Evolution of MRI of Brain Injury in Concussion Patients

Saeed Fakhran; Joseph Delic; Lea Alhilali

Disclosures

Future Neurology. 2014;9(5):517-520. 

Mild traumatic brain injury (mTBI), commonly referred to as 'concussion', affects over 1.7 million in the USA annually with costs of nearly US$17 billion.[1] Despite the name, these injuries are by no means mild, with approximately 15% of patients suffering persistent symptoms beyond 3 months.[2] This 'miserable minority' affects a large number of individuals in the prime of life and, until recently, no consistent correlation existed between clinical symptoms and radiological evidence of structural damage to the brain.

Conventional imaging methods have, in fact, had relatively little to contribute in the evaluation of patients with mTBI. Traditionally, only a small minority underwent imaging – usually patients with atypical symptoms, in order to exclude a more serious injury. Occasionally, MRI would demonstrate microhemorrhages, which were felt to be diffuse axonal injury, but of unclear clinical significance. In the majority of cases, however, not even microhemorrhages were seen. Therefore, the sine qua non of mTBI has come to be normal conventional brain imaging, with the causative white matter injury thought to be beyond the resolution of conventional imaging.

However, the advent of diffusion tensor imaging (DTI) has opened a new role for MRI in mTBI. DTI is a novel MRI technique to noninvasively assess the microstructural integrity of white matter. DTI examines the molecular diffusion of water, described by its fractional anisotropy (FA), which measures directional flow of water along white matter tracts.[3] In normal white matter, water diffuses readily along the orientation of axonal fibers. Any change in the microstructure of white matter and coherence of the white matter fiber tract – as may be seen after mTBI – will reduce the directional flow of water, and therefore reduce the FA, making DTI a prime tool in mTBI assessment.[3] Over the past several years, DTI has helped to elucidate imaging correlates for the cognitive impairments that often result from mTBI.

Unlike the microhemorrhages detected by conventional MRI, which remain of unclear clinical significance, abnormalities detected by DTI have been shown to correlate with clinical findings. In fact, when both microhemorrhages and FA abnormalities on DTI were compared with cognitive deficits after mTBI, only DTI findings correlated with abnormalities on cognitive tests. This finding cemented the role of DTI as a valuable tool in post mTBI analysis. Conversely, the lack of correlation of microhemorrhages with patient symptoms has validated the limited role of conventional MRI in the evaluation of mTBI patients.[3]

Despite this promising start, establishing exactly which of the white matter injuries detected by DTI are clinically significant has remained an elusive task. Numerous studies have demonstrated white matter abnormalities on DTI in mTBI patient relative to controls;[4,5] however, these have failed to correlate with symptoms.[2] Other studies have correlated postconcussive memory dysfunction with focal DTI white matter abnormalities, but failed to show a strong anatomic–pathologic correlation, with memory deficits correlating with regions as diverse as the occipital cortex and corticospinal tracts.[6,7]

This inability to obtain a strong structure/function relationship between DTI abnormalities and patient symptoms may be because DTI is very sensitive to white matter injuries, but not very specific. DTI detects any injury to white matter, regardless of etiology or acuity. As a result, vascular injuries cannot be distinguished from traumatic ones. Likewise, acute and chronic injuries cannot be distinguished. This lack of specificity is compounded in the mTBI patient population, where prior concussions are common.

It has, therefore, been difficult to determine which white matter injuries result in postconcussion symptoms from imaging alone. In an attempt to separate acute, symptomatic white matter injuries from background changes, studies have correlated white matter abnormalities on DTI with postconcussive cognitive dysfunction, in the hopes that only acute abnormalities would correlate with cognitive findings.[3,8] However, cognitive dysfunction is a suboptimal end point for evaluating mTBI patients, as it is difficult to assess for cognitive changes without knowing the baseline function prior to trauma. Somewhat predictably, these studies also did not find a strong structure/function relationship, with associated abnormalities seen in regions as disparate as the prefrontal cortex and internal capsule.

As opposed to cognitive dysfunction, patient-reported symptoms are more likely to be the result of the trauma itself. If patients report a new symptom after an injury, this is likely related to the recent event, since patients themselves recognize the symptom as new and acute. As a result, the most recent imaging studies in concussion have correlated injuries not with cognitive dysfunction, but rather new patient-reported symptoms, in order to detect only the acute, symptomatic injuries.

