from unpublished data.) Methods were otherwise as previously described (Mac Donald et al., 2007a, b).
precursor protein (APP), neurofilament) have supplanted traditional histology in the assessment of axonal injury, and there is some indication from experimental animals that newer methods of silver staining may be more sensitive still than immunohistochemical methods. Electron microscopy may be the most sensitive method, but clearly it is too labor intensive for most applications. In our view, an important guiding principle for future development of imaging methods with greater and greater sensitivity will be that validation of the imaging results should based on direct comparisons with a pathologic “gold standard.” We have taken this approach in a mouse model of pericontusional traumatic axonal injury (Mac Donald et al., 2007a, b). We found that the extent of reduction in the diffusion tensor imaging parameter relative anisotropy directly correlated with the number of APPstained axonal varicosities (Fig. 17.8). Interestingly, and in line with the discussion above, the least severe contusional injuries (1.0 mm impact depth) did not result in APP-stained axonal varicosities, but did cause reduction in relative anisotropy. Likewise, repetitive concussive injury in a mouse model causes extensive white matter silver staining abnormalities, electron microscopic evidence for axonal injury, and behavioral impairments, but no APP-stained axonal varicosities (Shitaka et al., 2011). Interestingly, DTI was abnormal in these white matter regions (Bennett et al., 2012).
Table 17.1 Role of specific modalities of brain imaging in the diagnosis and management of traumatic brain injury Advantages
Disadvantages
Typical use in TBI
Potential future use in TBI
Readily available 24/7 in the ER fast Detect most immediately lifethreatening traumatic lesions
Insensitive to axonal injury and early infarction Poor visualization of brainstem
Initial assessment Acute surgical decision making Evolution of intracranial lesion and hydrocephalus Assessment of vascular injury
Automated lesion volume measurements for prognostic evaluation
Good gray-white tissue contrast Good visualization of brainstem
Anatomical visualization of brainstem, especially cisternal effacement Visualization of arterial dissection
Automated volumetric analysis of regional and global atrophy for prognostic and pharmacodynamic evaluations
T2-weighted MRI
Sensitive to edema
FLAIR MRI
Sensitive to edema
Diffusion-weighted MRI T2* blood-sensitive MRI Susceptibility-weighted MRI Magnetic resonance spectroscopy
Sensitive to early infarction Sensitive to microhemorrhage Fast Even more sensitive to microhemorrhage Sensitive to selected aspects of brain chemistry
Detection of small contusions and some axonal injury Detection of small contusions and some axonal injury Detection of early infarction Assessment of microhemorrhages for medicolegal purposes Assessment of microhemorrhages for medicolegal purposes Research only
Diffusion tensor MRI
Sensitive to axonal injury
High angular resolution diffusion MRI
May be even more sensitive to axonal injury, especially at crossing fibers and close to gray-white junction Potential use in tractography
Noncontrast CT
CT with contrast, including CT angiography T1-weighted MRI
Anatomical distortions Slower to acquire Slow to acquire Poor spatial resolution Extensive analysis required Relatively poor spatial resolution Insensitive to injury at crossing fibers and close to gray-white junction Extensive analysis required Relatively poor spatial resolution Slow acquisition Extensive analysis required
Research only
Research only
TBI, traumatic brain injury; CT, computed tomography; FLAIR, fluid-attenuated inversion recovery; ER, emergency room; MRI, magnetic resonance imaging.
Prognostic planning, especially in children Assessment of injured large white matter tracts for prognostic and pharmacodynamic evaluations Assessment of large and moderate sized injured white matter for prognostic and pharmacodynamic evaluations Assessment of injured white matter for prognostic and pharmacodynamic evaluations
CURRENT AND FUTURE DIAGNOSTIC TOOLS FOR TRAUMATIC BRAIN INJURY
CONCLUSIONS Clearly, brain imaging plays a major role in the diagnosis and management of traumatic brain injury (Table 17.1). This role is evolving rapidly as new imaging methods are developed and the focus turns towards the less severe injuries which make up the vast majority of cases. In our view, rigorous preclinical and clinical validation of each imaging method should be a top priority.
