Interval-censored survival analysis of mild traumatic brain injury with outcome based neuroimaging clinical applications

Abstract

Objective

The purpose of this study was to assess the relationship between MRI findings and clinical presentation and outcomes in patients following mild traumatic brain injury (mTBI). We hypothesize that imaging findings other than hemorrhages and contusions may be used to predict symptom presentation and longevity following mTBI.

Methods

Patients (n = 250) diagnosed with mTBI and in litigation for brain injury underwent 3T magnetic resonance imaging (MRI). A retrospective chart review was performed to assess symptom presentation and improvement/resolution. To account for variable times of clinical presentation, nonuniform follow-up, and uncertainty in the dates of symptom resolution, a right censored, interval censored statistical analysis was performed. Incidence and resolution of headache, balance, cognitive deficit, fatigue, anxiety, depression, and emotional lability were compared among patients. Image findings analyzed included white matter hyperintensities (WMH), Diffusion Tensor Imaging (DTI) fractional anisotropy (FA) values, MR perfusion, auditory functional MRI (fMRI) activation, hippocampal atrophy (HA) and hippocampal asymmetry as defined by NeuroQuant ® volumetric software.

Results

Patients who reported LOC were significantly more likely to present with balance problems (p < 0.001), cognitive deficits (p = 0.010), fatigue (p = 0.025), depression (p = 0.002), and emotional lability (p = 0.002). Patients with LOC also demonstrated significantly slower recovery of cognitive function than those who did not lose consciousness (p = 0.044). Patients over the age of 40 had significantly higher odds of presenting with balance problems (p = 0.006). Additionally, these older patients were slower to recover cognitive function (p = 0.001) and less likely to experience improvement of headaches (p = 0.007). Abnormal MRI did not correlate significantly with symptom presentation, but was a strong indicator of symptom progression, with slower recovery of balance (p = 0.009) and cognitive deficits (p < 0.001).

Conclusion

This analysis demonstrates the utility of clinical data analysis using interval-censored survival statistical technique in head trauma patients. Strong statistical associations between neuroimaging findings and aggregate clinical outcomes were identified in patients with mTBI.

Keywords

Interval-censored survival analysis advanced neuroimaging mTBI clinical outcomes white matter hyper-intensities diffusion tensor imaging post-concussive syndrome

Introduction

Traumatic Brain Injury (TBI) is a major health challenge around the world, with an estimated burden ranging from 4.6 million cases per year in North America to 18.3 million annual cases in southeast Asia.¹ Severity of a TBI is generally categorized as mild, moderate, or severe. The most common type of TBI is mild, making up approximately 81% of TBIs worldwide.¹ Diagnosis of mTBI is complicated by the fact that individuals with mTBI may not seek immediate medical attention. Additionally, there is currently no definitive set of validated diagnostic tests for mTBI.² Advancements in neuroimaging technology offer great promise in providing objective evidence for diagnosing and predicting the prognosis for patients with mTBI.^3–5

Characterizing the clinical outcomes of patients following mTBI also comes with a unique set of challenges. Symptoms of mTBI are numerous, sometimes nonspecific, and typically rely on self-reporting.^6,7 Examples of symptoms frequently associated with mTBI include headache, dizziness, irritability, fatigue, cognitive problems, and poor concentration.⁸ Studies of symptom duration in patients with mTBI report inconsistent patterns; some studies have found that most symptoms of mTBI are transient and resolve within weeks or months, while others report that mTBI patients can retain cognitive, physiological, and clinical symptoms for years.^9–12 These inconsistencies as well additional pathophysiologic, neuroanatomic, and diagnostic challenges are well described by Bigler in the 2008 publication Neuropsychology and clinical neuroscience of persistent post-concussive syndrome.¹⁰ For example, the terms ‘concussion’ and ‘mild traumatic brain injury’ are often used interchangeably, but exact criteria for both of these terms varies among authors. In general, post-concussive syndrome (PCS) is less explicitly defined and describes symptoms in the acute stages of TBI whereas ‘persistent post-concussive syndrome’ (PPCS) is defined in terms of symptoms lasting more than three months.¹⁰ This study focuses on symptoms commonly described in concussion, mTBI, PCS and PPCS; headache and deficits in balance, emotion, and cognition.

The purpose of this study was to add to the current scientific understanding of mTBI symptomology and prognosis with additional consideration for sex, age, MRI normalcy, and loss of consciousness (LOC). To accurately estimate symptom longevity from our retrospective dataset, this study utilized a statistical method to incorporate the uncertainties associated with irregular time between follow-up visits, irregular time of clinical presentation and patients who either fail to improve or who fail to return for follow-up visits.

Material and methods

Subjects

All subjects were diagnosed with mTBI by a board certified neurologist specializing in head trauma (EF) utilizing standard DSM V diagnostic criteria.¹³ This definition defines a mTBI as:

An impact to the head or other mechanisms of rapid movement or displacement of the brain within the skull, with one or more of the following: loss of consciousness, posttraumatic amnesia, disorientation and confusion, neurological signs such as a new onset of seizure, a marked worsening of a preexisting seizure disorder, visual field cuts, anosmia or hemiparesis with neurocognitive complaints that present immediately after the occurrence of the traumatic brain injury or immediately after recovery of consciousness along with associated symptoms such as disturbances in emotional function (irritability, easy frustration, tension and anxiety, affective lability), personality changes (disinhibition, apathy, suspiciousness, aggression), physical disturbances (headache, fatigue, sleep disorders, vertigo or dizziness, tinnitus or hyperacusis, photosensitivity, anosmia, reduced tolerance to psychotropic medications) and in more severe TBI associated neurological symptoms and signs (seizures, hemiparesis, visual field disturbances, cranial nerve deficits) and evidence of orthopedic injuries. ¹³

