K.N.R. Fa ga lde
Dept of Biomedical Engineering a nd Phys ics, Aca demic Medical Center, Amsterdam MSc. Forensic Science, University of Amsterdam, Amsterdam
: k.n.kalapnathsing@amc.uva.nl Prof. M.C.G. Aa l ders (Supervisor)
Dept of Biomedical Engineering a nd Phys ics, Aca demic Medical Center, Amsterdam Prof. R. va n Ri jn (Examiner)
Dept. Of Ra diology, Aca demic Medical Center, Amsterdam O RIGINAL ARTICLE
Dating of Traumatic Brain Injury
Spectroscopic measurements and Magnetic Resonance Images of
hemoglobin derivatives in Subdural Hematomas, a phantom study
Kim N.R. Fagalde
∙
Rick R. van Rijn
∙
Maurice C. Aalders
Period: February-November 2015
Received: 2 November 2015 / Revised: 10 November 2015 / Accepted: 23 November 2015 ©Springer – Verlag Berlin Heidelberg 2015
Abstract
Traumatic brain injury (TBI) is associated with physical abuse. There is an urgent need to date TBI since this will provide important and relevant information. One way of doing this is to determine the age of Subdural Hematomas (SDH) which are frequently formed as a consequence of TBI. Several forensic studies attempted to do this by determining the intensity of SDH on Magnetic Reasonance (MR) Imaging. However, this method is proven to be unreliable and highly subjective. Hence, there is currently no reliable and objective method to date TBI by estimating the age of SDH. This preliminary study aimed for a theoretical underpinning of the aging process of SDH by measuring the Hb derivatives and obtain MR images of phantom SDH’s. A strong correlation was found between the Hb derivatives and relaxation time, which is a quantitative measure of the intensity on MR images. Moreover, mice models show that further research should be performed on them due to the promising results.Keywords
Subdural Hematomas∙
Traumatic Brain Injury∙
Hemoglobin Derivatives∙
Reflectance Spectroscopy∙
Magnetic Resonance ImagingIntroduction
Severe cases of Traumatic Brain Injury (TBI) are the leading cause of death worldwide [1], and TBI is associated with physical abuse [2]. Physical abuse is a huge social problem, affecting every age group. Indeed, surveys performed in 2012 and 2014 in the Netherlands showed that one out of 20 elderlies is
victim of domestic abuse and in 2012 there were 431 physical abuse cases. Furthermore, every year there are 130.000 individuals between 18-70 years old that are victim of domestic physical abuse [3]–[5].
Page | 2 Moreover, a Subdural Hematoma (SDH), of the
physically abused children whom were diagnosed with head trauma, was described in 77%-89% (20/26, Estonia [6]; 41/46, Switzerland [7]) when examined with neuroimaging. With autopsy a SDH was observed in 83%-92% (24/29, USA [8]; 12/13, USA, [9]) of the children diagnosed with head trauma due to abuse. However, the main concern is that almost a quarter of the cases regarding physical abuse of children, is dismissed by the OM due a lack of evidence (24% of 178 cases [10]). Also a concern is the low percentage of arrests in adult physical abuse cases (18% of 87.323 reports). Although not investigated, the reason for this is perhaps also due to a lack of evidence [3]. Objective evidence provided by a reliable method, in which an evidence-based time profile of TBI is obtained, is essential for the forensic-medical field because this can lead to a suspect and/or exonerates an innocent, and aid the reconstruction of activities. Thus, this information can be used in court cases.
Indeed, the need for such evidence is acknowledged by the forensic-medical field. For instance, researchers have studied whether the T1* and T2* Magnetic Resonance (MR) Images can be used to determine the age of SDH [11]–[15] in order to date TBI [16], [17]. To obtain a better understanding of the method, a short description of the MR image properties need to be given. The T1* MR images represents spin-latice relaxation and is influenced by the dipole-dipole interaction between Proton (of hydrogen) and Electron (of iron in hemoglobin in our case), which is called the PEDD interaction. To facilitate this, the protons and electrons have to be in close proximity of each other (within 3Ǻ, i.e. 3E-10 m). The shorter the T1* relaxation time is, the better the energy transfer is between the protons and electrons, which can be due to the PEDD interaction, which is visible as a hyperintense signal [15]. On the other hand, the T2* MR image depends on the spin-spinrelaxation and is influenced by the magnetic susceptibility of a tissue, i.e. the magnetic properties of a tissue depend on the presence of unpaired electrons. The shorter the T2* relaxation time is, the faster the protons defase due to local magnetic fields, which is caused by the magnetic properties, which is visible as a hypointense signal [13]. Hemoglobin (Hb), which is often imaged with MR, consists of a heme and a
globin part and is the oxygen carrier in blood. Importantly, it has derivatives (see Figure 1), some enable PEDD interaction and contain unpaired electrons [15]. As such, different stages of hemorrhage have different intensities viewed on MR images [11], [15], e.g. a hyperacute hemorrhage (<12h) will appear iso/hypointense on a T1* and hyperintense on a T2* MR image due to the presence of the diamagnetic HbO2. Subsequently, an acute hemorrhage (12-72h) will appear hypointense on T1* and T2* MR images due to the presence of the paramagnetic Hb. A subacute hemorrhage can be either an early subacute (3-7 days) hemorrhage, which will appear hyperintense on T1* and hypointense on a T2* MR image due to the paramagnetic intracellular methemoglobin (metHb), or a late subacute (7-31 days) hemorrhage, which will appear hyperintense on T1* and T2* MR images due to the paramagnetic extracellular metHb. Finally, when the hemorrhage is chronic (>1 month), it will appear hypointense on the T1* and hyperintense on the T2* MR images due to the diamagnetic HC. When rebleeding occurs, the paramagnetic hemosiderin could also be present in the rim of a SDH, which will appear hypointense on T1* and T2* MR images [15]. Even though the relation between the Hb derivatives and the corresponding intensities visible on a MR image is assumed rather than objectively tested, it is still frequently used by medical professionals [11]–[15].
