A REVIEW OF THE ADVANTAGES AND PITFALLS OF THE TECHNIQUES AND METHODS WHICH MAY PROVIDE INSIGHT IN TEMPORAL CHANGES OF TRAUMATIC BRAIN INJURY

(1)

UNIVERSITY OF AMSTERDAM

MASTER IN FORENSIC SCIENCE

LITERATURE THESIS

TOPIC:

A REVIEW OF THE ADVANTAGES AND PITFALLS OF THE

TECHNIQUES AND METHODS WHICH MAY PROVIDE INSIGHT IN

TEMPORAL CHANGES OF TRAUMATIC BRAIN INJURY

SUBMITTED BY: S. DIVYA 11390476

(2)

2

ABSTRACT

Traumatic brain injury (TBI) exists around the globe as a serious issue. The morbidity and mortality attached to it is high. TBI is an acquired injury resulting from external physical force impacting the head, consequently injuring the brain. The impacting forces arises from variety of causes. This thesis focuses on one such cause – physical abuse. Physical abuse is prevalent in both children and adults. Thus, detection of TBI and the dating of injuries are crucial for different reasons, like treatment, forensic casework (corroboration or refutation of victim/suspect/witness statements and event reconstruction). TBI is deemed to be very complex and heterogeneous as a process. A scale of mild, moderate or severe constitutes the injury severity. The pathophysiology of TBI has been generally classified into primary injuries, occurring at time of impact and subsequently developed secondary injuries. Subdural haematoma (SDH) and diffuse axonal injury (DAI) are predominant primary injuries, which will be discussed in this thesis. Recently, characteristics of SDH have been studied to determine the time of injury and thus, possibility of physical abuse – a form of backtracking to identify physical abuse from one of its pathological consequences. Array of methods are available for TBI detection and for dating. These have been broadly categorized as histomorphology and immunohistochemistry (IHC), biomarkers, neuroimaging and spectroscopic methods in this thesis. The usability of these methods varies based on many factors such as soundness. This thesis strives to determine the advantages and disadvantages of current TBI-based methods and if one sound method would suffice for TBI detection (and possibly dating as well), whilst being acceptable for forensic settings. A list of criteria, derived from the requirements of a good scientific method and the forensically-relevant Daubert guidelines have been used to compare the methods. A review of literature scrutinizing these methods was also made. As indicated by multitude of these studies, there is no one ideal gold-standard for TBI detection/dating and rather a combination of these methods would allow more accurate and reliable detection as well as dating of TBI.

INTRODUCTION

Traumatic brain injury (TBI), deemed to be a type of devastating injuries, causes morbidity and mortality, on a global-scale. Being an acquired brain injury, TBI occurs when an external physical force impacts the head and injures the brain (Hoshizaki et al., 2013). Consequently, the normal brain physiology is disrupted. A TBI varies in its severity, which ranges from mild to severe, based on the physical force causing the trauma. Mild TBI denotes transitory change in consciousness/ mental status while severe TBI indicates prolonged unconsciousness/ amnesia (Centers for Disease Control and Prevention, 2015). TBI results from external physical forces (Leo and McCrea, 2016) such as when the head,

 Is hit by an object, for instance during a fight by bats or fists

 Itself hits stationary objects, for example against windshield in vehicle crashes or against a wall

 Is violently shook, as in Shaken Baby Syndrome (child abuse, CA) or domestic violence (DV)

 Is penetrated, for instance gunshot

It is, however, key to note that not all peripheral forces to the head lead to TBI.

Thus far, the prime causes of TBI are motor vehicle accidents (MVAs), falls, paediatric and adult abuse, sports-related and war-related injuries (Brazinova et al., 2016). Meta-analysis

(4)

4

of European TBI studies show that abusive head trauma (in children) and violence/abuse-related TBI is far less common than TBI caused by MVAs or falls, although the occurrence of the former seems to be increasing over the years (Brazinova et al., 2016). Many population studies have denoted that TBI incidence is generally predominant in early childhood (0-4 years), late adolescence (15-19 years), and in the elderly (above 65 years). TBI from being struck by an object was high in early childhood while TBI from assault was high in late adolescence (Leo and McCrea, 2016).

In 2013, TBI resulting from being struck by or against an object was majorly prevalent in in the United States. Within this category of TBI-mechanism, assault was the chief cause of death in children of 0-4 years while non-fatal TBI was observed in 15-24-year-old people (Centers for Disease Control and Prevention, 2015). A global study showed that Asia had the highest rate of violence-related TBI followed by Europe and developing countries portrayed greater violence-based TBI than developed countries (Li et al., 2016).

A study of 25 European countries showed that Netherlands had very low mortality rates from violence-related TBI incidences while Latvia, Estonia and Lithuania had the highest. Almost no incidences were observed in Turkey, Germany, Austria, Denmark and UK (Majdan et al., 2016). Another study focused on children below 5 years and AHT confirmed by forensic physicians, indicated presence of only 89 cases from 2005 – 2010. These examples suggest the low occurrence of TBI in the Netherlands (Sieswerda-Hoogendoorn et al., 2013).

The different causes of TBI lead to varying pathological changes, one of which is subdural haematoma (SDH). SDH is a type of primary injury caused by TBI. Presently, several researches on SDH shows association between subdural bleeding and physical force-related TBI. SDH is observed commonly in AHT, especially as part of the triad (SDH, retinal bleeding and hypoxic encephalopathy) of the Shaken Baby Syndrome (Wright, 2017). It has also been highly implicated in abuse-related TBI in adults as demonstrated by many studies (Li et al., 2016, Currie et al., 2015 and Adrian et al., 2016).

SDH, and other pathological changes in TBI are considered useful for the detection and dating of TBI, which is important for prognosis, treatment and, even in forensics. Determining the age of the injury is critical to decide whether suspect’s statements are acceptable or refutable and also aid in the events reconstruction, in physical abuse (medicolegal) cases (Finnie, 2016).

This thesis discusses the variety of techniques used for detection and dating of TBI. The pathological features of TBI caused by physical abuse is first described then the different methods for TBI are discussed and compared. Primarily, methods involving SDH are reviewed followed by DAI and general TBI changes. The research question for the thesis is,

“What pros and cons arise from current TBI-based methods and would one sound method suffice for TBI detection (and possibly dating), whilst being acceptable for forensic settings?”

In an attempt to answer this question, a criteria list has been formulated using the most relevant indicators for an apt scientific method and the Daubert guidelines. These guidelines determine the usability of a technique for evidence investigation and use in court. The advantages and disadvantages vary with each method depending on the criteria. Upon comparison of methods and the vast literature investigating these methods, it is evident that

(5)

5

one technique is not always sufficient to determine and date TBI reliably. Rather, a combination of methods would be better in producing accurate results and avoid subjectivity.

PATHOLOGICAL FEATURES OF TBI DUE TO PHYSICAL ABUSE

As described by Xiong, Mahmood and Chopp (2013), TBI involves complicated pathophysiological processes rather than a singular and straightforward process. This dynamic and heterogeneous process begins at time of impact and progresses over time, whereby sequelae are mostly seen even several years post-impact (Currie et al., 2015). TBI can be either closed (blunt/non-missile/non-penetrating) or open (missile/penetrating). The former, typical in physical abuse, does not breach the dura mater whereas in the latter, the scalp, skull, meninges and brain tissue are penetrated. Heterogeneity of brain damage in TBI arises from differing nature and severity of injury, but is basically categorized as primary and secondary insults (Finnie, 2016).

The primary and secondary injury mechanisms account for the eventual structural and functional loss of the brain (Figure 1). At the moment of external mechanical forces impacting the brain, the brain tissue (usually the frontal, parietal and temporal lobes) undergoes immediate disruption. This leads to the primary injuries that are further classified as focal or diffuse injuries (Figure 2):

 Focal injuries: contusions (i.e. damage at the site of impact), lacerations and frequently hematoma like SDH.