The median time between Date of Injury (DOI) and initial consultation was 70 days (range: 2 to 1,126). All subjects had a single MRI brain exam with advanced imaging, typically obtained after the initial clinical consult. Exclusion criteria were: incomplete records [N = 7], canceled/incomplete MRI [N = 71], >3 years between DOI and consultation [N = 7], demyelinating disease [N = 10], history of a cerebrovascular accident [N = 13], prior head trauma [N = 12], chronic epilepsy [N = 1], prior brain surgery [N = 1], imaging findings of brain contusions [N = 4], hemorrhagic diffuse axonal injury (DAI) [N = 4], sequela of intracranial hemorrhages [N = 0], cerebral vascular malformations [N = 0], cavernoma [N = 2], or brain tumors [N = 0]. Out of 375 mTBI patients in litigation for brain injury, 250 were included in the study. Of the 250 included patients, documented preexisting conditions were as follows: hypertension [N = 12], diabetes [N = 3], and migraines [N = 6]. Additionally, 32 subjects were prescribed a medication commonly taken for a psychiatric condition and 15 of these had a documented preexisting psychiatric condition (i.e. anxiety and depression). There was no documented evidence of preexisting balance or dizziness, including Meniere’s disease or benign positional vertigo (BPV), in any patients. Since de-identified data were used, no consent was necessary, and this study was determined exempt from further oversight following an initial review by the Touro University Nevada Institutional Review Board (IRB4-5-17D).

Measures

Seven common mTBI symptoms were considered: headache, balance, cognitive deficit, fatigue, anxiety, depression, and emotional lability.^8,14 Cognitive deficit was defined as impairment in attention, learning and memory, as well as frontal executive functions and language and communication.¹⁵ Cognitive symptoms were assessed using a standard written questionnaire as well as assessment by the neurologist. Balance deficits were defined as new onset symptoms including ‘unsteadiness on the feet’, ‘feeling of falling’ and ‘coordination problems’ as assessed by the neurologist as well as confirmatory physical exam such as the Romberg test for balance. Anxiety, depression and emotional lability were diagnosed using standard DSMV diagnostic criteria.¹³

For the chart review, the status of symptoms was noted for each office visit. Clinical symptoms were reported as sustained, improved, or resolved. Rarely, symptoms were reported as worsening, but such cases were combined with the ‘sustained’ category. Gender, LOC, age (<40 or ≥40), and MRI findings (normal or abnormal) were all considered as factors which might mediate the presentation or duration of symptoms.

MRI acquisition and interpretation

Patients underwent an MRI protocol with a 3.0 Tesla MR Signa HDxt system (G.E. Healthcare Milwaukee WI) equipped with an 8-channel head coil. Standard sequences included Axial T2 and FLAIR and sagittal FLAIR imaging with in-plane resolutions of 0.45 mm and a slice thickness of 4.0 mm. A gradient echo imaging (GRE) sequence with an in-plane resolution of 0.45 mm × 0.45 mm and a slice thickness of 5.0 mm (TE 6.976 ms, TR 400 ms) or a susceptibility weighted imaging (SWI) sequence with a voxel size of 0.39 mm x 0.39 mm x 1.2 mm (TE 25.024 ms, TR 46.8 ms) was also acquired. A 3D T1-weighted sequence with a voxel size of 0.93 mm x 0.93 mm x 1.2 mm (TE 2.208 ms, TR 5.396 ms) was also performed on each patient. The T1 images were subsequently analyzed with NeuroQuant® version 3.0 volumetric analysis software.

Single Shell DTI was acquired with an in-plane resolution of 1.0156 mm and a slice thickness of 2.4 mm. Thirty-three directions were sampled with a b-value 1000 s/mm² compared against a b0 image. Experienced technologists used an Olea software equipped free-standing workstation to draw six regions of interest across the corpus callosum (total, anterior-inferior, anterior, mid-body, posterior, and posterior-inferior), to calculate FA values.

Dynamic whole brain perfusion imaging was performed with a voxel size of 1.875 mm x 1.875 mm x 4.0 mm (TE 14.0 ms, TR 2000.0 ms, flip angle 90°) with a timed injection of Optimark intravenous contrast. Evaluations of regional cerebral blood volume (rCBV), regional cerebral blood flow (rCBF), mean transit time (MTT), time to peak flow (TTP), and time-to-maximum (Tmax) were performed. Post processing analysis was completed on the software equipped Olea free-standing workstation with Tmax considered primarily a measure of macrovascular characteristics reflecting a combination of delay, dispersion, and mean transit time.

fMRI was performed utilizing an echo planar imaging sequence with a voxel size of 3.75 mm x 3.75 mm x 4.0 mm (TE 35.0 ms, TR 3000.0 ms, flip angle 90°) and analyzed with BrainWave software to localize the BOLD signal. Separate auditory tasks were performed on the right ear, the left ear, and bilaterally. During each of these three blocks, simple computer-generated nouns (each word repeated 3 times) were presented through headphones for 30 seconds with a 30 second rest period. Each block had 8 cycles, and each cycle contained 4 presentations of the auditory stimulus and 4 rest periods with no auditory stimulus. Audiologic evaluations for hearing abnormalities were not performed.

An abnormal MRI was defined as any one of the following: unusual or atypical white matter hyper-intensities (WMH), hippocampal atrophy (HA) or asymmetry (NeuroQuant®), abnormal fractional anisotropy (FA) values on Diffusion Tensor Imaging (DTI), abnormalities in MR perfusion, and abnormal auditory functional MRI (fMRI). Abnormal white matter hyper-intensities were defined as white matter foci in subcortical locations without deep or periventricular hyperintensities. WMH patterns consistent with chronic small vessel disease and microvascular changes of aging were not considered abnormal for the purposes of this TBI study or quantified as these are relatively common in older populations. Hippocampal atrophy was defined as either hippocampus <5 percentile in total overall volume as compared to age matched controls and/or in the <5 or >95 percentile asymmetry (NeuroQuant®). Abnormal DTI was defined as decreased FA values relative to age, determined by radiologists’ combined 10 years of experience with this protocol, spot references to a trauma database from the same scanners, and a large normative database synchronized with human phantom.^16,17 To qualify as an abnormal DTI, at least one FA value of the 6 regions of the corpus callosum did not conform to age expected values, which was determined subjectively by the neuroradiologists in a clinical manner. Cutoff values were not utilized as FA values are decreased in children¹⁸ and then again decrease in the elderly^19,20 and therefore exact normative values are not well established. The range of the callosal FA values generated using our technique was between 0.45 and 0.65.