However, Sieswerda-Hoogendoorn et al. [18] showed that the time intervals of the SDH are too wide and overlapping, and based on these phenomena concluded that this method is not reliable enough for forensic cases. Postema et al. [19] showed that the radiologists are too subjective (high variation and uncertainty) and are therefore not consistent in the age determination. Since physical abuse is a major social problem, (severe) TBI is the leading cause of death worldwide and at the moment there are limited methods, a novel and reliable dating method is extremely essential. The current study will theoretical underpin the aging of SDH by measuring the Hb derivatives quantitatively with reflectance spectroscopy which is a well-developed method by the AMC forensic photonics group [20] to provide a temporal profil, and correlate the Hb derivatives with the relaxation times of the T1* and T2* MR images.
Page | 3 In summary, this paper contains the
preliminary results of a pilot study in order to develop a revolutionary approach to TBI by examining the temporal profile of phantom SDH. The first question investigated is: What is the quantitative temporal profile of the Hb derivatives in phantom SDH. The second question investigated is, is there a relation between the Hb derivatives and the T1* and T2* relaxation times? Lastly preliminary results of a small study in which the Hb derivatives in a TBI mouse and non-TBI mice models are investigated. Even though the results seem promising, the protocol needs to be optimized and the measurements have to be performed again.
Methods
Study design
This study was performed according to three consecutive steps in order to answer the two previously dictated research questions. The steps consisted of optimization of the reflectanc e spectroscopy measurement protocol, measuring the temporal profile of Hb derivatives of the phantom SDH’s, and measure the relaxation times of the phantom SDH’s in order to correlate them with the Hb derivatives. All the measurements were performed at room temperature in the dark. Approval of the METC of the AMC in the Netherlands was obtained to include TBI patients whom are diagnosed with chronic SDH in a future study (not published yet).
Optimization of reflectance spectroscopy measurement protocol
The purpose of this step was to optimize the measurement protocol for measuring SDH by establishing the best settings such as: distance (between the probe and sample), interval time (time between emitting and receiving light), drying time (time between deposition and measuring), clot forming time (time between collecting blood and clot formation), substrate, background, form of sample, source of the sample (i.e. living or deceased persons). the condition with the highest goodness of fit (R²) was chosen. The goodness of fit can be used in order to determine how accurate the
measurements resemble the theoretical outcomes of the measurements (which is developed by Edelman et al. [20]). The maximum value for R² is 1 and represents a very good resemblance, while a minimum value of 0 for R² represents a very poor resemblance. The spectra are obtained by measuring the reflectance between 500-700 nm.
The distance depended on the type of samples that were measured. The distance for measuring blood that was deposited onto a white cotton cloth (bloodstains) in duplo and measured during multiple time points (up till a period of 267 hours) was fixed at 0.9 mm [21], but the interval time was free. The drying time was varied between 10-40 minutes (measured every 10 minutes). The distance for measuring phantom SDH’s was fixed at 0.9 mm, however, the interval time was not fixed. The drying time was not varied and kept constant at 30 minutes [21]. The condition of the clot was examined at different clot forming times (2,5, 20, 93 and 96 hours).
Phantom SDH’s were first created by allowing blood (from one healthy participant) that was drawn and collected, to clot in blood tubes (1.25 ml/tube). Thereby creating five simulated SDH’s. Subsequently, different forms of the simulated SDH’s were prepared, which included whole SDH, thick slices, thin slices, a third, and smeared onto a substrate. Furthermore, the types of substrates were also varied and resulted in testing four substrates such as glass, plastic, petri-dish and
HbO2 metHb “Reversible” HC “Irreversible” HC Hb
Figure 1. Adapted from wintrobe’s clinical
hematology [23]. The assumed aging scheme of Hb and its derivatives in blood. Reversible HC can be converted back into metHb, while irreversible HC cannot.
Page | 4 white cotton. Since the background can also have an
effect on the reflectance signal, three different backgrounds were tested such as white cotton, white paper and black paper. This process resulted in different conditions that were tested (some samples were too dense and therefore further tests were not possible, see appendix A Table 3). Regardless the condition, all the samples were measured 30 minutes after deposition. Furthermore, the distance as well as the interval time were kept free.