REFERENCES Annegers JF, Hauser WA, Coan SP et al. (1998). A populationbased study of seizures after traumatic brain injuries. N Engl J Med 338: 20–24. Arfanakis K, Haughton VM, Carew JD et al. (2002). Diffusion tensor MR imaging in diffuse axonal injury. AJNR Am J Neuroradiol 23: 794–802. Armonda RA, Bell RS, Vo AH et al. (2006). Wartime traumatic cerebral vasospasm: recent review of combat casualties. Neurosurgery 59: 1215–1225, discussion 1225. Bennett RE, Mac Donald CL, Brody DL (2012). Diffusion tensor imaging detects axonal injury in a mouse model of repetitive closed-skull traumatic brain injury. Neurosci Lett 513: 160–165. Blumbergs PC, Scott G, Manavis J et al. (1994). Staining of amyloid precursor protein to study axonal damage in mild head injury. Lancet 344: 1055–1056. Duhaime AC, Gean AD, Haacke EM et al. (2010). Common data elements in radiologic imaging of traumatic brain injury. Arch Phys Med Rehabil 91: 1661–1666. Gean AD (1994). Imaging of Head Trauma. Lippincott Williams and Wilkins, Philadelphia. Gennarelli TA, Thibault LE, Adams JH et al. (1982). Diffuse axonal injury and traumatic coma in the primate. Ann Neurol 12: 564–574. Gilliam F, Bowling S, Bilir E et al. (1997). Association of combined MRI, interictal EEG, and ictal EEG results with outcome and pathology after temporal lobectomy. Epilepsia 38: 1315–1320. Gurdjian ES, Gurdjian ES (1976). Cerebral contusions: re-evaluation of the mechanism of their development. J Trauma 16: 35–51. Haacke EM, Duhaime AC, Gean AD et al. (2010). Common data elements in radiologic imaging of traumatic brain injury. J Magn Reson Imaging 32: 516–543. Henry TR, Roman DD (2011). Presurgical epilepsy localization with interictal cerebral dysfunction. Epilepsy Behav 20: 194–208. Hogan RE, Lowe VJ, Bucholz RD (1999). Triple-technique (MR imaging, single-photon emission CT, and CT) coregistration for image-guided surgical evaluation of patients with intractable epilepsy. AJNR Am J Neuroradiol 20: 1054–1058. Holbourn A (1943). Mechanics of head injuries. Lancet 2: 438–441. Jellinger KA (2002). Alzheimer disease and cerebrovascular pathology: an update. J Neural Transm 109: 813–836.
275
Jiang T, Gao Y, Fu Y (2004). Clinical research and surgical treatment of posttraumatic epilepsy. J Huazhong Univ Sci Technolog Med Sci 24: 392–395. Lee H, Wintermark M, Gean AD et al. (2008). Focal lesions in acute mild traumatic brain injury and neurocognitive outcome: CT versus 3T MRI. J Neurotrauma 25: 1049–1056. Lee M, Saver JL, Alger JR et al. (2012). Blood–brain barrier permeability derangements in posterior circulation ischemic stroke: frequency and relation to hemorrhagic transformation. J Neurol Sci 313: 142–146. Mac Donald CL, Dikranian K, Bayly P et al. (2007a). Diffusion tensor imaging reliably detects experimental traumatic axonal injury and indicates approximate time of injury. J Neurosci 27: 11869–11876. Mac Donald CL, Dikranian K, Song SK et al. (2007b). Detection of traumatic axonal injury with diffusion tensor imaging in a mouse model of traumatic brain injury. Exp Neurol 205: 116–131. Mac Donald CL, Johnson AM, Cooper D et al. (2011). Detection of blast-related traumatic brain injury in US military personnel. N Engl J Med 364: 2091–2100. Mac Donald C, Johnson A, Cooper D et al. (2013). Cerebellar white matter abnormalities following primary blast injury in US military personnel. PLoS One 8: e55823. Mathern GW, Babb TL, Vickrey BG et al. (1994). Traumatic compared to nontraumatic clinical-pathologic associations in temporal lobe epilepsy. Epilepsy Res 19: 129–139. Niogi SN, Mukherjee P (2010). Diffusion tensor imaging of mild traumatic brain injury. J Head Trauma Rehabil 25: 241–255. Pierpaoli C, Jezzard P, Basser PJ et al. (1996). Diffusion tensor MR imaging of the human brain. Radiology 201: 637–648. Saatman KE, Duhaime AC, Bullock R et al. (2008). Classification of traumatic brain injury for targeted therapies. J Neurotrauma 25: 719–738. Shitaka Y, Tran HT, Bennett RE et al. (2011). Repetitive closed-skull traumatic brain injury in mice causes persistent multifocal axonal injury and microglial reactivity. J Neuropathol Exp Neurol 70: 551–567. Smith DH, Chen XH, Xu BN et al. (1997). Characterization of diffuse axonal pathology and selective hippocampal damage following inertial brain trauma in the pig. J Neuropathol Exp Neurol 56: 822–834. Taylor PA, Ford CC (2009). Simulation of blast-induced earlytime intracranial wave physics leading to traumatic brain injury. J Biomech Eng 131: 061007. Tong KA, Ashwal S, Holshouser BA et al. (2004). Diffuse axonal injury in children: clinical correlation with hemorrhagic lesions. Ann Neurol 56: 36–50. Tong KA, Ashwal S, Obenaus A et al. (2008). Susceptibilityweighted MR imaging: a review of clinical applications in children. AJNR Am J Neuroradiol 29: 9–17. Viswanathan A, Chabriat H (2006). Cerebral microhemorrhage. Stroke 37: 550–555. Zee C-S, Go JL (2002). Imaging of Head Trauma, WB Saunders, Philadelphia.