 Diffuse injuries: brain swelling, ischemic brain damage and axonal tearing (diffuse axonal injury - DAI)

SDH (focal) and DAI (diffuse) are prevalent and cause high mortality, thus they have been the more commonly studied primary injuries (Hoshizaki et al., 2013). SDH results from shearing of bridging veins, when acceleration-deceleration head movements cause sudden inertial brain movement inside the cranial cavity. DAI results from rotational brain movements causes the shearing of white matter fibres and axons which, with time, progresses into axonal swelling, degeneration and other secondary injuries such as axonal membrane depolarization and cytoskeletal damage. Subdural haematomas can be identified in the acute, subacute, or chronic stages depending on primarily the time of onset and also the severity, location and pathological features. Generally, acute SDH occurs within 2 days, subacute SDH from 2 days to roughly 3 weeks and chronic SDH after 3 weeks of TBI incident. SDH also leads to secondary injury mechanisms (Hoshizaki et al., 2013).

The primary insults evolve and progress into secondary injuries, over a few minutes to months. Mechanisms of secondary insults (type, extent and duration) vary depending on the type of TBI. Nonetheless, all mechanisms end in cellular death, tissue damage and atrophy of the brain. These mechanisms comprise of metabolic, molecular and cellular processes such as mitochondrial dysfunction, free radical production, greater excitatory neurotransmitter release, neuroinflammatory response, cerebral oedema, amyloid plaques and hyper-phosphorylated tau (Zollman, 2016). Additionally, microglia proliferate then migrate to the injury site and secrete cytokines, such as tumour necrosis factor (TNF-α), interleukins 1 and 6 (IL-1, IL-6, in conjunction with the healing processes. These cytokines not only repair and regenerate tissues, they also tend to be neurotoxic (Bogoslovsky et al., 2016).

(6)

6

METHODS

Many methods and techniques for the detection and dating of TBI exist. In this part of the review, the methods concerning SDH are given greater importance followed by DAI and other type of injuries. The methods are grouped based on their principles – histomorphology and immunohistochemistry (HC), biomarkers, imaging and spectroscopy.

I. Histomorphology and IHC

Histomorphological analysis of SDH involves the microscopic observation of pathological changes in the (stained) affected region of the brain, at different time intervals. The observations are utilized in dating the SDH. IHC exploits antigen-antibody interaction to target antigens in tissues using specific antibodies (staining), subsequently, fixing tissues, sectioning and visualizing through a light or fluorescence microscope as in histomorphological analysis.

Only few studies have dated SDH via histomorphology (van den Bos, Zomer and Kubat, 2014; Bokka and Trivedi, 2016; Rao et al., 2016). They fixed the brain tissue with dura mater or specific portions of the affected brain such as the outer membrane of the SDH. Typically, formalin and paraffin were used for fixation and embedding, respectively. 5-mm paraffin sections were subjected to various histological stains - Haematoxylin & Eosin, elastic van Gieson, Perl’s iron stain, Periodic Acid Schiff, Masson’s Trichrome and immunohistochemical stains –CD45, CD34 and CD68 (Table 1). Thereafter, light microscopy was used for observation.

In 2014, van den Bos, Zomer and Kubat developed a simple system for SDH dating using the histomorphological features and IHC. They investigated and upgraded the pre-existing dating system (Walter, Meissner and Oehmichen, 2009). This study provides a checklist (Table 2)for the SDH morphological features that have been clustered according to their appearance-time periods. The results, as well as comparison of estimated vs real post-trauma interval (PTI) proved to be in concert with those of Walter, Meissner and Oehmichen (2009).

The histomorphology of the outer membrane of chronic SDH was investigated by Bokka and Trivedi (2016). They categorized the features as per the intensity of inflammation and haemorrhage, into 4 types. It was concluded that these features could be used for the dating of SDH.

Rao et al (2016) conducted a study similar to van den Bos and team but utilized SDH cases with shorter actual PTI. Despite having some differences in results from Walter, Meissner and Oehmichen (2009), they demonstrated a positive and statistically significant correlation between the occurrence of histomorphological features and duration of PTI. Furthermore, they concluded that histomorphology deems to be the primary technique for SDH dating.

Many studies have utilized IHC for observing the general pathological changes in TBI but not specific to SDH. For example, Wang, Michiue and Mae (2012) discusses the IHC analysis of TBI, particularly neuronal apoptosis using markers such as GFAP, bFGF, S100 and ssDNA. Krohn et al (2015) also observed the potential of S100 in dating TBI, along with NSE through IHC in 47 autopsy cases.

(7)

7

Chen et al (2014) employed whole-slide imaging (WSI) to enable the quantitative histological and immunohistochemical assessment of TBI. They compared the traditional histological techniques with WSI different image analysis algorithms and proved that the latter would be highly efficient and promising when handled and validated properly.

Recently, a study focused on the IHC staining of degenerate neurons in brain injury (Gajavelli et al., 2017). Two rodent models with acute SDH and the markers, UCH-L1 and FJB were used. They detected an inverse correlation between these two markers as UCH-L1 was absent while FJB was present in degenerate neurons. However, UCH-L1 was seen in the extracellular fluid, which caused the authors to propose measurement of duration of brain injury by observing loss of UCH-L1 from degenerating neurons.

II. Biomarkers

A myriad of studies revolves around the usage of biomarkers for the detection of TBI and quantitative assessment of such injuries. Most biomarkers are yet to be validated and FDA-approved for use in adults and children. Table 3 denotes many of the potential biomarkers investigated thus far (Reis et al., 2015) and their targets (Figure 3) (Wang, 2016). These TBI biomarkers are generally classified according to the type of body fluid or the type of cells/tissue they are derived from. A brain injury biomarker should ideally fulfil the listed criteria (taken from Reis et al., 2015):

NUMBER CRITERIA

1 Show a high specificity and sensitivity for brain injury.

2 A passive release from the CNS without any stimulated active release. 3 Lack of specific effects on CNS cells interfering with the initial injury. 4 Stratify patients by severity of injury.

5 Have a rapid appearance in accessible biological fluids. 6 An unlimited passage through the BBB.

7 Provide information about injury mechanisms. 8 Have well defined bio-kinetic properties.

9 Monitor progress of disease and response to treatment. 10 Predict functional outcome.

In 2010, Saatman et al indicated a lack of panels of biomarkers specific to a type of injury (SDH vs contusion, for example) and such panels are still being investigated. Nonetheless, multiple studies have described the potential blood, CSF and serum based biomarkers for varying types of injured brain constituents (astroglia, neurons and axons), the largely studied ones will be discussed below. Some of the discussed biomarkers are associated to DAI (Table 4). SBDPs, UCHL1 and neuron-speciﬁc enolase (NSE) are examples of “neuronal and axonal protein biomarkers” while GFAP and S100B are “glial-specific markers” as categorized by Guingab-Cagmat et al (2013).

S100B (B – Beta), a calcium-binding protein that is plentiful in the astroglia, has been very extensively investigated in TBI. Increased S100B levels were observed as biomarker of acute astroglial injury and for the distinction between mild and severe TBI. The often-cited

(8)

8

studies, in the serum and urine, by Rodríguez-Rodríguez et al (2012) have demonstrated S100B peak at roughly <6hours post-TBI with a steady decrease, till 96hours in serum and till 48hours in urine. In the urine, from 48hours to 96hours, S100B mildly increased. Ercole et al (2016) also showed similar S100B peak in serum roughly at 27.2hours post-TBI. Since S100B is found in oligodendrocytes and non-neural cells (adipocytes, chondrocytes and melanoma cells), the specificity of this marker is affected. Moreover, elevated S100B occurs in traumatic cases without head injuries (Papa et al., 2016).

GFAP is a structural protein found almost exclusively in astrocyte and astroglial cytoskeleton. It is present in both white and gray brain matter. GFAP is an indicator of astroglial injury and has been vastly studied in TBI, wherein latter, it is exhibited to differentiate between mild and severe TBI. GFAP has much greater specificity to brain injuries than other biomarkers such as S100B and NSE, as indicated by many studies. It has also been studied in children with TBI, in which GFAP serum levels peak less than 12hours post-injury (Žurek and Fedora, 2011). In severe-TBI, GFAP levels peaked around 0.5 to 4 hours post-injury while in mild/moderate-TBI, GFAP peaked roughly 20hours post-injury (Papa et al., 2016). The latter study also detected GFAP in serum within 1hour of injury.