Another step in the optimization of the measurement protocol was measuring blood of a deceased and living persons. As such, there were blood sample of a deceased person (n=1) and blood samples of living persons (n=5) all taken from the antecubital area of the arm (either the basilica, cephalic or medial cubital veins). The blood sample of the deceased person was deposited in duplo on a white cotton cloth, and 30 minutes afterwards the blood stains were measured. Blood samples of the healthy persons that were deposited in duplo on a white cotton cloth were measured 2,5 h after deposition. The distance was fixed at 9 mm, and the interval time was kept free.
Temporal profile and correlations of phantom SDH In order to measure the temporal profile of phantom SDH’s, blood of a healthy participant (n=1) was drawn and collected into four tubes (1.25 ml/tube) after which the blood was allowed to clot. The age of the phantom SDH varied between 2,5 hours and 91 days (see
Table
1
). By having different ages of the phantom SDH’s, different stages of hemorrhagecould be simulated: one hyperacute (number 1), one acute (number 2), one early subacute (number 3), four late subacute (number 4-7) and seven chronic (number 8-14). The Hb derivatives of the phantom SDH were measured with reflectance spectroscopy and the relaxation times was determined with MR images. The phantom SDH’s in the first tubes (gently shaken, to mimic a living person) and the phantom SDH’s in the second tubes (not shaken, to mimic a deceased person) were used (see Figure 3). Subsequently, these tubes were placed in green foam (see Figure 2) and placed into a 3T unit to make MR images. The phantom SDH’s in the third tubes were used to measure the Hb derivatives by smearing the phantom SDH’s onto a white cotton cloth with the same measurement settings according to the optimized measurement protocol.
Table 1. Overview of the ages of the created phantom SDH’s used in this study
N U MBER 1 2 3 4 5 6 7 8 9 10 11 12 13 14
A G E Days 0,1 1 3 7 14 21 25 35 42 49 56 63 67 91
Hours 2,5 24 72 168 336 504 600 840 1008 1176 1344 1512 1608 2184
Figure 3. Phantom SDH’s of different ages in tubes. The tubes on the left image were shaken, the tubes on the right image were not shaken.
Figure 2. Overview of the set up for the MR image: tubes with phantom SDH were inserted in foam
Page | 5 Small study: Distinguish TBI from non-TBI
The mice models that were used to examine the Hb derivatives in blood of TBI (n=1) and non-TBI (n=2), were prepared in a previous study [22]. Arterial blood was drawn and the subsequent plasma and serum samples were stored at -80°C. Serum and plasma stains were created on a white cotton cloth, and the measurement settings were according to the bloodstains of the optimized measurement protocol.
Measurement setup
The set up (see Figure 4) for the reflectanc e spectroscopy measurements consisted of one non-contact probe (QR400-7-UV/BX, Ocean Optics, Duiven, The Netherlands) with six delivery fibers (400 µm) surrounding a similar central collection fiber. The delivery fibers and a tungsten halogen light source (H-2000, Ocean Optics, Duiven, The Netherlands) are connected to each other, and the collection fiber is connected to the spectrograph (USB 4000; Ocean Optics, Duiven, The Netherlands). The fraction of the Hb derivatives were determined with a FTS program containing a logarithm developed by Edelman et al. [20], between 500-700 nm. The MR images were made with a 3T unit (Philips 3T Ingenia), the imaging protocol included a 2D-T1*-weighted sequence (TR 10 sec, TE 80 msec), 2D-T2*-weighted gradient-echo sequence (TR 1 sec, TE 7 msec), with a head coil, and the MR data was analyzed by a MRI specialist. Subsequently, the intensity (hypo-, iso- and hyperintense) and T1* and T2* relaxation times were obtained.
Statistical analysis
The data was analyzed with SPSS 17.0 for Windows and with Excel (windows 10), and was expressed as mean ± standard deviation (M ± SD). Data was found significant when p was less than 0.05 (p <.05). All the data was first examined for normality. No statistical test was applied to check the set up (step one). The data of the healthy participants and deceased participant (step two) could not be analyzed. The data of the mice (step three) could also not be
analyzed. The intensities of the MR images were not statistically analyzed. However, the relaxation times of the MR images and the fractions of the Hb derivatives were tested with Spearman’s correlation coefficient.
Results
Optimization of reflectance spectroscopy
measurement protocol
The drying time of 30 minutes of bloodstain gave a R² of 0.99. The interval time was not fixed and varied between 3800-8100 msec. The drying time of 30 minutes, the free interval time and the distance gave the highest R² scores, therefore further bloodstains were measured with these settings . Figure 6 shows the measurement bloodstains over a period of time with the above mentioned settings. The interval time of the measured phantom SDH’s varied between 4260-5700 msec, while the distance was kept constant (0.9 mm) as well as the drying time (30 minutes). The condition of the clot was best when the clot forming time was 96 hours. These settings gave a R² of 0.99 and was therefore used in further measurements.