Abnormal perfusion was defined as significant asymmetry in the normal expected perfusion color maps. Abnormal fMRI was defined as decreased activation in the temporal cortices or asymmetric activation following auditory word stimulation.

Double blind MRI interpretation were performed by two board certified neuroradiologists with disagreements (which occurred for 21.3% of readings) reached by consensus.

Statistical analysis

The effects of gender, LOC, age, and MRI findings on the occurrence of symptoms and improvement or resolution of symptoms were described using odds ratios and 95% confidence intervals. Comparisons between groups were made using a Fisher’s exact test for each symptom-characteristic pair. The Fisher’s exact test is more robust than a chi-square test for data that have low counts in one cell, and it is valid for all 2×2 table count data. Independence among gender, LOC, age, and MRI findings was also assessed using a Fisher’s exact test.

Symptom longevity was calculated from the date of injury to the time when improvement or resolution (both referred to as ‘improvements’ hereafter) was reported during follow-up visits. Time elapsed between follow-up appointments resulted in non-instantaneous reporting of symptom progression, since improvements could only be reported at scheduled visits. The resulting data are interval-censored, which means that the number of days the symptom persisted after the date of injury are known to be within a time interval but cannot be localized to an exact date. Each patient in the analysis was counted only once for the analysis, and the uncertainty associated with the elapsed time between visits is built into the maximum likelihood estimate of the survival function. To make the situation more complicated, many patients never reported improvement or resolution. In such cases, the data are right-censored, which means that the symptom might improve at some future time, but it had not improved by the date of the last visit to the physician. Data that are both interval-censored and right-censored are relatively common in medical studies, but often statistical methods fail to account for the associated uncertainty.²¹ To estimate longevity of symptoms reported in patients, a Cox proportional hazards model was fit which allowed interval censoring and right censoring. No parametric options provided a satisfactory fit for the observed hazard function, so semi-parametric Cox proportional hazards models were used to estimate and compare treatments.²² The analysis assumes that, for an interval of time between two appointments which can be expressed as (L, R], where ‘R’ is the physician visit at which the patient reported improvement and ‘L’ is the physician visit immediately preceding ‘R’, it is equally probable that the symptom abated or resolved across the entire time interval.

Models for groups defined by sex, LOC at the time of injury, age, and MRI findings were estimated separately for each symptom. Standard errors were derived from bootstrapping (1,000 samples), and differences in symptom persistence were tested by using the bootstrapped standard errors to form z-scores. Analyses were conducted in R 3.5.1 software with the ‘icenReg’ package.

Results

The sample population had a relatively high percentage (62.4%) of females, with the primary cause of injury for all patients being motor vehicle collisions (Table 1). Headaches, cognitive disabilities, and balance issues were the most commonly reported symptoms, and they were also the most likely to improve or resolve (Figure 1; Table 2). Anxiety and depression affected approximately half of the patients. Fatigue and emotional lability were both the least commonly noted symptoms and the least likely to improve or resolve (Figure 1; Table 2). On average, patients presented with 4.2 symptoms (presenting for evaluation a median 70 days post injury), with 77.6% of patients having 3–6 symptoms. Some pairs of symptoms tended to co-occur, most notably anxiety with depression or cognitive problems (Table 3).

Table 1.

Demographics of 250 analyzed patients. Mean age of surveyed population was 43 years (range: 12 to 82).

		%
Age	12–19	6
	20–39	42
	40–59	38
	60–82	14
Race/ethnicity	Asian	9
	Black or African American	7
	Hispanic or Latino	17
	White	43
	N/A	24
Gender	Female (mean age = 40.6)	62
	Male (mean age = 43.5)	38
Cause of injury	Motor vehicle collision	80
	Fall	7
	Assault	5
	N/A	8

Figure 1.

Count of patients who presented symptoms divided into groups based on their characteristics, separated by symptom improvement or resolution during the study. A ‘^’ indicates that the sample size of patients who improved was too small for further analysis.

Table 2.

Prevalence of symptoms and symptom resolution by number of patients (percent of patient population for occurrence and percent based on affected patients for percent improved or resolved) in analyzed patient population.

	Occurrence	Improved	Resolved
Headache	237 (97%)	62 (26%)	11 (5%)
Balance	203 (83%)	47 (23%)	9 (4%)
Cognitive	221 (90%)	59 (27%)	5 (2%)
Fatigue	76 (30%)	2 (3%)	1 (1%)
Anxiety	122 (49%)	10 (8%)	3 (2%)
Depression	112 (46%)	10 (9%)	2 (2%)
Emotional lability	59 (24%)	2 (3%)	0 (0%)

Table 3.

Pairwise co-incidence of symptoms after mTBI. Odds ratios are expressed as odds of having both symptoms/odds of having only one of the symptoms.

	Balance	Cognitive	Fatigue	Anxiety	Depression	Emotional lability
Headache	7.1	1.5	2.7	2.4	2.1	1.9
	1.2, 50.5	0.0, 13.4	0.3, 124.8	0.4, 26	0.3, 22.8	0.2, 89.9
Balance		4.9	2.5	4.3	6.4	4.9
		1.8, 12.8	1.0, 6.9	1.9, 10.8	2.5, 19.5	1.5, 25.7
Cognitive			*	5.8	24.0	8.5
				1.9, 23.8	3.8, 999.7	1.3, 356.1
Fatigue				3.5	3.0	5.6
				1.9, 6.5	1.7, 5.6	2.9, 11
Anxiety					66.0	2.0
					28.4, 170.8	1.1, 3.8
Depression						2.1
						1.1, 4.1

Note: Odds ratios and 95% confidence intervals are reported. Values with Bonferroni-adjusted p-value <0.05 are in bold. Note that fatigue always co-occurred with cognitive symptoms, indicated on this table as *.