Ubiquitin C-terminal hydrolase-L1 (UCH-L1), abundant in neuronal cytoplasm, has recently deemed to be a promising biomarker of neuronal brain injury. Serum UCH-L1 rises within 1hour post-injury and peaks 8hours but readily declines thereafter (Papa et al., 2016). It has also found to be increased in children with moderate/severe TBI and there is good correlation between serum and CSF UCH-L1 levels (Berger et al., 2012).

Neuron-specific enolase (NSE) is an isoenzyme of the glycolytic enzyme enolase located in the cell bodies of the central and peripheral neurons. NSE, having a half-life of 24-48hours, has been expansively investigated as a biomarker of acute TBI. It has shown to be implicated in DAI. Serum NSE peaks approximately 15hours after injury and steadily decreases till 5th_{day (Olivecrona, Bobinski and Koskinen, 2014).}

Spectrin, a cytoskeletal protein, is highly abundant in axons and presynaptic terminals and that maintains membrane and cytoskeleton integrity. Spectrin to spectrin breakdown products (SBDPs) cleavage by caspase and calpain occurs upon cellular injury. SBDPs have been implicated in DAI and increased SBDP levels in serum and CSF have been observed in adults with severe-TBI (Papa et al., 2012). Serum SBDP levels also increased in children with moderate/severe TBI (Berger et al., 2012).

Neurofilaments (NF) are integral heteropolymeric proteins of the axonal cytoskeleton conatining NF light (NF-L), medium (NF-M) and high (NF-H) chains. Upon calcium influx following neuronal injury, NF-H becomes phosphorylated and this pNF-H is specific to axons. Being detectable in blood, pNF-H is a potential biomarker of DAI. Žurek and Fedora, (2011) found high serum pNF-H levels in children with DAI, which continued increasing throughout a 6-day period study. Gatson et al (2014) found highest serum pNF-H in mild-TBI/DAI patients around 4th_{day onwards over a 10-day study. This portrays the different kinetic profile of}

pNF-H whereby it continues to increase post-injury, unlike other biomarkers that peak and decline few days after injury.

The aforementioned biomarkers are more effective in TBI assessment when used in combinations (Reis et al., 2015 and Berger et al., 2012). The use of biomarkers is also extended to aid in imaging of the different types of injuries in TBI. Imaging techniques are expensive

(9)

9

thus a pre-assessment of TBI using biomarkers, prior to selection of imaging technique (if necessary), has been widely agreed upon (Zhang, Puvenna and Janigro, 2016).

A very recent study by Sharma et al (2017) employed this concept to successfully develop a panel of blood-based biomarkers to accurately discriminate mild-TBI cases (with intracranial bleeding vs without intracranial bleeding). They utilized the serum biomarkers creatine kinase B type (CKBB), type IV collagenase (MMP-2) and C-reactive protein (CRP). These biomarkers are involved in neuronal injury, inflammation, blood-brain barrier breakdown, and oxidative stress post-TBI and they were collected within 24 hours of the TBI. The use of such an assay of blood-based biomarkers would offer a minimally invasive and rapid detection of mild-TBI. Thereafter, the ideal neuroimaging method could be decided for determination of injury extent.

III. Imaging

There is a high abundance of studies dedicated to imaging of TBI and the related injury types. Neuroimaging is not done for all TBI cases since it is expensive and might instead be used for patients with more critical conditions. As previously mentioned, biomarkers can help in pre-assessing the injury severity and type. Many types of imaging techniques are available for detection and dating of TBI (Table 5). In this thesis, the recent literature regarding the differing methods of imaging of TBI including those specific to neuroimaging in adult brain injuries and paediatric AHT. Since SDH and DAI are the more common injuries, imaging of these will be of focus (Reis et al., 2015 and Wintermark et al., 2015).

Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) are the frontline methods for imaging TBI (Currie et al., 2015). CT involves the X-ray imaging of specific areas of the body from different angles and generating cross-sectional images of that area via computer processing. MRI involves strong magnetic fields, field gradients and radio waves to generate detailed, structural images of organs, based on the proton content of tissues. For TBI studies and cases, these conventional methods are used for initial assessment of injuries.

CT is the preferred technology for preliminary evaluation of TBI extent and dismiss serious intracranial damages. This is due to the fast results, easy accessibility and high sensitivity for acute haemorrhages of CT. MRI is employed when CT findings are inconsistent with neurological symptoms and for determination of these lesions (Reis et al., 2015).

SDH is classified into 4 stages on CT (Adamsbaum et al., 2014) as tabulated below:

STAGE APPEARANCE APPROXIMATE AGE OF _SDH Hyperacute Isodense Few hours

Acute Hyperdense 3-10 days Subacute Isodense 3 weeks Chronic Hypodense >3 weeks

(10)

10

Many SDH exhibit mixed density as well, particularly in the frontotemporal region. As for MRI, SDH appears differently as per T1 or T2 weighted MRI (Figure 4)(Vázquez et al., 2014):

Vázquez et al (2014) describe the differing features of SDH, caused by AHT, on CT. SDHs are frequently found to spread over hemisphere convexity or present in posterior fossa or inter-hemisphere and might occur opposite to impact site (contrecoup). The authors indicate MRI to have higher sensitivity for SDH membranes (indicator of old SDH) than CT. Prior abuse incidents are signified by chronic SDH, sometimes with extra acute damage. Multiple SDH with varying density are also common and represent repetitive AHT. AHT leads to acute, mixed-density SDH due to the hyperacute-acute blood and blood-CSF mixture coexistence.

It has been also demonstrated that such hypodensities appear on initial CT scans performed as early as 1.2 hours after the report of injury, and become evident within 27 hours in all cases. This finding is not specific for abuse, but seen most often in association with abuse However, other studies have showed that mixed density can also occur from spontaneous rebleeding and does not definitively imply repetitive AHT (Hseih et al and Wright, 2017). Moreover, small SDH (subacute and chronic) are hard to determine as CSF and SDH exhibit similar CT appearance (Wright, 2017).

As such, MRI is deemed to provide more specific dating of SDH. MRI is sensitive to the red blood cell (RBC) degradation and the haemoglobin (Hb) derivatives are used for dating SDH (Figure 5). Caveats also exist for MRI since the SDH signal intensity is influenced by many factors such as size and location of SDH, RBC concentration, RBC degradation rate, sedimentation rate and mixture of CSF-haematoma constituents (Hsieh et al., 2015). This suggests that the age estimation of SDH exclusively via MRI must very carefully be done.

Collectively, CT/MRI evaluation for aging of SDH has proven to be unreliable for forensic usage since Sieswerda-Hoogendoorn et al (2014). demonstrated that time intervals for various appearances of SDH overlap and are too wide. Similarly, Adamsbaum et al (2014) recommended dating (or rather staging) of SDH via CT to be only a descriptive approach as there is no strict protocol for terminology to describe the density and corresponding age of SDH between studies, thereby causing confusing. For example, studies tend to mention “old SDH” and “acute SDH” without indicating the density. They also warned about the subjectivity of the CT/MRI radiologists towards the staging of SDH.

STAGE

APPEARANCE

APPROXIMATE AGE OF SDH

T1 T2

Hyperacute Isointense/hypointense Hyperintense <12 hours Acute Hypointense 12-72 hours Subacute Early Hyperintense Hypointense 3-7 days

Late Hyperintense 7-31 days Chronic Hypointense Hyperintense >1 month

(11)

11

MRI has shown to have higher sensitivity for white matter irregularities, most often DAI (not easily discriminated through CT). There are advanced MRI techniques aimed at increased sensitivity for,

 detecting coexisting functional and structural injury – DTI and fMRI  detecting lesions (haemorrhages etc) – SWI and DWI,

which will be briefly reviewed below (Reis et al., 2015).