The type of conditions that gave the highest R² was smeared phantom SDH’s onto glass (R²=0.9985) and plastic (R²=0.9993) with a white paper background (see Appendix A, Table 3Error! Reference source not found. ). The samples were measured 30 minutes after deposition, the distance varied between 5-9 mm, and the interval time varied between 3800-28700 msec. However the FTS Figure 4. Reflectance spectroscopy set up used in this study. Set up and im age derived from Edelm an et al., [20].
Page | 6 program developed by the AMC photonics group
[20], that is used to determine the Hb derivatives, is developed for a single layer (white cotton). Thus, not for a given background with a given substrate on top of it. Therefore it was chosen to smear the phantom SDH’s onto a white cotton cloth.
Bloodstains that were made with blood of healthy participants were measured with a fixed distance (9 mm) and a free interval time (3800-4050 msec), after a deposition time of 30 minutes to
produce a R² of 0.999. Bloodstains made of blood of a deceased person produced a R² of 0.999 as well with a fixed distance (9 mm), an interval time of 3800 msec and a drying time of 30 minutes. Unfortunately, only one blood sample of a deceased participant was obtained, therefore statistical analysis could not be performed. Figure 5 shows the fractions of the Hb derivatives in blood of a deceased individual and healthy participants. The Hb derivatives were higher in the blood of the deceased individual than the Hb derivative in the blood of the healthy participants.
Temporal profile and correlations of phantom
SDH
Statistically, there was no significant differenc e between the phantom SDH in the tubes that were shaken, and the phantom SDH in the tubes that were not shaken (p >.05). Therefore the average between the two conditions are taken for further analysis. In Table 2 (Appendix A) an overview is given of the results dictated in literature and the results of this study of the corresponding variables. The T1* intensities obtained in this study were overall similar to the literature results values except for the
phantom SDH’s that were 72 hours (number 3) and 336 hours (number 5) old. For the T2* intensities the opposite is true, all the T2* intensities obtained in this study differed from the literature results except for the phantom SDH’s that were 24 hours (number 2), 72 hours (number 3), 504 hours (number 6) and 1512 hours (number 12) old. In Figure 7 the intensity maps of the phantom SDH are shown for the T1* and T2* MR images. It is noticeable that the intensity of the T1* MR image decreases with increasing age. This effect is not seen in the T2* MR images, the relaxation time seems to increase with increasing age. More importantly, the black spots visible on the intensity maps were not included in the analysis.
Figure 5. Bloodstains of healthy and a deceased person were measured with reflectance spectroscopy. The obtained HB derivatives of a deceased individual and five healthy participants are shown here.
Figure 6. Aging of HbO2, metHb and HC in blood of a healthy participant. The bloodstains were measured with reflectance spectroscopy.
Page | 7 Figure 8. The right image portrays the temporal
profile of the T1* relaxation time and T2* relaxation time.
The Hb derivative HbO2 is significantly correlated with age (rs= -0.93, p <.001), as are the derivatives
MetHb (rs=0.94, p <.001) and HC (rs=0.86, p <.001).
Thus, a steep decrease of HbO2 and increase of metHb and HC is noticeable as age increases. Furthermore, there is a significant relationship
between T1* relaxation time (rs= -0.99, p < .001,)
and age, and T2* relaxation time (rs=0.68, p <.001)
and age (see Figure 9), which is seen as a steep decrease of T1* relaxation time and an increase of T2* relaxation time as time progresses.
2000 0 1000 200 0 100
Figure 7. On the left (T1*) image, the intensities of the phantom SDH’s are visible at time 0. The middle image depicts the mean T1* relaxation time (0-2000) of the phantom SDH’s. The right image depicts the mean T2* relaxation time (0-200) of the phantom SDH’s. All the images were obtained through MR imaging.
Figure 9The image depicts the temporal profile of the Hb derivatives.
Page | 8 There are highly significant correlations between the
Hb derivatives and T1* and T2* relaxation times (see Figure 12 A-F). Specifically, there is a significant relationship between HbO2 and T1* relaxation time, rs=0.91, p <.001, and with T2* relaxation time, rs=
-0.71, p <.05. MetHb was significantly related to the T1* relaxation time, rs=-0.92, p <.001, and T2*
relaxation time, rs=0.71, p <.05. Furthermore, HC is
significantly correlated with T1* relaxation time, rs
=-0.86, p <.001, and T2* relaxation time , rs=0.60, p
<.05.
Small study: Distinguish TBI from non-TBI
Figure 10 illustrates the results of the fractions of Hb derivatives in the plasma of the brain injured and healthy mice. The fraction of HbO2 in the plasma of the brain injured mice seem to be higher than the fraction of HbO2 in the plasma of the healthy mice. Furthermore, The fraction of HC in the plasma of the brain injured mice seem to be lower than the fraction of HbO2 in the plasma of the healthy mice. Due to the large standard deviation it is visually not possible to see any further differences between the Hb derivatives of the TBI and non-TBI model.
Figure 10. The plasma stains were measured with reflectance spectroscopy. The HbO2 and HC differences are visible here.
