Abstract
Neuroprognostication following diffuse axonal injury (DAI) has historically relied on neuroimaging techniques with lower spatial resolution and contrast than techniques currently available in clinical practice. Since the initial studies of DAI classification and prognosis in the 1980s and 1990s, advances in neuroimaging have improved detection of brainstem microbleeds, a hallmark feature of Grade 3 DAI that has traditionally been associated with poor neurologic outcome. Here, we report clinical and radiologic data from two patients with severe traumatic brain injury and grade 3 DAI who recovered functional independence and returned to work within 4 months of injury. Importantly, both patients were scanned using 3 Tesla MRI protocols that included susceptibility-weighted imaging (SWI), a technique that provides enhanced sensitivity for detecting brainstem microbleeds. These observations highlight the importance of developing approaches to DAI classification and prognosis that better align with contemporary neuroimaging capabilities.
Introduction
Grade 3 diffuse axonal injury (DAI) has historically been believed to confer an unfavorable prognosis. 1 The initial pathological reports of this condition came from patients who did not survive an acute traumatic brain injury (TBI)2-4 or who developed chronic disorders of consciousness (DoC). 5 Imaging studies using computed tomography (CT) 6 and magnetic resonance imaging (MRI)6,7 similarly suggested an association between Grade 3 DAI and outcomes such as chronic DoC, severe disability, or death. Reports of recovery were rare, 8 particularly recovery of independence and return to work or school. It is against this backdrop that imaging techniques have rapidly advanced, with new MRI sequences and increased spatial resolution providing far more sensitive detection of Grade 3 DAI than their predecessors.9,10 It remains unknown whether patients with Grade 3 DAI diagnosed by today’s imaging techniques should be considered in the same pessimistic prognostic light as patients treated over previous decades.
Here, we report clinical and radiologic data from two young patients with severe TBI and Grade 3 DAI who recovered functional independence and returned to work within 4 months. Importantly, both patients were scanned using 3 Tesla (3T) MRI protocols that included susceptibility-weighted imaging (SWI), a technique that provides enhanced sensitivity for detecting brainstem microbleeds, 9 a distinguishing finding of Grade 3 DAI. 11 These observations add to a growing body of evidence 12 that prior approaches to DAI classification and prognosis should be reevaluated in the setting of improved neuroimaging capabilities, which may detect smaller brainstem lesions that are not necessarily associated with poor outcome.
Case Descriptions
Case 1
A 26-year-old left-handed male engineer sustained a head-first fall down a flight of concrete stairs. Emergency medical services were called and found the patient to have a Glasgow Coma Scale (GCS) score 13 of 4 (E1M2V1). He was hypertensive but oxygenating well and protecting his airway. The patient was transferred to a regional trauma center where his GCS score was 6 (E1M4V1), and he was subsequently intubated.
A head CT revealed multiple frontotemporal hemorrhagic contusions, a small left frontal subarachnoid hemorrhage, a 10 mm left hemispheric subdural hematoma, and a minimally displaced fracture of the right parietal bone extending through the temporal bone. There was left-to-right midline shift of 10 mm at the septum pellucidum. An emergent left hemicraniectomy was performed with placement of an intracranial pressure (ICP) monitor. ICP was controlled at ≤ 20 mmHg, and cerebral perfusion pressure was maintained at 60–70 mmHg for the next two days, but his neurological exam did not improve. On day three post-injury, the patient was airlifted to the Neurosciences Intensive Care Unit (NeuroICU) at a Level I Trauma Center.
Upon arrival to the NeuroICU, the patient’s GCS score was 7T (E2M4V1T). The next day, a 3T brain MRI revealed evolution of the multi-compartmental intracranial hemorrhages with no new sites of hemorrhage or evidence of acute infarction. The SWI sequence revealed punctate hypointensities, indicating traumatic microbleeds, in the inferior longitudinal fasciculi, splenium and body of the corpus callosum, left thalamus, right and left dorsal midbrain (Figure 1), and right dorsal pons. He began following commands on post-injury day 6 and was extubated on day 11. After extubation, he was consistently oriented to self and hospital, and occasionally to month and year. On post-injury day 17, he was discharged to an inpatient rehabilitation hospital where he received two weeks of physical, occupational, and speech therapy before returning home. Brainstem lesion mapping. The brainstem microbleeds are visualized as punctate hypointensities on susceptibility-weighted imaging. A dorsal midbrain microbleed in patient 1 is shown in (A), with a zoomed in view of the microbleed shown in (C), indicated by a red arrow. A ventral midbrain microbleed in patient 2 is shown in (B), with a zoomed view of the microbleed shown in (D), indicated by a red arrow. The respective microbleeds were traced on the Harvard Ascending Arousal Network Atlas,
14
superimposed on the Montreal Neurological Institute (MNI 152) 1 mm T1 template. In patient 1, the microbleed (red) overlaps part of the periaqueductal grey (purple), but not the midbrain reticular formation (blue) or the ventral tegmental area (pink), as shown in (E). Additional microbleeds (not shown) were identified in the right dorsal pons and overlapping a different part of the periaqueductal gray in the right dorsal midbrain. In patient 2, the microbleed does not overlap any of the arousal nuclei located within the midbrain, as shown in (F).
During evaluations in outpatient teleneurology and speech/language therapy clinics over the following month (6–8 weeks post-injury), he did not report any deficits. He was initially taking brief naps during the day due to fatigue, a change from his pre-injury habits. On neurological examination, he was awake, alert and fully oriented. His speech was fluent and non-dysarthric, with normal rate, syntax, grammar, and prosody. Attention, working memory, and visuospatial skills were grossly intact. There were no cranial nerve, motor, sensory, or cerebellar abnormalities. By 8 weeks post-injury, his fatigue resolved, and he returned to work in a limited administrative role. By 16 weeks post-injury, he returned to work as an engineer full time.