Susceptibility-Weighted Imaging (SWI) is 3D, volumetric MRI that enhances contrast between haemorrhagic blood products and adjacent brain tissue by merging phase and magnitude data. Thus, microbleeds and small quantity of blood products can be detected (Vázquez et al., 2014). SWI recognizes brain susceptibility variations and is very sensitive for products of Hb oxidation and iron deposits thereby detecting older shearing TBI (Wright, 2017). Microbleeding is suggestive of DAI, thus SWI can detect DAI. SWI is known to have 3 to 6 times higher sensitivity than conventional MRI for DAI-related haemorrhages (Reis et al., 2015). It has been shown that SWI detected mild-TBI in patients even after a year of injury whereas conventional MRI/CT failed to detect any microbleeds (Reis et al., 2015). DAI with microbleeds has been associated to poor prognosis in AHT patients and detects the microhaemorrhages in 30% of AHT cases (Wright, 2017 and Hsieh et al., 2015).

Diffusion-weighted imaging (DWI) depends on the differing diffusion of water molecules and the disruption of the Brownian motion of water molecules in the brain resulting from injury (Vázquez et al., 2014). The disrupted region appears bright white in DWI. DWI allows detection of DAI, which is difficult to be imaged with conventional MRI, due to the diffuse nature of DAI and the lack of localized lesions (Zhang, Puvenna and Janigro, 2016). DWI allows detection of early (acute) mild-TBI and extent of injury better than conventional T2-MRI (Reis et al., 2015). In most AHT patients, it has been seen that there is watershed pattern of restricted diffusion (indicative of cytotoxic oedema), albeit other patterns of DAI existing in AHT (Wright, 2017).

Diffusion tensor imaging (DTI) relies on the diffusion of water molecules along white matter fibres to determine their structural integrity through 3D images (Zhang, Puvenna and Janigro, 2016). Typically, water molecules diffusion occurs along the axonal length and this directional movement is called anisotropy. Apparent diffusion coefficient (ADC), the average diffusion rate in all directions, and fractional anisotropy (FA), the fraction of entire diffusion, constitute DTI imaging (Zhang, Puvenna and Janigro, 2016). Injured axons are represented by high ADC and low FA. A meta-analysis of 20 non-penetrating TBI studies in children proved DTI as a biomarker of TBI by showing reduced white matter integrity upon mild to severe-paediatric-TBI (Reis et al., 2015). Another meta-analysis denoted that increased FA corresponds to acute mild-TBI while reduced FA corresponds to chronic mild-TBI (Eierud et al., 2014).

Functional MRI (fMRI) depends on the oxygen level in the blood flowing through the brain. In fMRI, patients are instructed to execute particular tasks, for which corresponding blood oxygen level-dependent (BOLD) signals are measured. Difference in OxyHb to DeoxyHB ratio causes highlighting of brain regions that are active during these tasks (Reis et al., 2015). fMRI thus enables determination of abnormal functionality of the brain when injuries deter activation of certain brain areas. Resting state-fMRI (rs-fMRI) allows detection of brain activity at rest. It has been shown that rs-fMRI can be used as mild-TBI radiological

(12)

12

marker during semi-acute phase of injury (<3 weeks post-injury), especially when T1/T2-MRI observations are absent (Wintermark et al., 2015).

IV. Spectroscopy

Apart from the already discussed neuroimaging techniques, advanced imaging of TBI also involves spectroscopy techniques, such as the commonly studied magnetic resonance spectroscopy (MRS) and near-infrared spectroscopy (NIRS).

MRS relies on changes in metabolite levels in the brain upon injury. The atypical ratio of relative concentrations of specific brain-metabolites indicate possible TBI. N-acetylaspartate (NAA), choline (Cho) and creatine (Cr) are the most frequently measured metabolites, while glutamate/glutamine (Glx) and myo-inositol are also used (Reis et al., 2015). Wintermark et al (2015) have reviewed studies that demonstrated extensively reduced NAA (most common observation), higher Cho, enhanced Cr and Glx in the white/gray matter, parenchyma and white matter (lowered Glx in gray matter) respectively, in mild-TBI. However, certain studies have showed no significant change in Cho. In chronic-TBI, they mention about studies also demonstrating reduced NAA and increased Cho. Increased Cho (total) concentrations are considered to possibly be associated to DAI, in child-TBI. Reis et al (2015) indicated combined MRS-DTI measurements leads to highly enhanced sensitivity and specificity for prediction of poor prognosis.

NIRS is based on unique features of brain chromophores in absorption of near-infrared light (NIR). NIR can successfully penetrate deep (certain depth) into the brain, thus allowing the above absorption. The intensity of absorption is detected to give indication of intracranial characteristics and physiology (Davies et al., 2015). The spatial-temporal NIRS dynamics enable this detection thus determining TBI. Each chromophore in the brain has unique absorption characteristics thus the intensity signals can be “unmixed”, as mentioned by Davies et al (2015), to measure the relative quantities of each chromophore in brain. Intracranial haemorrhages have effectively been identified in German army-men with TBI while abnormalities and acute intracranial haemorrhages were correctly detected with high sensitivity in all cases from a study of 21 children with acute intracranial haemorrhages (Sen, Gopinath and Robertson, 2016).

COMPARSION

The various methods discussed above can be compared to each other in many ways – discriminating the fundamental differences or burrowing in-depth to the technical details. This thesis aims to compare the methods based on the fundamental differences and concentrating the comparison to a forensic context for TBI detection. A list of criteria as tabulated below has been made, entailing the major points of comparison for scientific methods (Figure 6), merged with the forensics-related Daubert guidelines for scientific techniques (Fournier, 2016). The methods are discussed as a category (I, II, III or IV) and, also as particular techniques wherever necessary.

(13)

13

CRITERION DESCRIPTION

Validity (rigor of research)/ testability

Have the studies (and methods themselves) achieved the expectations/desired outcomes of TBI detection? Have appropriate experiments and observations been done for that? Are the observations quantitative or qualitative? Accuracy How exactly do the methods detect TBI from non-TBI _{abnormalities?} Precision (reproducible and

repeatable) Are the methods usable easily and do they produce similar results every/most of the time? Sensitivity How easily do the methods detect TBI (even the smallest _{extent of injury)?}

Specificity Do the methods detect only TBI or also other abnormalities/diseases?

Statistical approach and error rate

Are the outcomes/results from using the methods

statistically significant and how prone are they to errors? General acceptance and

peer-review

How well have researchers/clinicians responded to the use of the methods? Have the pros and cons of these methods been acknowledged by these researchers?

Forensic usability Are these methods forensically relevant, usable and how effective have they been for forensic purposes?

Validity (rigor of research)/ testability

All reviewed methods are valid, and most have been validated for the use in TBI detection. Those yet to be validated concern mostly the biomarkers. All studies have employed the renowned scientific method (Figure 11) to determine the effectiveness of these methods in detecting TBI. Relevant research questions have been indicated either passively (Sharma et al., 2017) or directly (Sieswerda-Hoogendoorn et al., 2014) and have thereafter been answered. Most studies have demonstrated the methods to identify TBI as a whole or specific injuries like SDH and DAI while some have had conflicting results. Although the methods have been implicated in detecting TBI successfully, some do not detect certain TBI with equal effectiveness as other methods do, for instance CT vs MRI and biomarkers vs spectroscopy. The other criteria are also crucial for the validity of a method thus it can be generally considered that the method that better fulfils the remaining criteria would be suited for TBI determination (and dating) in a forensic context.

When comparing the type of analysis, it seems that histomorphological evaluation, IHC, conventional CT and MRI are qualitative. SWI tends to be more qualitative than quantitative while the remaining techniques allow both quantitative and qualitative analyses.

(14)

14

Accuracy

A high accuracy is preferred as it represents the method distinguishing TBI from head injuries via other means. The accuracy of methods tends to sometimes vary for detection and dating of TBI. The above methods are all deemed to be at-least mildly accurate in TBI detection – they are useful in determining certain pathological features of TBI like axonal injury or SDH. However, no method has shown to have extremely high accuracy for TBI dating, due to limitations. For example, van den Bos, Zomer and Kubat, (2014) indeed showed that TBI detection and post-trauma interval estimation using IHC are both accurate but the latter was only in comparison with another study. As for biomarkers, S100B has differing accuracy for TBI detection, depending on factors like age of victim/patient and ethnicity. GFAP seems to have better accuracy than UCH-L1 and GFAP is the most accurate compared to other mentioned biomarkers (Bogoslovsky et al., 2016). Contrarily, NSE shows accuracy. Moreover, any other tissue or blood-related markers can contaminate/mask the observation of the desired biomarker. Among imaging techniques, DTI has better accuracy followed by SWI (Papa et al., 2016 and Vázquez et al., 2014). Though CT generally shows good accuracy in TBI detection, recent studies have proved CT and MRI to be inaccurate in TBI dating due to overlapping and wide intervals (Sieswerda-Hoogendoorn et al., 2014). MRS shows good accuracy but the measurements of metabolite ratio tend to be not accurate as the ratio may result from other metabolite changes rather than the metabolite of interest, such as when Cr changes with mild-TBI (Wintermark et al., 2015). Overall, neuroimaging and spectroscopic techniques are evidently more accurate as they target specific and clinically significant features of TBI with better resolution.