Page | 9 Figure 11. The following significant correlations are shown: [A]The correlation between HbO2 and T1* relaxation time, [B]The correlation between HbO2 and T2* relaxation time, [C]The correlation between metHb and T1* relaxation time, [D]The correlation between metHb and T2* relaxation time, [E]The correlation between HC and T1* relaxation time, [F]The correlation between HbO2 and T1* relaxation time.
A B
C D
E
Page | 10
Discussion
This pilot study started with the findings of Sieswerda-Hoogendoorn [18] and Postema [19], namely that the age of SDH and therefore dating of TBI, cannot be objectively nor reliably determined with MR images. The relation between Hb derivatives and intensity shown on the MR images is never objectively tested. The purpose of this study was to theoretically underpin the aging of phantom SDH and to provide a basis for the future development of a revolutionary method that enables dating of TBI by studying the aging of phantom SDH. This was done by measuring the Hb derivatives of the phantom SDH with reflectanc e spectroscopy and the T1* and T2* relaxation times of the phantom SDH were determined with MR images. This was done in order to answer the following research questions: What is the quantitative temporal profile of the Hb derivatives in phantom SDH and is there a relation between the Hb derivatives and the T1* and T2* relaxation times?
In summary, there was no statistical difference between the phantom SDH’s in the tubes that were shaken and the ones that were not. It appears as though reflectance spectroscopy is not sensitive to the state of aliveness and could potentially be used on every person. It also appeared that overall, the T1* relaxation time of the phantom SDH measured with MR images, resembled that of the results mentioned in literature obtained from SDH in patients [15]. However, the T2* relaxation times were different than expected, this could be due to the fact that phantom SDH’s do not resemble actual SDH’s of TBI patient’s or that our results are more reliable than the results mentioned in literature. Furthermore, the results from this study indicate that different Hb derivatives are formed during the hemorrhage stages and that they age differently than previously assumed in literature [15], [23]. Most importantly, the Hb derivatives that are measured, are highly correlated to the T1* and T2* relaxation times. The T1* relaxation time seems to be positively correlated with HbO2, and negatively with metHb and HC. With increasing age, T1* relaxation time decreases. This indicates that energy transfer from the hydrogen protons to the surrounding tissue (Hb in our case) is not quick and therefore a long relaxation time is observed. This can
be due to several reasons. First, the movement frequency of the molecules in the surrounding tissue (Hb) changes over time. This would mean that the molecules that are formed over time will have a movement frequency closer to the lamorfrequenc y and this would facilitate energy transfer and is seen as shorter relaxation time. The kind of tissue that has this property are fatty chains and proteins. The most likely explanation is that proteins are formed or are released, and the latter is mostly due to cell lysing. Second, a conformation change of the surrounding molecules (Hb) has occurred and therefore the PEDD interaction can occur. This would facilitate energy transfer, which would result in a short relaxation time. The most probable explanation would be that metHb is formed over time, which is indeed formed in older phantom SDH’s. Third, the decrease of T1* relaxation time could also due to the phase change, i.e. the blood will clot after a certain period. A liquid will not facilitate energy transfer due to the rapid movement of the molecules. However, energy transfer is facilitated when the clot is formed which is visible as a short relaxation time. However, controlled concentrations of these Hb derivatives need to be measured in the future to study their influence on the T1* relaxation times.
On the other hand, T2* relaxation time seems to be negatively correlated with HbO2, and positively with metHb and HC. Unfortunately, to explain the relaxation time of the T2 MR images, no coherence answer can be provided. The curve of T2* relaxation time can be interpreted in two different ways. First, there are two data points (age 7 and 14 days) that seem to be outliers (although statistically not proven). If this is the case, we can interpret the relaxation time curve as increasing with increasing age. If this is the case, then two explanations can be given. First, the relaxation time is short for young phantom SDH’s, and becomes longer as the age increases. This would suggest that first paramagnetic molecules are present and afterwards diamagnetic molecules are formed. This would imply that in the beginning metHb is the dominant Hb derivative, and afterwards HbO2 or HC is formed which is not consistent with our Hb derivative temporal profile data. Therefore this explanation does not satisfies. Second, the relaxation time curve can also indicate
Page | 11 that the youngest phantom SDH was a clot (a clot has
large differences in field strengths which causes the differences in precession frequency and therefore the protonspins will defase quickly), and the oldest phantom SDH is a liquid (the independent magnet fields will cancel each other out and less quickly defasing will occur). However, this is also not consistent with our findings. Second, if the two data points (7 and 14 days) are not considered being outliers (which is statistically more justified), then the relaxation time shows a short, shorter, long, short pattern. There are again two possibilities to explain this phenomenon. First, to account for this observation, first we would see a clot, then this clot would become even more solid, then it would become liquid and then solid again to account for the relaxation times that are observed. However, this phase pattern of the phantom SDH’s is not observed. Second, to account for the relaxation times, the Hb derivatives would have the following pattern of magnetic properties: paramagnetic , paramagnetic, diamagnetic and paramagnetic . Which would translate in finding metHb first as the dominant Hb derivative, then HbO2 or HC had to be formed, and then metHb would have to return again. This theory is also not consistent with our Hb derivative temporal profile finding. Thus, no clear explanation can be given for the shape of the T2* relaxation time curve. Therefore, it is mandatory that this study is performed again.