Static improvement statistics were complicated by the fact that some patients were followed for over 1,000 days post injury, and others were followed less than three months. The median time between the date of Injury (DOI) and initial consultation was 73 days (range: 0 to 1,536; IQR: 34 to 148); the median number of days between DOI and imaging was 107 (range: 12 to 1,854; IQR: 56 to 196). Each patient had a median of 2 (range: 1 to 9; IQR: 2 to 3) clinical follow-up appointments. The median time between DOI and the last clinical visit was 150 days (range: 0 to 1,175; IQR: 91 to 251).

It would have been insightful to analyze symptom resolution and improvement separately, but few patients reported complete resolution of one or more symptoms (Table 2). Therefore, symptom resolution and improvement were combined for the symptom longevity analysis. Analyses including symptoms which had very few patients who saw improvement or resolution had lower statistical power and produced less reliable descriptors of actual symptom longevity.

There was no evidence the independent variables of Sex, LOC, Age, or MRI findings tended to co-occur (analyses not shown; Fisher’s exact test between pairs p > 0.265).

Effect of sex

There were no gender-based differences in clinical symptom presentation (Figure 2), nor were there any differences in symptom longevity observed between genders (Table 4, Figure 3(a)).

Figure 2.

Odds of presenting symptom based on characteristics, including sex, LOC, age (less than or equal to 40, greater than 40), and abnormal MRI findings. In the tabular data, p-values from the Fisher’s exact test are reported, and ‘Inf' indicates infinitely large log odds ratio because all individuals with the characteristic had the symptom. In the plot, the log odds and 95% confidence intervals are provided. The numbers on the far-right side of the plot indicate the upper 95% confidence limit, which was truncated on the plot for readability.

Table 4.

Statistics for survival functions comparing symptom retention between patient subgroups.

	Symptom	Est. effect	Hazard ratio	SE	z-value	P-value
Sex (Male/female)	Headache	0.05	1.05	0.25	0.2	0.840
Sex (Male/female)	Balance	–0.17	0.84	0.30	–0.6	0.570
	Cognitive	–0.36	0.70	0.29	–1.2	0.214
	Anxiety	0.14	1.15	1.06	0.1	0.896
	Depression	–0.19	0.82	1.37	–0.1	0.888
Loss of consciousness (with/without)	Headache	–0.47	0.63	0.30	–1.6	0.114
	Balance	–0.57	0.56	0.34	–1.7	0.094
	Cognitive	–0.63	0.53	0.31	–2.0	0.044
	Anxiety	–0.59	0.55	1.51	–0.4	0.695
	Depression	–0.46	0.63	1.52	–0.3	0.760
Age (<40 years/>40 years)	Headache	–0.68	0.51	0.26	–2.7	0.007
Age (<40 years/>40 years)	Balance	–0.34	0.71	0.29	–1.2	0.238
	Cognitive	–0.89	0.41	0.27	–3.3	0.001
MRI status (normal/abnormal)	Headache	–0.46	0.63	0.24	–1.9	0.060
	Balance	–0.75	0.47	0.28	–2.6	0.009
	Cognitive	–0.91	0.40	0.26	–3.5	<0.001

Note: The ‘Est Effect’ coefficient indicates the effect size on symptom retention. The hazard ratio (hazard in the group with the characteristic minus hazard in the group without the characteristic), bootstrapped standard errors (SE), z value, and P-value are also reported.

Figure 3.

Estimated symptom longevity based on characteristics, which are denoted in the upper-right key, Statistics are provided in Table 4.

Effect of loss of consciousness

Patients that lost consciousness at the time of injury had higher odds of balance issues, cognitive problems, fatigue, depression, and emotional lability (Figure 2). All patients who lost consciousness had headaches. The only statistically significant difference in recovery time in those with LOC was in cognitive function (Table 4, Figure 3(b)).

Effect of age (>40)

Notable differences in symptom presentation and longevity were identified in age dichotomized patient groups. Patients over the age of 40 had significantly higher odds of reporting balance problems at presentation (Figure 2). Patients over the age of 40 were also slower to recover cognitive function and from headaches (Table 4, Figure 3(c) and (d)). Recovery or improvements associated with anxiety and depression were too rare to allow analysis of recovery time.

Effect of abnormal MRI

There were no significant differences in symptom presentation based on abnormal MRI findings, although there were trends in depression and anxiety (Figure 2). Patients who had one or more abnormal MRI findings had slower recovery from balance and cognitive deficits (Table 4, Figure 3(f) and (g)). Other symptoms, such as depression, also appeared to show slower recovery in those with abnormal MRI findings (Figure 3(h)), but the sample sizes were not large enough to test statistical significance.

Sixty percent of patients [151] had abnormal MRI findings. Counted independently of other MRI abnormalities, the prevalence of abnormalities in the five categories was: 25% [62] atypical WMH, 15% [38] hippocampal atrophy or asymmetry, 30% [75] abnormal DTI, 6% [16] abnormal perfusion and [87] 35% abnormal auditory fMRI. 57 patients had one abnormal MRI finding, 65 had two abnormal MRI findings, 25 had three abnormal MRI findings and four had four abnormal MRI findings. No patients had all five abnormal MRI findings. The average ages of subjects with a normal MRI versus abnormal MRI was 41.3 and 41.9 years respectively. Examples of imaging findings in each of the 5 categories is shown in supplemental Figures 1 to 6.