Precision

All the discussed methods do require specific technical knowledge to generate good results. Thus, it is key to note that precision also relies partly on the user’s expertise. The easy repeatability of the methods is evident from an array of studies being able to use these methods for investigative studies. The methods have been mentioned by the various studies to be precise with similar results produced upon every use. Although the methods can detect TBI in almost every case, the dating of TBI would differ in terms of results owing to the complexity and heterogeneity of TBI as a process. Hence, precision of methods in TBI dating can be subjective. Moreover, studies investigating these methods do indicate similar results (range of intervals, approximate concentration of markers, characteristic TBI features) being obtained between them but the experimental design may vary (for example sample size). As such, determination of precision of the methods from varying studies with different conditions, is not entirely ideal.

By comparing studies with similar design as prior studies, the study by van den Bos, Zomer and Kubat (2014), showed that histomorphological evaluation of TBI does produce similar results as older studies (for example, Walter, Meissner and Oehmichen, 2009). Similarly, neuroimaging studies denote good precision in identifying TBI as with spectroscopic techniques and biomarkers as well.

Additionally, there are studies that have contradictory or variable results for the methods, when compared to other studies. For instance, Rao et al (2016) found certain histomorphological features to diverge from the studies of Walter, Meissner and Oehmichen, (2009). Some studies showed that UHC-L1 did not discriminate well between TBI and

(15)

non-15

TBI abnormalities (Adrian et al., 2016). No alteration of initial NAA levels was observed in paediatric TBI in one study using MRS (Wintermark et al., 2015).

Sensitivity

Sensitivity and specificity constitute the accuracy of a technique. Sensitivity broadly refers to the capacity of a technique to identify the desired molecules/features. The higher the sensitivity, the better the accuracy and increased positive outcomes of using that method for TBI detection and/or dating.

It has been suggested that a single biomarker has poorer sensitivity compared to usage of a panel of biomarkers (Reis et al., 2015). For instance, S100B might be less sensitive to mild-TBI while NSE has restricted sensitivity, depending on the experimental conditions/ TBI complexity. However, a combination of both had greater sensitivity for TBI detection and evaluation of prognosis. Likewise, combination of GFAP and UCH-L1 displayed increased sensitivity than being used individually. It is also crucial to realize that the mentioned biomarkers, like S100B, might have excellent sensitivity to the general brain abnormalities and pathologies but in terms of TBI, their sensitivity is deemed to be debatably lower. This is due to lack of specificity (Bogoslovsky et al., 2016).

CT is known to be sensitive for intracranial haemorrhages, cerebral oedema, skull fractures and any foreign matter/objects. CT has lower sensitivity to non-haemorrhagic injuries (particularly in acute stage). MRI however is much more sensitive to the extra-axial accumulations, intracranial haemorrhages, contusions and axonal injuries (particularly T2W). T2-weighted-MRI is more sensitive to the haemorrhages during acute and subacute phase than CT. The higher sensitivity of MRI occurs generally in all stages of injuries and even in complicated acute brain injuries (Wintermark et al., 2015). SWI shows very high sensitivity for products of differing stages in the intracranial blood and deoxyHB, in cerebral veins. fMRI also has good sensitivity for neural functioning modifications caused by TBI. The sensitivity of CT and MRI for DAI is low (some studies indicate insensitivity) but DWI compensates for this by having good sensitivity for early DAI. As for MRS and NIRS, high sensitivities have been reported in both adult and paediatric TBI cases. The former showed at-least 70% sensitivity (Reis et al., 2015) while latter showed at-least 80% sensitivity (Sen, Gopinath and Robertson, 2016). Reis et al (2015) furthermore mention studies that demonstrated increased sensitivity (at-least 85%) when DTI measurements were added to MRS.

Specificity

The specificity of a method would be to determine TBI (and characteristic features) in presence of other diseases or abnormalities with similar changes. Histomorphological evaluation and IHC are dependent on the cellular and molecular changes in the brain and affected regions in TBI. These changes have been characterized and studied for many years. Thus, the specificity of these methods in detecting TBI is of acceptable range as long as the observations are made with close attention to the well-accepted list of characterized changes. Immuno-stains and other basic stains themselves have varying specificities depending on the cells and molecules being stained. For example, immuno-stains also show positive results with ischemic hypoxic injury and not just traumatic injury in the brain. Thus, this should also be carefully considered when using IHC and histomorphology for TBI detection and dating.

(16)

16

The specificity seems to be a major issue for the use of biomarkers. This results from the difficulty in determining a single biomarker that is present only in TBI or certain brain injuries like SDH only or DAI only and not in any other brain-related injuries or other tissues. GFAP has thus far has been signified to have highest specificity among other brain-biomarkers owing to its selective presence in astroglial cells. UCH-L1 and pNF-H are also portrayed to have high specificities due to specific and abundant presence in neuronal cell body and axons respectively (Adrian et al., 2016). Contrarily, NSE and S100B are both implicated in many other diseases apart from TBI, other types of brain irregularities and are found in several types of cells (some unrelated to TBI) (Adrian et al., 2016).

The specificity of CT/MRI has been largely debated as the specificities are unsatisfactory when individually used. However, opposing studies indicate CT and MRI have sufficient specificity for initial detection of TBI, where MRI is more specific. Nonetheless, agreement lies in use of CT/MRI along with advanced neuroimaging techniques. The exact specificity of DTI is also under investigation since pattern of DTI measurements (FA and ADC) as well as the modified integrity white matter are not restricted to TBI. Similar situation persists for SWI since patterns of microbleeding are not precluded to TBI. Yet, studies do indicate high specificity of DTI for DAI. A high specificity was also observed of about 90% was observed when DTI was combined with MRS (Zhang, Puvenna and Janigro, 2016).

NIRS, as stand-alone technique has exhibited high specificity of at-least 90% in several studies involving both adults and children. Its specificity relates to detection of SDH and epidural haematoma via Hb saturation (Davies et al., 2015).

Statistical approach and error rates

The statistical approach and error rates differ based on the experimental parameters such as sample size and measurement intervals. Therefore, this criterion will be used to compare the studies with similar parameters and the methods, as generically as possible. Firstly, most studies utilized a small to moderate sample size – as low as 10 to roughly 250 people. Thus, there will be large variation in data across studies, which would affect the weighing of exact statistical power of a method on a general scale. Nonetheless, Rao et al (2016) and van den Bos, Zomer and Kubat (2014) who utilized 100 cases and 89 cases respectively, both indicated statistical significance of their results from using histomorphological features in dating TBI. These studies predominantly show high successful TBI interval estimation rate of at-least 90%. The features displayed distribution-free tolerance interval of more than 90%. There were some outliers, polymorphonuclear leukocytes in Rao et al (2016) and hematoidin in Walter, Meissner and Oehmichen (2009).

Though biomarkers show significant results in TBI detection, they face the issue of false-positives (FP) and false-negative (FN). For example, Kou et al (2013) reported FP for NSE, due to hemolysis while Posti et al (2017) show FP of GFAP and UCH-L1 in orthopaedic trauma patients. S100B was seen to have FN due to being expressed more non-brain tissues/cells.