With our results, a start has been made to better understand the aging of phantom SDH’s due to the high and significant correlations obtained between the Hb derivatives and T1* relaxation time and T2* relaxation time. If this is also seen in TBI patients, the kinetics of the Hb derivatives should be further studied. Ultimately, a situation is expected in which a SDH with a given relaxation time is correlated to a Hb derivative which tells something about the age of TBI due to the known kinetics. The results from this study look promising, and it is expected that a distinction between TBI and non TBI can be made. The results of this study and the prospects it provides are very valuable, because it can ultimately be used in court cases. Dating of TBI can lead to a suspect and/or exonerate an innocent and it can aid the reconstruction of activities.
To our knowledge, this is the first study worldwide in which Hb derivatives are quantitatively measured in phantom SDH and are correlated to the
T1* and T2* relaxation times and therefore contain a lot of strengths. Such as the fact that the HB derivatives are objectively measured with a validated set up which is developed by the highly experienced AMC forensic photonics group. Another strong point of this study is that the T1* and T2* relaxation times of the phantom SDH resembled that (for the most part) of the SDH measured in living patients, therefore the phantoms are probably of good quality.
With that said, to every study there are limitations to be found as is valid in this case. Although the spread of the age of the phantom SDH was large, there was only one data point for every time available. This prevented the possibility to use more sophisticated statistical analysis. The fact that the data was not normal stresses the fact that the correlations have to be interpreted with caution. Instances in which the T1* relaxation time and T2* relaxation time were different than expected, it is most likely caused by the absence of the appropriate volume of air, absence of CSF and proteins, and/or the presence of other Hb derivatives, such as hemosiderin and deoxyhemoglobin, that were not measured with reflectance spectroscopy. Even though, it seems that TBI can be distinguished from non TBI, this effect was observed in a small sample size consisting of mice which does not automatically mean that this situation can be extrapolated to humans. Furthermore, our study did not account for re-bleedings and the measurements were performed at room temperature. Finally, for each reflectance spectroscopy measurement, a correction for the wavelength-dependent reflectance values has to be applied in order to correct for intensity fluctuations of the light source [20]. Unfortunately, this was renounced, but after examination it was concluded that this had a minimal impact on the results.
As previously mentioned, physical abuse is a major social problem and (severe) TBI as a consequence is also the leading cause of death worldwide. These facts stress the urgency for a solution to this problem. Since there are no reliable methods to date TBI, this study is the first worldwide who attempts to theoretically underpin the aging of SDH’s. The results of the current study raises several possibilities, such as further exploring the age estimation of SDH with MR images. Indeed, the high and significant correlations found in this study
Page | 12 stresses a compelling need to test these methods
with a large sample size containing SDH of TBI patients, and blood of healthy participants. Further studies should also measure phantom SDH’s with mixed ages to mimic re-bleedings and perform the measurements at 37°C. Even though the renounced wavelength-dependent reflectance values had a minimal impact on the results, this should be taken into account in future measurements. Furthermore, controlled concentrations of the Hb derivatives have to measured again with the protocols suggested in this study. Currently, there are several studies [24]– [27] that are attempting to determine TBI with a hand held device which is held against the patients’ skull and measures the absorption of the hematomas. Coupling this technique with reflectance spectroscopy, establishing the rate and manner of aging of the SDH, gives possibilities to use this hand held device to date TBI by estimating the age of SDH.
Conclusion
This phantom study has major implications for crime related and non-crime related TBI. This study is the first in the world to quantitatively determine the Hb derivatives during different stages of SDH. Furthermore, this method is based on reliable setups and shows also robustness regarding the aliveness of
the individuals. There are high correlations between the Hb derivatives and T1* and T2* relaxation times, as well as indications regarding Hb derivatives differences between TBI and non-TBI. The results provide a better understanding of the aging of phantom SDH. Further investigation should indicate whether these results can be extrapolated to SDH’s of TBI patients. This matter is urgent and should be studied sooner rather than later in order to aid the court and medical professionals.
Ac knowledgements
I would like to thank the whole AMC biomedical engineering & physics department and in particular Prof. A a lders. Thank you for the wise lessons, your guidance, jokes, positive encouragement, supporting me at 8:00 AM, and many more. Ri chelle, Leah, Angela and Annemieke, I hope you know I am very grateful! Also a big thank you to my fellow interns who kept listening to my stories. Furthermore, I want to thank Prof. van Rijn and Bram. Bram thank you for the MR images and the analysis. Rick thank you for arranging the appointment to make the MR images, your humor, encouraging words, explanations and for always answering my emails. Probablemente esto es muy raro, pero quiero las gracias a mí mismo. Tal vez más importante, mi amor F i del. Muchas gracias, sin ti nada de esto es posible. Gracias por su bromas, apoyo, amor, paciencia eterna y por escuchar mis historias aburridas. Te amo!