Discussion

To our knowledge, this is the first study to examine symptom presence and longevity after mTBI within a sizeable civilian patient population using analysis techniques capable of handling retrospective clinical data (even with irregular visit intervals and follow-up). This study is also the largest review of mTBI patients in litigation. A study of 147 severe TBI patients found litigation to be an independent significant predictor of poorer outcomes.²³ A larger meta-analysis found a ‘moderate effect’ of litigation on outcomes in TBI patients.²⁴ However, other review articles found little effect on cognition and headaches in litigants versus non litigants²⁵ with reported widespread continuation of post-traumatic headaches despite litigation resolution.²⁶

The subjects in this study were more than half female and included a broad age range, from young children to elderly patients. Most studies in the literature have matched controls, whereas this is a retrospective cohort view, which may allow a more representative sample to be included. Many of the disparities between this study and others may be a consequence of the different study populations, study designs and sample sizes.²⁷

Results of this study are generally concordant with already published work. Although there is some limitation in comparison, presenting symptoms of our subjects appear similar to those described in the literature, though we report lower rates of fatigue and higher rates of headache.^28–30 Some studies have found mTBI patients typically experience a complete recovery and may resume a normal lifestyle in a matter of weeks or occasionally even days.^31,32 Losoi et al. found that by 6 months nearly 90% of the mTBI patients had reported “good recovery”.³³ However, many other studies have found significant deficits and symptoms greater than a year following injury.^30,34–37 McInnes et al. conducted a large meta review and found a majority of subjects showed cognitive impairment >12 months after injury.³⁸ A 2019 meta-analysis found less than half (47.2%) of patients experienced full recovery in 12 months.³⁹ These recent studies are more consistent with our results relative to frequency of symptom presentation and prolonged experience of mTBI symptoms.⁴⁰

Sex associated sequelae of mTBI

TBI is more prevalent in males,^41,42 likely due to behavioral/risk taking patterns.^42,43 Evidence for gender differences in TBI symptomology and recovery is more mixed. Several studies have identified female subjects as having diminished recovery and increased severity of symptoms following mTBI.^42,44,45 This study found no differences between males and females in symptom presence or symptom longevity.

Clinical symptoms after loss of consciousness

The results of this study indicate LOC plays a stronger role in initial symptom presentation and a lesser role in symptom longevity, this is concordant with the results and review of Roy et al, who studied 407 mTBI subjects and found much higher rates of post concussive symptoms in mTBI patients who lost consciousness evaluated one month following head trauma, but no significant associations at a 6 month evaluation.⁴⁶ However, contrary to Roy et al, our subjects who lost consciousness were more likely to present with depression. This study and our study indicate LOC may not be as important for long term outcomes as previously thought. The known dementia risk from mTBI does not require LOC.⁴⁷

Age as a factor in mTBI outcomes

Age is a well-recognized determinant of outcome after head trauma, with older patients suffering worse outcomes.^48–51 In this study, older age (≥40) was adversely related to both symptom presentation and symptom improvement.

MRI abnormalities and mTBI outcomes

Hemorrhagic diffuse axonal injury, cerebral contusions, and intra and extra-axial hemorrhages are well described in the trauma literature as correlating with poorer prognosis; however, the 5 categories comprising our ‘abnormal MRI group’ have been less well studied, particularly in mTBI. This study suggests that the MRI abnormalities studied may be associated with slow recovery of balance and cognition.

Non-hemorrhagic shearing injuries/white matter hyper-intensities secondary to head trauma are well described in the literature.^52–54 There is a reported 30% to 50% resolution of traumatic WMH in the weeks to months following head trauma.¹⁰ Discerning post traumatic WMH from incidental WMH, changes of aging and microvascular disease, and other etiologies is challenging in clinical practice, particularly in older patients. Incidence of WMH in normative populations is varied, with some reporting low incidence^55–57 and others higher incidence.^54,58 Comparing WMH in TBI versus controls, some studies have found no difference, while some have found changes in subcortical predominance of lesions⁵⁹; others have documented trends approaching significance.⁵⁴ Tate et al., did not find significant differences in WMH between controls and TBI subjects, but did find that TBI patients with WMH had significantly worse cognitive deficits than those without WMH.⁶⁰ 25% [62] of our subjects has atypical WMH.

It is well established that brain trauma may result in hippocampal atrophy with the degree of atrophy related to severity of injury.^61,62 Hippocampal atrophy appears to correlate with clinical outcomes in mTBI.^63,64 NeuroQuant® is an FDA-approved software program for measuring brain MRI volumes and can be used in the clinical setting to assess for hippocampal atrophy following traumatic brain injury and may be more effective than radiologists simply “eyeballing” brain structures.^65,66 15% [38] of our subjects had hippocampal atrophy or asymmetry.

Diffusion Tensor Imaging (DTI) measures the ability of water molecules to move through tissue, with notable directional constraints associated with white matter axons.⁶⁷ DTI is reported to be sensitive to diffuse axonal injury.⁶⁸ Many studies have found differences between mTBI patients and controls.^16,69,70 Abnormalities in DTI are reported to correlate with clinical outcomes.^16,71 Wilde et al. found DTI abnormalities associated with LOC in mTBI.⁷² 30% [75] of our subjects were classified as having an abnormal DTI, with FA values decreased as expected per age by two neuroradiologists with experience interpreting FA values on these scanners.

Perfusion imaging assesses blood flow to regions of the brain. Non-uniform cerebral perfusion has been identified in patients with a variety of brain injuries including head trauma. Reduced cerebral blood flow has been demonstrated following head trauma with improvement in cerebral blood flow correlating with clinical recovery.^73,74 6% [16] of our subjects had abnormal perfusion.

In functional magnetic resonance imaging (fMRI) blood oxygen-level-dependent (BOLD) contrast is altered in response to local blood flow changes associated with increased regional neuronal activity.^75,76 Auditory fMRI has been studied for over 20 years.^77–79 Although less well investigated than the other imaging measures, abnormalities in auditory fMRI may correlate with central auditory processing disorder, and a recent study suggested mild differences between mTBI patients and controls in a military population.⁸⁰ Therefore, our auditory fMRI protocol was performed using this same protocol and we found 87 (35%) of our subjects had an abnormal auditory fMRI.⁸⁰

To our knowledge, existing literature has addressed outcomes rather than symptom presentation relative to MRI findings in mTBI.

Conclusions and limitations

The study had several limitations. Costs associated with advanced imaging limited us to those patients undergoing litigation cases in a retrospective manner and without a control group. The lack of a control group is the primary limitation of this study. For example, the rates of central auditory processing disorders and abnormal auditory fMRI in a normative population have not been well established. Symptom reporting may be subjective and in part relies on self-reporting. Time between DOI and first presentation and length of follow-up was quite variable, limiting conclusions based only on symptom presentation and improvement without accounting for time. We therefore attempted to restrict our statistical interpretation to differences among groups. ‘Improvement’ in symptoms was not quantified or scaled. Standardized cognitive tests with quantifiable results were not administered. Quantifiable vestibular testing was not performed. Time of loss of consciousness was not quantified. Secondary gain and other factors in the medical legal process have potential to distort data. It is also possible that patients with worse mTBI would be more likely to enter litigation.