Su Kim found FN of brain CT in few cases of mild intracranial injury and recommended caution to be exercised even when MRI is used. Wintermark et al., 2015 and Zhang, Puvenna and Janigro (2016) also discussed about the FN and FP in DTI and SWI caused by the

(17)

17

complexity of TBI which act as confounding factors. Eierud et al (2014) mentioned that errors in fMRI result from anisotropy variability within TBI and among protocols as well. Wintermark et al., 2015 moreover indicated the FP and FN of neuroimaging to be caused by artifacts, absence of standardized protocols for using methods as well as gauging technical variations/irregularities. Sen, Gopinath and Robertson (2016) warned about FP and FN in

NIRS, which was shown to have FN and FP in two (separate) of the 21 acute intracranial haemorrhage cases investigated.

General acceptance and peer-review

Numerous studies are build up on previous studies and tend to be extensions of older studies. Hence, for this criterion, the major methods which are branched out from prior ones as well as the peer-review of the methods are discussed. It is apparent from the discussions thus far that the methods have been widely researched upon to be applied in realistic circumstances. The basic principles of each method have been accepted by the scientific community but the standards for data collection and analysis components are still being debated. For instance, the reason for using MRI or biomarkers in TBI detection and dating is valid but the type of features to be focused and the conditions for analysing these features are still under research. This is also evident from the lack of a single well-accepted gold-standard for TBI detection and/or dating.

Limitations and advantages also are critical for general acceptance since there is not a single method with no disadvantages. Some of these advantages and disadvantages, apart from those discussed previously in this thesis, are elaborated below.

The main concern for histomorphological evaluation and IHC as well as CT/MRI is the subjectivity involved when lab technicians and radiologists make visual observations of the results. The absence of a globally-accepted standard for these visual contributions (well-described terminologies for example) contributes to the subjectivity. Similarly, (Fagalde, van Rijn and Aalders (2015) indicated frequent usage of MRI despite the relationship between Hb-derivatives and respective MRI intensities not being verified objectively. Granacher, (2013) suggests that collection of any details about prior TBI and/or other head-related or non-head injuries and diseases would be helpful towards making precise diagnosis. However, unlike what the authors say, these details should be scrutinized only after initial TBI diagnosis to avoid any biasness.

Albeit being the first desired modality, the radiation exposure implicated in CT causes researchers to caution and limit the use of CT to the most relevant of paediatric cases (Zhang, Puvenna and Janigro, 2016). CT is beneficial for its low-cost, easy availability and fast imaging duration. MRI, though being informative, is costly, time-consuming, sensitive to motion of individual and unsuitable for patients with implants and ferromagnetic substances like pacemakers (Wintermark et al., 2015). DTI, DWI, SWI and fMRI are MRI-based methods. DWI has been refined to generate DTI while fMRI is constructed from the functional aspect of MRI (Reis et al., 2015). MRS is also a variant of functional MRI which detects cellular injuries that are otherwise detected only neuropathologically instead of conventional neuroimaging (Wintermark et al., 2015). NIRS is non-invasive as with other neuroimaging methods but fails in 2 ways – no localization of the brain haematomas precisely and no reliable distinction between bilateral haematomas or chronic SDH (Sen, Gopinath and Robertson, 2016).

(18)

18

Forensic usability

In a forensic setting, the results used from these techniques might end up being used as evidence in the court, to corroborate or refute the alleged offence. In this thesis, the focus has been towards detection and dating of TBI caused by physical abuse (in adults and children, more in former). Currently, there is no ideal gold-standard for an efficient method to do the above. As such, forensic scientists continually investigate the methods to improve the reliability and increase possibility of a sound technique. Furthermore, patients with physical abuse are first investigated for diagnosis and treatment instead of diagnosis and forensic outcomes. The Daubert guidelines (testability, peer-review & publication, error rate, general acceptance) incorporated into the comparison criteria in this thesis allows the methods to be evaluated with a forensic perspective.

As with any field, in order to construct a sound method, investigation needs to be done with varying subjects, time intervals and other variables that affect the results. Such repetitive experimentation ensures that the validity and accuracy (constituting the soundness of a technique) of a given technique is certain and not coincidental or an error (Katz and Halámek, 2016). A forensic method should also produce results, efficiently and less tediously as compared to methods used for biomedical or pathological research due to the need to promptly solve cases.

Taking into account these notions and looking at the current use of forensic methods in TBI, CT and MRI have been the most popular techniques. This is due to the relative balance of the advantages and disadvantages of these techniques.

Histomorphological and IHC evaluation of TBI is not widely used more probably due to the underlying subjectivity associated with these methods. Although the accuracy seems to be good for both, other criteria are not significant enough to use these techniques on their own. They could be used as a preliminary or supplementary investigation with other methods.

Biomarkers have already been suggested to be used with imaging techniques, particularly DTI to give more informative and decisive results about TBI presence (Zhang, Puvenna and Janigro, 2016). The quantification of biomarker levels would be useful in determining the extent of different types of injuries. The tissue-specific biomarkers such as UCH-L1 and GFAP would provide more accurate discrimination of injury types. However, a single biomarker is not advised thus validation of sound and effective panel of biomarkers for specific injuries would be of better advantage in forensics. This will save time in TBI identification and characterization.

As formerly mentioned, CT/MRI have been the key choices of imaging modalities despite their flaws. It had been investigated very extensively in forensic pathology, radiology and even law. Granacher (2013) quotes CT to be an invaluable instrument in forensics – for the detection and indication of brain injuries, detection of gas and foreign bodies as well. A similar praise exists for MRI. These techniques also enable 3D reconstructions of TBI that will certainly increase the possibility of imaging evidence being used in court. The members of the court would have easier visualization of the location and extent of injuries while it tends to be easier for expert witnesses to explain their conclusions. However, Granacher (2013) also denotes that radiologists conducting the CT/MRI analyses are not well-studied about forensic

(19)

19

implications of the results. This, as per the authors, is a crucial issue when neuroimaging is applied to forensic.

Likewise, Sieswerda-Hoogendoorn et al (2014) describes this issue with regards to SDH determination and dating. The authors conducted a survey among radiologists, whereby the latter had to answer questions concerning SDH estimation with certainty. The results portrayed variation in the approach of radiologists, which the authors trace back to the absence of proper medical literature or absence of good studies regarding SDH dating. This suggests that apart from the power of efficient utility and accurate results, a forensic technique needs to be handled by skilled technicians to avoid blunders and also technicians who do not fear reporting inconclusive results.

The remaining methods of imaging as well as spectroscopy tend to be promising platforms for TBI detection and dating. Nonetheless, these techniques have only recently been advocated to be used in court, like DTI in the US, while some forensic scientists question their usability in non-clinical cases (Granacher, 2013). As such, more studies should evaluate the effectiveness of these techniques in TBI-related forensic cases. Based on this thesis, it seems more logical to use combinations of these techniques for example, DTI and SWI since the corresponding strengths increase the overall accuracy of TBI detection and the disadvantages may get relatively negated.

CONCLUSION

TBI is seemingly a serious global problem and physical abuse is one of the many causes of TBI. There are different types of TBI injuries, based on the impacting object/force and the effect of such an impact. In forensic contexts, physical abuse and the TBI resulting from it has been researched widely. There are multiple methods to detect and even estimate the age of the TBI. This thesis focused on the primary injuries of TBI, with special consideration of SDH in certain methods. The predominant research question of “What pros and cons arise from current TBI-based methods and would one sound method suffice for TBI detection (and possibly dating), whilst being acceptable for forensic settings?”, was explored through the comparison of the various methods.

For the first part of this question, the formulated list of criteria delves into the advantages and disadvantages provided by the differing TBI detection/dating methods. The chosen comparison criteria enabled the methods to be evaluated based on critical factors entailing any scientific method and moreover, in a forensic aspect (via Daubert guidelines). It is quite evident that second part of this question cannot be concretely answered based on studies that have varying parameters and differing protocols for investigation. This has also been stated by Finnie (2016), who discusses the difference in type of models (animal/human), variables (exact extent of injury for e.g.,) and resulting TBI intervals being unreliable for real cases. It is unrealistic Moreover, the heterogeneity and complexity of TBI makes it difficult for a single method to very precisely detect and further, date the TBI. In view of this, and as proposed by many of the reviewed studies, the usage of combination of methods would be much better.

For instance, histomorphology and MRI, biomarkers and DTI, or combination of different markers can be utilized to provide more holistic, informative and accurate results. This also reduces the time and expenses for failing to produce results with one technique and then investing in another one, separately. There should more research dedicated to using combined methods in TBI cases, especially in a forensic setting.