Page | 13
References
[1] H. Daoud, I. Alharfi, I. Alhelali, T. Charyk Stewart, H. Qasem, and D. D. Fraser, “Brain injury biomarkers as outcome predictors in pediatric severe traumatic brain injury,” Neurocrit. Care, vol. 20, pp. 427–435, 2014. [2] Z. N. Burkhart, C. J. Thurber, A. Z. Chuang,
K. S. Kumar, G. H. D. Mph, and J. Kellaway, “Risk factors associated with retinal hemorrhage in suspected abusive head trauma,” J. AAPOS, vol. 19, no. 2, pp. 119– 123, 2011.
[3] Movisie, “Feiten en cijfers Huiselijk geweld,” 2014.
[4] Movisie, “Registraties
ouderenmishandeling,” 2012. [5] Movisie, “Huiselijk geweld: Aard en
omvang, gevolgen, hulpverlening en aanpak,” no. november, 2013.
[6] K. TALVIK, I., Metsvaht, T., Leito, “Inflicted traumatic brain injury ( ITBI ) or shaken baby syndrome ( SBS ) in Estonia,” no. July 2005, 2006.
[7] M. Fanconi and U. Lips, “Shaken baby syndrome in Switzerland : results of a prospective follow-up study , 2002 – 2007,” pp. 2002–2007, 2010.
[8] C. Brennan, L.K., Rubin, D., Christian, “Neck injuries in young pediatric homicide victims,” vol. 3, no. March, pp. 232–239, 2009.
[9] L. E. Thibault, D. Sc, and D. A. Bruce, “The shaken baby syndrome,” vol. 66, pp. 409– 415, 1987.
[10] V. Montfoort, “De inzet van het strafrecht bij kindermishandeling,” 2010.
[11] C. Scheau, A. E. Ghergus, G. Popa, E. M. Preda, and I. G. Lupescu, “Intracranial hemorrhage made easy - a semiological approach on CT and MRI,” 2014. [12] T. A. G. M. Huisman, “Intracranial
hemorrhage : ultrasound , CT and MRI findings,” pp. 434–440, 2005.
[13] G. Bradley, “MR Appearance,” pp. 15–26. [14] G. Vezina, “Assessment of the nature and
age of subdural collections in nonaccidental head injury with CT and MRI,” Pediatr. Radiol., vol. 39, pp. 586–590, 2009. [15] E. Parizel, P.M., Makkat, S., van Miert,
“Intracranial hemorrhage : principles of CT and MRI interpretation,” pp. 1770–1783, 2001.
[16] P. J. Vinchon, M., Noule, N., Tchofo, “Imaging of head injuries in infants: temporal correlates and forensic implications for the diagnosis of child
abuse,” vol. 101, no. Pediatrics 2, pp. 44– 52, 2004.
[17] E. Petrovic, A., Ogris, K., Hassler,
“Characterization of the time course of MR relaxation parameters for ageing blood,” p. 5780, 2007.
[18] T. Sieswerda-Hoogendoorn, “Chapter 6: Age determination of subdural hematomas with CT and MRI: a systematic review,” 2014.
[19] F. a M. Postema, T.
Sieswerda-Hoogendoorn, C. B. L. M. Majoie, and R. R. Van Rijn, “Age determination of subdural hematomas: Survey among radiologists,” Emerg. Radiol., vol. 21, pp. 349–358, 2014. [20] G. Edelman, T. G. van Leeuwen, and M. C.
G. Aalders, “Hyperspectral imaging for the age estimation of blood stains at the crime scene,” Forensic Sci. Int., vol. 223, pp. 72– 77, 2014.
[21] R. H. Bremmer, A. Nadort, T. G. van Leeuwen, M. J. C. van Gemert, and M. C. G. Aalders, “Age estimation of blood stains by hemoglobin derivative determination using reflectance spectroscopy,” Forensic Sci. Int., vol. 206, pp. 166–171, 2011.
[22] N. Achetib, “Detecting of ICAM-5 using immunoassays : ‘ The search for potential biomarkers to diagnose subdural
hematomas .,’” 2014.
[23] G. M. Lee, G.R., Foerster, J., Lukens, J., Paraskevas, F., Greer, J.P., Rodgers, Wintrobe’s Clinical Hematology Volume 1, 10th ed. Williams & Wilkins, 2009. [24] H. Ghalenoui, H. Saidi, M. Azar, S. T.
Yahyavi, H. B. Razavi, and M. Khalatbari, “Near-Infrared Laser Spectroscopy as a Screening Tool for Detecting Hematoma in Patients with Head Trauma,” vol. 23, no. 6, pp. 4–7.
[25] B. Zhang, Q., Ma, H., Nioka, S., Chance, “Study of near infrared technology for intracranial hematoma detection,” vol. 5, no. 2, pp. 206–213, 2000.
[26] C. Kahraman, S., Kayali, H., Atabey, “The Accuracy of Near-Infrared Spectroscopy in Detection of,” no. December, 2006. [27] C. S. Robertson, E. L. Zager, R. K. Narayan,
N. Handly, A. Sharma, D. F. Hanley, H. Garza, E. Maloney-wilensky, J. M. Plaum, C. H. Koenig, and A. Johnson, “Clinical Evaluation of a Portable Near-Infrared Device for Detection of Traumatic Intracranial Hematomas,” vol. 1604, no. September, pp. 1597–1604, 2010.