Excluding the NeuroQuant® hippocampal volumetric analysis, all the additional imaging findings interpreted by the neuroradiologists contain an element of subjectivity. For example, study design did not include quantified predetermined cutoff DTI FA values or comparison to a normative DTI database with age and sex matched controls and z scores. While unusual in young and middle-aged patients, hearing loss estimates in patients over 65 are 25–40% and therefore the auditory fMRI results may be less reliable in the older patients.⁸¹ Resting state fMRI, a component of the auditory fMRI performed, has shown promise in detecting abnormal connectivity in head trauma patients, but this analysis was not performed. We did not assess which abnormal MRI categories played the largest role in the detected adverse clinical outcomes due to limited sample size within categories.

Despite these limitations, our observations agree with existing literature regarding patient age, LOC, and MRI findings.

The purpose of this study was to assess if further investigation of these advanced neuroimaging and white matter findings was warranted using more objective measures and to assert the validity of this statistical technique to analyze the incidence of clinical outcomes. Given the strong association of abnormal advanced imaging and white matter findings with clinical outcomes, further investigation is justified.

A more detailed understanding of symptoms and the prognosis following mTBI has important implications for health care costs and quality.⁸² Interval censored survival statistical analysis appears to be a valid method to assess head trauma cohorts and predict clinical outcomes based on quantitative clinical data. The diagnostics of a mild traumatic brain injury may still rely in part on patient self-reporting and keen clinical assessment; however, this study indicates neuroimaging techniques are likely underutilized in contributing to diagnosis and prognosis of mTBI.

Supplemental Material

sj-pdf-1-ccn-10.1177_2059700220947194 - Supplemental material for Interval-censored survival analysis of mild traumatic brain injury with outcome based neuroimaging clinical applications

Supplemental material, sj-pdf-1-ccn-10.1177_2059700220947194 for Interval-censored survival analysis of mild traumatic brain injury with outcome based neuroimaging clinical applications by Cheryl Vanier, Trisha Pandey, Shaunaq Parikh, Alan Rodriguez, Thomas Knoblauch, John Peralta, Amanda Hertzler, Leon Ma, Ruslan Nam, Sami Musallam, Hallie Taylor, Taylor Vickery, Yolanda Zhang, Logan Ranzenberger, Andrew Nguyen, Mike Kapostasy, Alex Asturias, Enrico Fazzini and Travis Snyder in Journal of Concussion

Footnotes

Acknowledgments

We would like to thank Touro University Nevada, its medical students, and staff that participated in the research of this manuscript. We would like to acknowledge the late Neuroradiologist William Orrison Jr. M.D. for his vision and guidance in this project.

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.

Disclosure statement

The data reviewed and analyzed were from a cohort of mild traumatic brain injury patients all of whom were in litigation during the time of physician documentation.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Thomas Knoblauch

Supplemental material

Supplemental material for this article is available online.

References

Michael

Abbas

Saksham

, et al. Estimating the global incidence of traumatic brain injury. J Neurosurg 2018; 130: 1080–1097.

Prince

Bruhns

ME.

Evaluation and treatment of mild traumatic brain injury: the role of neuropsychology. Brain Sci 2017; 7: 1-2.

Ashwal

Tong

Ghosh

, et al. Application of advanced neuroimaging modalities in pediatric traumatic brain injury. J Child Neurol 2014; 29: 1704–1717.

Kou

Tong

, et al. The role of advanced MR imaging findings as biomarkers of traumatic brain injury. J Head Trauma Rehabil 2010; 25: 267–282.

Shenton

Hamoda

Schneiderman

, et al. A review of magnetic resonance imaging and diffusion tensor imaging findings in mild traumatic brain injury. Brain Imaging Behav 2012; 6: 137–192.

Hoge

McGurk

Thomas

, et al. Mild traumatic brain injury in U.S. Soldiers returning from Iraq. N Engl J Med 2008; 358: 453–463.

Stein

McAllister

TW.

Exploring the convergence of posttraumatic stress disorder and mild traumatic brain injury. Am J Psychiatry 2009; 166: 768–776.

National Center for Injury Prevention and Control. Report to congress on mild traumatic brain injury in the United States: steps to prevent a serious public health problem. Atlanta, GA: Centers for Disease Control and Prevention: National Center for Injury Prevention and Control, 2003.

Bazarian

Wong

Harris

, et al. Epidemiology and predictors of post-concussive syndrome after minor head injury in an emergency population. Brain Injury 1999; 13: 173–189.

10.

Bigler

ED.

Neuropsychology and clinical neuroscience of persistent post-concussive syndrome. J Int Neuropsychol Soc 2008; 14: 1–22.

11.

Rimel

Giordani

Barth

, et al. Disability caused by minor head injury. Neurosurgery 1981; 9: 221–228.

12.

Vanderploeg

Curtiss

Luis

, et al. Long-term morbidities following self-reported mild traumatic brain injury. J Clin Exp Neuropsychol 2007; 29: 585–598.

13.

American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 5th ed. Washington DC: American Psychiatric Association, 2013.

14.

World Health Organization. ICD-10: international statistical classification of diseases and related health problems. 2nd ed. Geneva: World Health Organization, 2004.

15.

Silver

McAllister

Yudofsky

SC.

Textbook of traumatic brain injury. Arlington, VA: American Psychiatric Publishing, Inc., 2005.

16.

Hulkower

Poliak

Rosenbaum

, et al. A decade of DTI in traumatic brain injury: 10 years and 100 articles later. AJNR Am J Neuroradiol 2013; 34: 2064–2074.

17.

Venkatraman

Gonzalez

Landman

, et al. Region of interest correction factors improve reliability of diffusion imaging measures within and across scanners and field strengths. NeuroImage 2015; 119: 406–416.

18.