(20)

20

REFERENCES

1. Adamsbaum, C., Morel, B., Ducot, B., Antoni, G. and Rey-Salmon, C. (2014). Dating the abusive head trauma episode and perpetrator statements: key points for imaging. Pediatric Radiology, 44(S4), pp.578-588.

2. Adrian, H., Marten, K., Salla, N. and Lasse, V. (2016). Biomarkers of Traumatic Brain Injury: Temporal Changes in Body Fluids. eNeuro, 3(6).

3. Berger, R., Hayes, R., Richichi, R., Beers, S. and Wang, K. (2012). Serum Concentrations of Ubiquitin C-Terminal Hydrolase-L1 and αII-Spectrin Breakdown Product 145 kDa Correlate with Outcome after Pediatric TBI. Journal of Neurotrauma, 29(1), pp.162-167.

4. Bogoslovsky, T., Gill, J., Jeromin, A., Davis, C. and Diaz-Arrastia, R. (2016). Fluid Biomarkers of Traumatic Brain Injury and Intended Context of Use. Diagnostics, 6(4), p.37.

5. Bokka, S. and Trivedi, A. (2016). Histopathological study of the outer membrane of the dura mater in chronic sub dural hematoma: Its clinical and radiological correlation. Asian Journal of Neurosurgery, 11(1), p.34.

6. Brazinova, A., Rehorcikova, V., Taylor, M., Buckova, V., Majdan, M., Psota, M., Peeters, W., Feigin, V., Theadom, A., Holkovic, L. and Synnot, A. (2016). Epidemiology of Traumatic Brain Injury in Europe: A Living Systematic Review. Journal of Neurotrauma.

7. Centers for Disease Control and Prevention (2015). Report to Congress on Traumatic Brain Injury in the United States: Epidemiology and Rehabilitation. [online] Atlanta, GA.: National Center for Injury Prevention and Control; Division of Unintentional Injury Prevention.

Available at:

https://www.cdc.gov/traumaticbraininjury/pdf/tbi_report_to_congress_epi_and_rehab-a.pdf. 8. Chen, Z., Shin, D., Chen, S., Mikhail, K., Hadass, O., Tomlison, B., Korkin, D., Shyu, C., Cui,

J., Anthony, D. and Gu, Z. (2014). Histological Quantitation of Brain Injury Using Whole Slide Imaging: A Pilot Validation Study in Mice. PLoS ONE, 9(3), p.e92133.

9. Currie, S., Saleem, N., Straiton, J., Macmullen-Price, J., Warren, D. and Craven, I. (2015). Imaging assessment of traumatic brain injury. Postgraduate Medical Journal, 92(1083), pp.41-50.

10. Davies, D., Su, Z., Clancy, M., Lucas, S., Dehghani, H., Logan, A. and Belli, A. (2015). Near-Infrared Spectroscopy in the Monitoring of Adult Traumatic Brain Injury: A Review. Journal of Neurotrauma, 32(13), pp.933-941.

11. Eierud, C., Craddock, R., Fletcher, S., Aulakh, M., King-Casas, B., Kuehl, D. and LaConte, S. (2014). Neuroimaging after mild traumatic brain injury: Review and meta-analysis. NeuroImage: Clinical, 4, pp.283-294.

12. Ercole, A., Thelin, E., Holst, A., Bellander, B. and Nelson, D. (2016). Kinetic modelling of serum S100b after traumatic brain injury. BMC Neurology, 16(1).

13. Fagalde, K., van Rijn, R. and Aalders, M. (2015). Dating of Traumatic Brain Injury Spectroscopic measurements and Magnetic Resonance Images of hemoglobin derivatives in Subdural Hematomas, a phantom study. International Journal of Legal Medicine, 18.

14. Finnie, J. (2016). Forensic Pathology of Traumatic Brain Injury. Veterinary Pathology, 53(5), pp.962-978.

15. Fournier, L. (2016). The Daubert Guidelines: Usefulness, Utilization, and Suggestions for Improving Quality Control. Journal of Applied Research in Memory and Cognition, 5(3), pp.308-313.

(21)

21 16. Gajavelli, S., Bregy, A., Spurlock, M., Diaz, D., Burks, S., Bomberger, C., J., C., Yokobori, S., Diaz, J., Sanchez-Chavez, J. and Bullock, R. (2017). Immunohistochemical Correlation of Novel Biomarkers with Neurodegeneration in Rat Models of Brain Injury. Applications of Immunocytochemistry.

17. Gatson, J., Barillas, J., Hynan, L., Diaz-Arrastia, R., Wolf, S. and Minei, J. (2014). Detection of neurofilament-H in serum as a diagnostic tool to predict injury severity in patients who have suffered mild traumatic brain injury. Journal of Neurosurgery, 121(5), pp.1232-1238.

18. Granacher, R. (2013). Forensic Issues in the Structural or Functional Neuroimaging of Traumatic Brain Injury. Novel Frontiers of Advanced Neuroimaging.

19. Guingab-Cagmat, J., Cagmat, E., Hayes, R. and Anagli, J. (2013). Integration of Proteomics, Bioinformatics, and Systems Biology in Traumatic Brain Injury Biomarker Discovery. Frontiers in Neurology, 4.

20. Hoshizaki, B., Post, A., Kendall2, Clara Karton2 and Susan Brien3, M., Karton, C. and Brien, S. (2013). The Relationship between Head Impact Characteristics and Brain Trauma. Journal of Neurology & Neurophysiology, 05(01).

21. Hsieh, K., Zimmerman, R., Kao, H. and Chen, C. (2015). Revisiting Neuroimaging of Abusive Head Trauma in Infants and Young Children. American Journal of Roentgenology, 204(5), pp.944-952.

22. Katz, E. and Halámek, J. (2016). Forensic Science: A Multidisciplinary Approach. Weinheim: Wiley-VCH Verlag.

23. Kou, Z., Gattu, R., Kobeissy, F., Welch, R., O’Neil, B., Woodard, J., Ayaz, S., Kulek, A., Kas-Shamoun, R., Mika, V., Zuk, C., Tomasello, F. and Mondello, S. (2013). Combining Biochemical and Imaging Markers to Improve Diagnosis and Characterization of Mild Traumatic Brain Injury in the Acute Setting: Results from a Pilot Study. PLoS ONE, 8(11), p.e80296.

24. Krohn, M., Dreßler, J., Bauer, M., Schober, K., Franke, H. and Ondruschka, B. (2015). Immunohistochemical Investigation of S100 and NSE in Cases of Traumatic Brain Injury and Its Application for Survival Time Determination. Journal of Neurotrauma, 32(7), pp.430-440. 25. Leo, P. and McCrea, M. (2016). Translational Research in Traumatic Brain Injury. [ebook] CRC Press/Taylor and Francis Group, p.Chapter 1. Available at: https://www.ncbi.nlm.nih.gov/books/NBK326730/.

26. Li, M., Zhao, Z., Yu, G. and Zhang, J. (2016). Epidemiology of Traumatic Brain Injury over the World: A Systematic Review. General Medicine: Open Access, 04(05).

27. Majdan, M., Plancikova, D., Brazinova, A., Rusnak, M., Nieboer, D., Feigin, V. and Maas, A. (2016). Epidemiology of traumatic brain injuries in Europe: a cross-sectional analysis. The Lancet Public Health, 1(2), pp.e76-e83.

28. Olivecrona, Z., Bobinski, L. and Koskinen, L. (2014). Association of ICP, CPP, CT findings and S-100B and NSE in severe traumatic head injury. Prognostic value of the biomarkers. Brain Injury, 29(4), pp.446-454.

29. Papa, L., Brophy, G., Welch, R., Lewis, L., Braga, C., Tan, C., Ameli, N., Lopez, M., Haeussler, C., Mendez Giordano, D., Silvestri, S., Giordano, P., Weber, K., Hill-Pryor, C. and Hack, D. (2016). Time Course and Diagnostic Accuracy of Glial and Neuronal Blood Biomarkers GFAP and UCH-L1 in a Large Cohort of Trauma Patients With and Without Mild Traumatic Brain Injury. JAMA Neurology, 73(5), p.551.