Page | 14
A p p e n d i x A
Figure 12. The following significant correlations are shown: [A]The correlation between HbO2, age and T1* relaxation time, [B]The correlation between HbO2, age and T2* relaxation time, [C]The correlation between metHb, age and T1* relaxation time, [D]The correlation between metHb, age and T2* relaxation time, [E]The correlation between HC, age and T1* relaxation time, [F]The cor relation between HbO2, age and T1* relaxation time.
B A
C D
Page | 15 Table 2. Overview of the study results, and the expected results obtained from literature
St age h em o r r h age H y per ac u t e Ac u t e Ear ly subac u t e L at e su b ac u t e C h r o n ic
Tu b e n u m b er 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Age (D ay s) 0,1 1 3 7 14 21 25 35 42 49 56 63 67 91 Age (H o u r s) 2,5 24 72 168 336 504 600 840 1008 1176 1344 1512 1608 2184 E x p e c te d [1 5 ], [ 1 7 ]
HB derivative HbO2 Hb Met-Hb Met-Hb Hemichrome°°°
T1* intensity Iso/hypo Iso/hypo hyper hyper hyper
T1* relaxation time (ms)°
1001-2000 1001-2000 0-1000 0-1000 0-1000
T2* intensity hyper Hypo hypo hyper hyper
T2* relaxation time (ms)°° 101-200 0-100 0-100 101-200 101-200 S tu d y r e su lt HB derivative 0.99 (HbO2); 0.01 (MetHb) 0.95 (HbO2) ; 0.01 (MetHb); 0.03 (HC) 0.99 (HbO2); 0.01 (MetHb) 0.99 (HbO2); 0.01 (MetHb) 0.99 (HbO2); 0.01 (MetHb) 0.95 (HbO2); 0.05 (MetHb) 0.80 (HbO2); 0.13 (MetHb); 0.08 (HC) 0.63 (HbO2); 0.21 (MetHb); 0.16 (HC) 0.54 (HbO2); 0.33 (MetHb); 0.13 (HC) 0.38 (HbO2); 0.35 (MetHb); 0.27 (HC) 0.33 (HbO2); 0.46 (MetHb); 0.21 (HC) 0.21 (HbO2); 0.47 (MetHb); 0.31 (HC) 0.40 (HbO2); 0.24 (MetHb); 0.36 (HC) 0.13 (HbO2); 0.56 (MetHb); 0.31 (HC) T1* intensity [15], [17]
Iso/hypo Iso/hypo Iso/hypo X hyper Iso/hypo X hyper hyper hyper hyper hyper hyper hyper hyper hyper
T1* relaxation time (ms)
1246,30 1149,01 1052,33 918,01 1004,89 923,28 751,80 552,86 449,20 430,39 418,14 405,69 356,39 326,42
T2* intensity [15], [17]
hypo X Hypo hypo hypo X hypo X hyper hypo X hypo X hypo X hypo X hypo X hyper hypo X hypo X
T2* relaxation time (ms)
65,65 63,86 76,03 34,74 37,75 102,98 91,10 85,27 86,32 96,77 97,29 106,93 94,39 90,55
°
hyperintense = 0-1000; iso/hypointense = 1001-2000 °° hypointense = 1-100 ; hyperintense = 101-200°°° in case of rebleeding, superparagmatic hemosiderin is expected at the rim
Page | 16
Figure 14. The image shows the best fit (M0-e-t/T2) of the 30 images obtained of 1 selected pixel in
each phantom SDH.
Figure 13. The left image depicts the T1* intensity of the phantom SDH’s at timepoint 0 ms, while the right image shows the intensity of the
phantom SDH’s on T1* at timepoint 850 ms. The middle image shows the best fit (A(1-Be-t/T*
)) of the 80 images obtained of 1 selected pixel in each phantom SDH.
Page | 17 Table 3. The table depicts the results of the optimization step of the measurement protocol in which different conditions were measured to obtain the highest R²
Sam p le So u r c e Bac k gr o u n d
R ²
Whole A third Thick
slices
Thin slices
smeare d
Glass Plastic Petri
dish White cotton White Cotton White paper Black paper X X X 0.285 X X 0.978 X X X 1 X X X 0.992 X X X 0.998 X X X 0.999 X X X 0.998 X X X 0.994 X X X 0.997 X X X 0.998 X X X 0.360 X X X X 0.965 X X X 0.986 X X X 0.262 X X X 0.872 X X X 0.930 X X X 0.786 X X X 0.913 X X X 0.733 X X X 0.247 X X X 0.977 X X X 0.518 X X X 0.967 X X X 0.869 X X X 0.154 X X X 0.79 X X X 0.781 X X X 0.959 X X X 0.976 X X 0.674 X X X 0.987 X X X 0.782 X X X 0.168 X X X 0.736 X X X 0.615 X X X 1 X X X 0 .999 X X X 0 .999 X X X 0 .999 X X X 1 X X X 0 .998 X X X 0 .998 X X X 0 .998