Case 2
A 32-year-old right-handed engineer sustained a head-first fall onto ice while snowboarding. He was found down by friends who alerted ski patrol. Ski patrol emergency medical personnel found him unconscious with facial bleeding and a GCS score of 3 (E1M1V1). He was intubated for airway protection and airlifted to a regional trauma center.
Shortly before arriving at the hospital, the patient had an episode of bradycardia, hypertension, and pupillary asymmetry, concerning for Cushing’s reflex. A bolus of hypertonic saline was administered, leading to normalization of his heart rate and pupillary symmetry. On arrival to the emergency department, his GCS score was 3T. Head CT demonstrated multiple facial fractures, a small right temporal contusion, and a punctate hyperdensity at the lateral border of the right putamen, suggesting hemorrhagic DAI.
The patient was admitted to the ICU, and on post-injury day 2 he underwent a 3T brain MRI. In addition to showing the right temporal contusion and putaminal lesion seen on head CT, the SWI sequence revealed numerous punctate hypointensities, indicating microbleeds, in the parasagittal frontal lobes, anterior temporal lobes, genu and splenium of the corpus callosum, right cerebellar hemisphere, and right cerebral peduncle (Figure 1). Several of these lesions were visible as hyperintensities on the T2-weighted fluid-attenuated inversion recovery (FLAIR) and diffusion-weighted imaging (DWI) sequences, including in the splenium of the corpus callosum, periventricular white matter, and subcortical frontal white matter, but not within the brainstem (Figure 2). He was extubated on post-injury day 2, at which time he was awake, alert, and cooperative, but disoriented to date, year, and circumstances. T2-weighted fluid-attenuated inversion recovery (FLAIR), diffusion-weighted imaging (DWI), and susceptibility-weighted imaging (SWI) sequences at the level of the rostral midbrain, where microbleeds were located, in patient 1 (top row) and in patient 2 (bottom row).
On post-injury day 6, he was discharged to an inpatient rehabilitation hospital where he received three weeks of physical, occupational, and speech therapy before returning home. The patient was amnestic to all events surrounding the fall and ICU admission but recalled arriving at the rehabilitation hospital and his subsequent course. Upon returning home, he returned to full functional independence.
During evaluation in outpatient teleneurology clinic the following month, he reported feeling “great,” although he described mild residual fatigue, gradually improving since discharge from the rehabilitation hospital, and decreased libido. On neurological examination, he was awake, alert, and fully oriented. Psychomotor speed was normal. There was some difficulty with serial sevens and delayed recall on first trial. Luria hand sequences were correct. Speech was fluent, non-dysarthric, with normal rate, syntax, grammar, and prosody. He had no difficulties with sentence repetition. Transitive and intransitive praxis were intact. There were no cranial nerve, motor, sensory, or cerebellar abnormalities. He returned to full-time work at his engineering firm 12 weeks following the injury.
Discussion
These two patients who returned to work as engineers within 4 months of severe TBI add to a rapidly growing body of evidence that Grade 3 DAI does not preclude recovery of neurological function. Despite devastating prognoses typically conferred to patients with Grade 3 DAI, it is crucial to recognize that neuroprognostication following DAI has historically been based on neuroimaging techniques with lower spatial resolution and contrast than current 3T SWI techniques available in clinical practice. Similarly, 3T MRI enhances the sensitivity of detecting non-hemorrhagic DAI lesions on other sequences, such as DWI and T2 FLAIR. Therefore, brainstem lesions seen on 3T MRI may not have the same association with outcome as lesions seen with CT or lower resolution MRI. It is also important to acknowledge that although previous studies suggested an association between brainstem lesions and poor neurologic outcome, it is not clear how to translate these data to individual patients due to sample heterogeneity, inconsistencies in outcome reporting, differences in MRI techniques, and the wide confidence intervals reported for functional outcome measures. 1 Thus, the presence of a brainstem lesion (i.e., Grade 3 DAI) should not be used as the sole premise for a pessimistic prognosis in an individual patient.
Our findings also highlight the potential clinical utility of precise neuroanatomic localization of brainstem lesions in neuroprognostication. Similar to prior studies of brainstem lesion mapping,8,11,15 the two patients described here had brainstem lesions that spared the majority of the arousal nuclei. Although the prognostic relevance of brainstem lesion mapping has yet to be demonstrated in large studies, preliminary evidence suggests that lesions involving arousal nuclei and their axons have a greater effect on long-term functional outcomes than do lesions involving non-arousal nuclei and axons.10,11
These findings underscore important opportunities to develop approaches to DAI classification and prognosis that concord with contemporary neuroimaging capabilities. The clinico-radiologic data from these two patients with grade 3 DAI are emblematic of the challenges of neuroprognostication in the ICU and the need to acknowledge uncertainty when communicating with families about prognosis. 16 Clinicians caring for patients with Grade 3 DAI should exercise caution before issuing prognoses derived from studies of prior decades that do not align with current neuroimaging techniques.
Footnotes
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Supported by the NIH BRAIN Initiative (F32MH123001), NIH National Institute of Neurological Disorders and Stroke (R21NS109627, RF1NS115268), NIH Director’s Office (DP2HD101400), Henry and Allison McCance Center for Brain Health/Mass General Neuroscience SPARC Award, AAN Palatucci Advocacy Leadership Grant, James S. McDonnell Foundation, and Tiny Blue Dot Foundation. The funders had no role in the design, analysis, preparation, review, approval, or decision to submit this manuscript for publication.