If only white matter lesions that correlate with patient-reported symptom severity are evaluated, abnormalities are only seen at the gray–white junction, primarily in the auditory cortex.[9] This is in contradistinction to when mTBI patients are compared with normal controls, where diffuse white matter abnormalities are seen. This indicates that the majority of the white matter injuries detected by DTI in mTBI patients are likely chronic, and symptomatic injuries are located in the auditory cortex. This correlates well with the known central auditory dysfunction that occurs after mTBI, producing the first strong structure/function relationship between the location of an injury on DTI and patient symptoms.

This symptom-based approach to white matter injuries after concussion not only removes the background of chronic injuries detected by DTI, but is changing the way we view concussions. If we focus on patient symptoms, we find that individual concussion patients present with different symptomatology, usually with an overall dominant symptom cluster.[10] Dominant symptom clusters fall into six categories: sleep–wake disturbances, migraine, anxiety, vestibulopathy, ocular dysfunction and cervicalgia.[10–12] This symptom-based approach asks if there are different injuries producing different symptoms in different patients. This could completely change our understanding of concussion from the notion of a single injury to multiple different injuries, each with their own unique symptoms.

Symptom-based imaging of mTBI with DTI has yielded startling and promising results, demonstrating pathologic–anatomic correlations lacking in prior studies. For example, evaluating mTBI patients with sleep–wake disturbances demonstrated DTI abnormalities in the parahippocampus, a region known to be important in sleep initiation and sleep-associated memory consolidation.[9]

By further clustering patients by symptoms, mTBI patients with vestibulopathy have been found to have decreased FA in the cerebellum and fusiform gyri.[13] The areas of cerebellar abnormality are known to play vitals roles in movement and balance while the fusiform gyri are central in stereoscopic vision and landmark recognition during navigation.[14,15]

Examining mTBI patients with the ocular dysfunction symptom cluster has shown DTI white matter abnormalities in the lateral geniculate nucleus and anterior thalamic radiations.[13] The optic radiations of the lateral geniculate nucleus are the major relay station for the accommodation circuit controlling convergence.[16] The anterior thalamic radiation is the principal tract responsible for processing speed.[17,18]

By determining which specific injured region is associated with a symptom, the injuries underlying mTBI can now be quantified, to determine how injury severity determines recovery. Imaging biomarkers have lagged far behind clinical findings in predicting outcomes after concussion and are not used clinically. However, prior attempts to correlate DTI findings with outcome have used a nonselective approach and ignored individual patient demographics and symptomatology. The lack of a reliable correlation may be because a given white injury may impact outcome in a certain group of patients, but not another. For example, when patients are separated by gender, a strong correlation is seen between poor outcome and injury to the uncinate fasciculus on DTI. The degree of injury to the uncinate fasciculus on DTI was a stronger predictor of recovery than even initial symptom severity.[19] This suggests a role for uncinate fasciculus FA values as a gender-neutral metric for evaluating the severity of mTBI. Similarly, when patients are separated by symptoms, injury to the cerebellum on DTI is found to correlate with prognosis in patients with vestibulopathy,[13] but not those without vestibular symptoms. This indicates knowing whether or not an injury detected by DTI will impact outcome requires knowing the individual patient characteristics and symptomatology.

Symptom-based imaging has advanced our understanding of mTBI from the initial knowledge that mTBI patients demonstrated regions of white matter injury relative to controls, towards an understanding that patients with a given symptom demonstrate a unique white matter injury pattern. This has helped us understand that specific symptoms after mTBI are the result of different injuries rather than different clinical manifestations of the same injury. Given the importance of using an individual's symptoms to determine the meaningful injuries on DTI, a detailed and complete clinical assessment is critical prior to MRI to maximize the potential for locating clinically relevant areas of brain injury in concussion patients. Our understanding of mTBI has evolved with our evolving methods of imaging the underlying injury. Now, with symptom-based DTI, we finally have a concept of mTBI, not as a single pathology, but rather a diverse set of pathologies, with associated unique injuries and symptoms, which share a common traumatic origin.

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....