30. Papa, L., Lewis, L., Silvestri, S., Falk, J., Giordano, P., Brophy, G., Demery, J., Liu, M., Mo, J., Akinyi, L., Mondello, S., Schmid, K., Robertson, C., Tortella, F., Hayes, R. and Wang, K. (2012). Serum levels of ubiquitin C-terminal hydrolase distinguish mild traumatic brain injury

(22)

22 from trauma controls and are elevated in mild and moderate traumatic brain injury patients with intracranial lesions and neurosurgical intervention. The Journal of Trauma and Acute Care Surgery, 72(5), pp.1335-1344.

31. Posti, J., Hossain, I., Takala, R., Liedes, H., Newcombe, V., Outtrim, J., Katila, A., Frantzén, J., Ala-Seppälä, H., Coles, J., Kyllönen, A., Maanpää, H., Tallus, J., Hutchinson, P., van Gils, M., Menon, D., Tenovuo, O. and TBIcare Investigators (2017). Glial Fibrillary Acidic Protein and Ubiquitin C-Terminal Hydrolase-L1 Are Not Specific Biomarkers for Mild CT-Negative Traumatic Brain Injury. Journal of Neurotrauma, 34(7), pp.1427-1438.

32. Rao, M., Singh, D., Vashista, R. and Sharma, S. (2016). Dating of Acute and Subacute Subdural Haemorrhage: A Histo-Pathological Study. JOURNAL OF CLINICAL AND DIAGNOSTIC RESEARCH, 10(7), pp.HC01-HC07.

33. Reis, C., Wang, Y., Akyol, O., Ho, W., II, R., Stier, G., Martin, R. and Zhang, J. (2015). What’s New in Traumatic Brain Injury: Update on Tracking, Monitoring and Treatment. International Journal of Molecular Sciences, 16(6), pp.11903-11965.

34. Rodríguez-Rodríguez, A., Egea-Guerrero, J., León-Justel, A., Gordillo-Escobar, E., Revuelto-Rey, J., Vilches-Arenas, Á., Carrillo-Vico, A., Domínguez-Roldán, J., Murillo-Cabezas, F. and Guerrero, J. (2012). Role of S100B protein in urine and serum as an early predictor of mortality after severe traumatic brain injury in adults. Clinica Chimica Acta, 414, pp.228-233.

35. Saatman, K., Duhaime, A., Bullock, R., Maas, A., Valadka, A. and Manley, G. (2010). Classification of Traumatic Brain Injury for Targeted Therapies. Journal of Neurotrauma, 25(7), pp.719-738.

36. Sen, A., Gopinath, S. and Robertson, C. (2016). Clinical application of near-infrared spectroscopy in patients with traumatic brain injury: a review of the progress of the field. Neurophotonics, 3(3), p.031409.

37. Sharma, R., Rosenberg, A., Bennett, E., Laskowitz, D. and Acheson, S. (2017). A blood-based biomarker panel to risk-stratify mild traumatic brain injury. PLOS ONE, 12(3), p.e0173798. 38. Sieswerda-Hoogendoorn, T., Bilo, R., van Duurling, L., Karst, W., Maaskant, J., van Aalderen,

W. and van Rijn, R. (2013). Abusive head trauma in young children in the Netherlands: evidence for multiple incidents of abuse. Acta Paediatrica.

39. Sieswerda-Hoogendoorn, T., Postema, F., Verbaan, D., Majoie, C. and van Rijn, R. (2014). Age determination of subdural hematomas with CT and MRI: A systematic review. European Journal of Radiology, 83(7), pp.1257-1268.

40. van den Bos, D., Zomer, S. and Kubat, B. (2014). Dare to date: age estimation of subdural hematomas, literature, and case analysis. International Journal of Legal Medicine, 128(4), pp.631-640.

41. Vázquez, E., Delgado, I., Sánchez-Montañez, A., Fábrega, A., Cano, P. and Martín, N. (2014). Imaging abusive head trauma: why use both computed tomography and magnetic resonance imaging?. Pediatric Radiology, 44(S4), pp.589-603.

42. Walter, T., Meissner, C. and Oehmichen, M. (2009). Pathomorphological staging of subdural hemorrhages: Statistical analysis of posttraumatic histomorphological alterations. Legal Medicine, 11, pp.S56-S62.

43. Wang, Q., Michiue, T. and Mae, H. (2012). Immunohistochemistry of Neuronal Apoptosis in Fatal Traumas: The Contribution of Forensic Molecular Pathology in Medical Science. Apoptosis and Medicine.

44. Wintermark, M., Sanelli, P., Anzai, Y., Tsiouris, A., Whitlow, C., Druzgal, T., Gean, A., Lui, Y., Norbash, A., Raji, C., Wright, D. and Zeineh, M. (2015). Imaging Evidence and

(23)

23 Recommendations for Traumatic Brain Injury: Conventional Neuroimaging Techniques. Journal of the American College of Radiology, 12(2), pp.e1-e14.

45. Wright, J. (2017). CNS Injuries in Abusive Head Trauma. American Journal of Roentgenology, 208(5), pp.991-1001.

46. Xiong, Y., Mahmood, A. and Chopp, M. (2013). Animal models of traumatic brain injury. Nature Reviews Neuroscience, 14(2), pp.128-142.

47. Zhang, J., Puvenna, V. and Janigro, D. (2016). Biomarkers of Traumatic Brain Injury and Their Relationship to Pathology. Boca Raton: CRC Press.

48. Zollman, F. (2016). Manual of traumatic brain injury management. New York, NY: Springer Publishing Company, pp.10-13.

49. Žurek, J. and Fedora, M. (2011). The usefulness of S100B, NSE, GFAP, NF-H, secretagogin and Hsp70 as a predictive biomarker of outcome in children with traumatic brain injury. Acta Neurochirurgica, 154(1), pp.93-103.

IMAGES:

1. Carolina Biological Supply Company (2017). Infographic: The Scientific Method and NGSS Dimension 1: Scientific Practices. [image] Available at: https://www.carolina.com/teacher-

resources/Interactive/the-scientific-method-and-ngss-imension-1-scientific-practices/tr41419.tr.

2. Hill, C., Coleman, M. and Menon, D. (2016). Figure 1 Traumatic Brain Injury Subtypes.. [image] Available at: http://www.cell.com/trends/neurosciences/fulltext/S0166-2236(16)00050-3.

3. Scheau, C., Ghergus, A., Popa, G., Preda, E., Capsa, R. and Lupescu, I. (2014). Fig. 1: Biochemistry of bleeding. Evolution of a typical cerebral hematoma.. [image] Available at:

http://www.myesr.org.

4. Wang, K. (2016). TBI Blood-Based Biomarkers: Impacts on diagnosis, prognosis and therapy

development. [online] Available at:

https://www.fda.gov/downloads/MedicalDevices/NewsEvents/WorkshopsConferences/UCM 490909.pdf.

5. Weegenaar, C. (n.d.). T1 vs.T2 weighted MRI. [image] Available at:

https://www.ole.bris.ac.uk/bbcswebdav/institution/Faculty%20of%20Health%20Sciences/MB

(24)

24

APPENDIX

Figure 1: Classification of TBI injury mechanisms leading to functional loss in brain. Taken from Xiong, Mahmood and Chopp, 2013.

Figure 2: Different types of TBI-caused primary injury. Taken from Hill, Coleman and Menon, 2016.

(25)

25

Figure 4: Differences between T1-weighted and T2-weighted MRI. Adapted from www.ole.bris.ac.uk.

Figure 5: Hb derivatives and their presence in progression of haematoma. Taken from Scheau et al, 2014.

Figure 6: Components of a good scientific method. Taken from Carolina Biological Supply Company, 2017.

STUDY BY STAINS USED van den Bos, Zomer and

Kubat, 2014

Haematoxylin-eosin (H&E), elastic van Gieson (EvG), Perl’s iron, CD34, CD45 CD68

Bokka and Trivedi, 2016 H&E and EvG

Rao et al, 2016 H&E, Perl’s stain, Masson’s Trichrome, Periodic Acid Schiff (PAS)