Feldman

Yeatman

Lee

, et al. Diffusion tensor imaging: a review for pediatric researchers and clinicians. J Dev Behav Pediatr 2010; 31: 346–356.

19.

Inano

Takao

Hayashi

, et al. Effects of age and gender on white matter integrity. AJNR Am J Neuroradiol 2011; 32: 2103–2109.

20.

Rathee

Rallabandi

Roy

PK.

Age-related differences in white matter integrity in healthy human brain: evidence from structural MRI and diffusion tensor imaging. Magn Reson Insights 2016; 9: 9–20.

21.

Eisenberg

Andrea

Meehan

, et al. Time interval between concussions and symptom duration. Pediatrics 2013; 132: 8–17.

22.

Sun

Interval censoring. In: Klein JP, van Houwelingen HC, Ibrahim JG, et al. (eds) Handbook of survival analysis. Boca Raton, FL: CRC Press, USA, 2014, pp.369–390.

23.

Bayen

Jourdan

Ghout

, et al. Negative impact of litigation procedures on patient outcomes four years after severe traumatic brain injury: results from the Paris – traumatic brain injury study. Disabil Rehabil 2018; 40: 2040–2047.

24.

Binder

Rohling

ML.

Matters: a meta-analytic review of the effects of financial incentives on recovery after closed-head injury. Am J Psychiatry 1996; 153: 7–10.

25.

Evans

RW.

Post-traumatic headaches. Neurol Clin 2004; 22: 237–249.

26.

Packard

RC.

Posttraumatic headache: permanency and relationship to legal settlement. Headache 1992; 32: 496–500.

27.

Schwab

Gudmudsson

Lew

HL.

Chapter 40 – long-term functional outcomes of traumatic brain injury. In: Grafman J and Salazar AM (eds) Handbook of clinical neurology. Amsterdam: Elsevier, 2015, pp.649–659.

28.

Shetty

Nguyen

Cogsil

, et al. Clinical findings in a multicenter MRI study of mild TBI. Front Neurol 2018; 9: 11–09.

29.

Kashluba

Paniak

Blake

, et al. A longitudinal, controlled study of patient complaints following treated mild traumatic brain injury. Arch Clin Neuropsychol 2004; 19: 805–816.

30.

Dikmen

Machamer

Fann

, et al. Rates of symptom reporting following traumatic brain injury. J Int Neuropsychol Soc 2010; 16: 401–411.

31.

Carroll

Cassidy

Peloso

, et al. Prognosis for mild traumatic brain injury: results of the WHO collaborating Centre task force on mild traumatic brain injury. J Rehabil Med 2004; 36: 84–105.

32.

Iverson

GL.

Outcome from mild traumatic brain injury. Curr Opin Psychiatry 2005; 18: 301–317.

33.

Losoi

Silverberg

Waljas

, et al. Recovery from mild traumatic brain injury in previously healthy adults. J Neurotrauma 2016; 33: 766–776.

34.

van der Naalt

Timmerman

de Koning

, et al. Early predictors of outcome after mild traumatic brain injury (UPFRONT): an observational cohort study. Lancet Neurol 2017; 16: 532–540.

35.

Dean

PJA

Sterr

Long-term effects of mild traumatic brain injury on cognitive performance. Front Hum Neurosci 2013; 7: 30.

36.

Peltz

Gardner

Kenney

, et al. Neurobehavioral characteristics of older veterans with remote traumatic brain injury. J Head Trauma Rehabil 2017; 32: E8–E15.

37.

Hiploylee

Dufort

Davis

, et al. Longitudinal study of postconcussion syndrome: not everyone recovers. J Neurotrauma 2017; 34: 1511–1523.

38.

McInnes

Friesen

MacKenzie

, et al. Mild traumatic brain injury (mTBI) and chronic cognitive impairment: a scoping review. PloS One 2017; 12: e0174847.

39.

Nelson

Temkin

Dikmen

, et al.; the TRACK-TBI Investigators. Recovery after mild traumatic brain injury in patients presenting to US level I trauma centers: a transforming research and clinical knowledge in traumatic brain injury (TRACK-TBI) study. JAMA Neurol 2019; 76: 1049.

40.

Nagae-Poetscher

Golay

, et al. Routine clinical brain MRI sequences for use at 3.0 tesla. J Magn Reson Imaging 2005; 22: 13–22.

41.

Mollayeva

El-Khechen-Richandi

Colantonio

Sex & gender considerations in concussion research. Concussion 2018; 3: CNC51.

42.

Munivenkatappa

Agrawal

Shukla

, et al. Traumatic brain injury: does gender influence outcomes? Int J Crit Illn Inj Sci 2016; 6: 70–73.

43.

Collins

Molcho

Carney

, et al. Are boys and girls that different? An analysis of traumatic brain injury in children. Emerg Med J 2013; 30: 675–678.

44.

Bazarian

Blyth

Mookerjee

, et al. Sex differences in outcome after mild traumatic brain injury. J Neurotrauma 2010; 27: 527–539.

45.

Kontos

Elbin

Schatz

, et al. A revised factor structure for the post-concussion symptom scale: baseline and postconcussion factors. Am J Sports Med 2012; 40: 2375–2384.

46.

Roy

Peters

Everett

, et al. Loss of consciousness and altered mental state predicting depressive and post-concussive symptoms after mild traumatic brain injury. Brain Inj 2019; 33: 1064–1066.

47.

Barnes

Byers

Gardner

, et al. Association of mild traumatic brain injury with and without loss of consciousness with dementia in US military veterans. JAMA Neurol 2018; 75: 1055–1061.

48.

Erlebach

Pagnamenta

Klinzing

, et al. Age-related outcome of patients after traumatic brain injury: a single-center observation. Minerva Anestesiol 2017; 83: 1169–1177.

49.

Chiang

Guo

Huang

, et al. Trajectories and associated factors of quality of life, global outcome, and post-concussion symptoms in the first year following mild traumatic brain injury. Qual Life Res 2016; 25: 2009–2019.

50.

Katz

White

Alexander

, et al. Recovery of ambulation after traumatic brain injury. Arch Phys Med Rehabil 2004; 85: 865–869.

51.

Senathi-Raja

Ponsford

Schönberger

Impact of age on long-term cognitive function after traumatic brain injury. Neuropsychology 2010; 24: 336–344.

52.

Chung

Park

Nam

, et al. Locations and clinical significance of non-hemorrhagic brain lesions in diffuse axonal injuries. J Korean Neurosurg Soc 2012; 52: 377–383.

53.

Orrison

Hanson

Alamo

, et al. Traumatic brain injury: a review and high-field MRI findings in 100 unarmed combatants using a literature-based checklist approach. J Neurotrauma 2009; 26: 689–701.

54.

Riedy

Senseney

Liu

, et al. Findings from structural MR imaging in military traumatic brain injury. Radiology 2016; 279: 207–215.

55.

Hopkins

Beck

Burnett

, et al. Prevalence of white matter hyperintensities in a young healthy population. J Neuroimaging 2006; 16: 243–251.

56.

Katzman

Dagher

Patronas

NJ.

Incidental findings on brain magnetic resonance imaging from 1000 asymptomatic volunteers. JAMA 1999; 282: 36–39.

57.

Morris

Whiteley

Longstreth

Jr , et al. Incidental findings on brain magnetic resonance imaging: systematic review and meta-analysis. Bmj 2009; 339: b3016.

58.

Weaver

Wilson

Lindblad

, et al.; NORMAL Study Team. Comprehensive evaluation of healthy volunteers using multi-modality brain injury assessments: an exploratory, observational study. Front Neurol 2018; 9: 1030.

59.

Trifan

Gattu

Haacke

, et al. MR imaging findings in mild traumatic brain injury with persistent neurological impairment. Magn Reson Imaging 2017; 37: 243–251.

60.

Tate

Gusman

Kini

, et al. Susceptibility weighted imaging and white matter abnormality findings in service members with persistent cognitive symptoms following mild traumatic brain injury. Mil Med 2017; 182: e1651–e1658.

61.

Tate

Bigler

ED.

Fornix and hippocampal atrophy in traumatic brain injury. Learn Mem 2000; 7: 442–446.

62.

Wilde

Bigler

Hunter

, et al. Hippocampus, amygdala, and basal ganglia morphometrics in children after moderate-to-severe traumatic brain injury. Dev Med Child Neurol 2007; 49: 294–299.

63.

Bigler

Anderson

Blatter

, et al. Temporal lobe morphology in normal aging and traumatic brain injury. AJNR Am J Neuroradiol 2002; 23: 255–266.

64.

Monti

Voss

Pence

, et al. History of mild traumatic brain injury is associated with deficits in relational memory, reduced hippocampal volume, and less neural activity later in life. Front Aging Neurosci 2013; 5: 41–41.

65.

Ross

Castelvecchi

Ochs

AL.

Brain MRI volumetry in a single patient with mild traumatic brain injury. Brain Inj 2013; 27: 634–636.

66.

Ross

Ochs

Seabaugh

, et al.; Alzheimer's Disease Neuroimaging Initiative. Man versus machine: comparison of radiologists' interpretations and NeuroQuant(R) volumetric analyses of brain MRIs in patients with traumatic brain injury. J Neuropsychiatry Clin Neurosci 2013; 25: 32–39.

67.

Huisman

Schwamm

Schaefer

, et al. Diffusion tensor imaging as potential biomarker of white matter injury in diffuse axonal injury. AJNR Am J Neuroradiol 2004; 25: 370–376.

68.

Ranzenberger

Snyder

Diffusion tensor imaging. Treasure Island, FL: StatPearls, 2019.

69.

Inglese

Makani

Johnson

, et al. Diffuse axonal injury in mild traumatic brain injury: a diffusion tensor imaging study. J Neurosurg 2005; 103: 298–303.

70.

Mac Donald

Johnson

Cooper

, et al. Detection of blast-related traumatic brain injury in U.S. military personnel. N Engl J Med 2011; 364: 2091–2100.

71.

Yin

Huang

, et al. Longitudinal changes in diffusion tensor imaging following mild traumatic brain injury and correlation with outcome. Front Neural Circuits 2019; 13: 28.

72.

Wilde

Hunter

, et al. Loss of consciousness is related to white matter injury in mild traumatic brain injury. J Neurotrauma 2016; 33: 2000–2010.

73.

Meier

Bellgowan

Singh

, et al. Recovery of cerebral blood flow following sports-related concussion. JAMA Neurol 2015; 72: 530.

74.

Yuh

Hawryluk

Manley

GT.

Imaging concussion: a review. Neurosurg 2014; 75 Suppl 4: S50–63.

75.

McAllister

Saykin

Flashman

, et al. Brain activation during working memory 1 month after mild traumatic brain injury: a functional MRI study. Neurology 1999; 53: 1300–1308.

76.

Belin

Zatorre

Hoge

, et al. Event-related fMRI of the auditory cortex. NeuroImage 1999; 10: 417–429.

77.

Leff

Iverson

Schofield

, et al. Vowel-specific mismatch responses in the anterior superior temporal gyrus: an fMRI study. Cortex 2009; 45: 517–526.

78.

Price

CJ.

A review and synthesis of the first 20 years of PET and fMRI studies of heard speech, spoken language and reading. NeuroImage 2012; 62: 816–847.

79.

Yang

Yeo

Pena

, et al. An FMRI study of auditory orienting and inhibition of return in pediatric mild traumatic brain injury. J Neurotrauma 2012; 29: 2124–2136.

80.

Santhanam

Orrison

Wilson

, et al. Central auditory processing disorders after mild traumatic brain injury. Undersea Hyperbaric Med 2019; 46: 1–2.

81.

Yueh

Shapiro

MacLean

, et al. Screening and management of adult hearing loss in primary care: scientific review. JAMA 2003; 289: 1976–1985.

82.

Taylor

Hagel Campbell

Nugent

, et al. Three year trends in veterans health administration utilization and costs after traumatic brain injury screening among veterans with mild traumatic brain injury. J Neurotrauma 2017; 34: 2567–2574.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.26 MB