Diffusion tensor brain imaging in forensic radiology

Diffusion tensor brain imaging in forensic radiology

Journal of Forensic Radiology and Imaging 3 (2015) 193–199 Contents lists available at ScienceDirect Journal of Forensic Radiology and Imaging journ...

1MB Sizes 2 Downloads 193 Views

Journal of Forensic Radiology and Imaging 3 (2015) 193–199

Contents lists available at ScienceDirect

Journal of Forensic Radiology and Imaging journal homepage: www.elsevier.com/locate/jofri

Diffusion tensor brain imaging in forensic radiology Nadav Berkovitz, Michael Abrahamy n, Paul Gottlieb, Margarita Vasserman, Sigal Tal Radiology Department, Assaf Harofeh Medical Center Affiliated with Sackler School of Medicine Tel-Aviv University, Zerifin 70300, Israel

art ic l e i nf o

a b s t r a c t

Article history: Received 5 July 2015 Received in revised form 13 September 2015 Accepted 17 September 2015 Available online 28 September 2015

Objectives: Evaluation of changes in DTI parameters in post-mortem (PM) cadaveric vs. antemortem (AM) healthy control brains based on eigenvalues at o24 h PMI. Methods and materials: The in-situ DTI brain scans of 10 PM subjects were compared to 10 AM controls. DTI metrics were measured from 25 ROIs for each brain in the gray and white matter. Results: PM eigenvalues (λ1, λ2, λ3) were significantly lower than in the AM controls, independent of tissue type. Longitudinal diffusivity was more affected than transverse (control vs. PM, for λ1, λ2 and λ3 in mm2/s  10  3: 0.96, 0.74, 0.58 vs. 0.27, 0.20, 0.15 for gray matter and 1.52, 0.44, 0.25 vs. 0.33, 0.10, 0.05 for white matter, po 0.0001). FA was significantly higher PM for gray matter (0.28 PM vs. 0.22 control, p¼ 0.003) but similar for white matter (0.73 PM vs. 0.75 control, p¼ 0.4). ADC values were lower for PM (0.73 control vs. 0.16 PM and 0.76 control vs. 0.21 PM, in white/gray matter respectively, p o0.0001; all units mm2/s  10  3). Conclusion: Both longitudinal diffusivity, transverse diffusivity and ADC are reduced PM. The lack of FA changes in white matter PM implies that FA changes in stroke are due to the ischemic cascade rather than direct cell death. Gray matter of the caudate showed an increase in FA similar to what is seen in a number of both degenerative and inflammatory pathologies. & 2015 Elsevier Ltd. All rights reserved.

Keywords: Diffusion tensor imaging Post-mortem Forensic radiology Brain

1. Introduction Diffusion-tensor imaging (DTI) is an MR sequence that provides information on biological tissue activity at the microstructural level by quantifying water diffusion in various tissues. DTI assesses diffusion magnitude and diffusion directionality, and can elucidate fine fundamental details on both healthy and damaged neurological tissue. DTI enables the determination of three perpendicular eigenvectors, whose magnitudes are given by their three corresponding eigenvalues, λ1, λ2, and λ3. The indices derived from DTI measurements such as fractional anisotropy (FA) and the apparent diffusion coefficient (ADC) are quantitative and objective measures of diffusion properties that can be measured in the human brain [1]. Decreased diffusivity has been described for cell death, particularly in cases of stroke. Post-mortem (PM) DTI analysis is designed to compare and contrast MRI-derived data with histological and forensic findings. Abbreviations: ADC, apparent diffusion coefficient; DTI, diffusion-tensor imaging; DWI, FA, fractional anisotropy; MR, magnetic resonance imaging; ROI, region of interest; PMI, post-mortem interval; ES, effect size n Corresponding author. Fax: þ 972 8 954 2113. E-mail addresses: [email protected] (N. Berkovitz), [email protected] (M. Abrahamy), [email protected] (P. Gottlieb), [email protected] (M. Vasserman), [email protected] (S. Tal). http://dx.doi.org/10.1016/j.jofri.2015.09.001 2212-4780/& 2015 Elsevier Ltd. All rights reserved.

In the only such study to date, Scheurer et al. [2] analyzed MRI data from cooled, non-fixated brains using a clinical-standard MRI system (1.5 T). They found that the ADC values of the brain were significantly decreased compared to normal controls. Increasing the post-mortem interval (PMI) decreased the ADC values of gray matter, whereas the white matter showed no significant changes. ADC values were significantly lower in cases of mechanical and hypoxic brain injuries (caused by heart failure), compared to traumatic brain injury. Post-mortem FA was not significantly different from the FA of live brains, and showed little influence of PMI. The authors emphasized the importance of MRI in general, and ADC in particular, as a novel tool in forensic medicine that can help determine cause of death. The present study expands upon by Scheurer et al. [2] by comprehensively evaluating changes in DTI parameters, and by including the eigenvalues, (λ1, λ2, λ3), FA and ADC in cadaveric full brains at a relatively short PMI compared to antemortem (AM) controls. Assessment of the eigenvalue data determines the main directional changes in the PM brains. Unlike the longer and varied mean PMIs examined in Scheurer et al., our study examined the progression of early (o24 h) PM damage to brain tissue by comparing the extent of change in different areas and directions of diffusivity (longitudinal vs. transverse). The DTI parameters in both white and gray matter were measured and compared with the corresponding anatomical locations in healthy live controls and compared and contrasted the data with known DTI changes in

194

N. Berkovitz et al. / Journal of Forensic Radiology and Imaging 3 (2015) 193–199

stroke ( o24 h) and neurological disease.

2.5. Data collection

2. Materials and methods 2.1. Subjects The in-situ DTI brain scans of 10 post-mortem human cadavers were compared to a antemortem control group consisting of the DTI scans of normal brains of living patients with no known pathologies. The 10 deceased cases, 3 females and 7 males, ranged in age from 30 to 62 years (mean 37.5 710(SD)). Circumstances of death included natural death, disease, suicide, and homicide (Table 1). The average PMI was 12 75.2 h. Prior to scanning, PM examination by a forensic pathologist was performed to ensure compliance with the inclusion criteria. Average core temperature before scanning was 14.8 76.4 °C, measured rectally with a digital thermometer. The 10 control subjects, 5 females and 5 males, ranged in age from 24 to 39 (mean age 31 75). The demographics are presented in Table 1. IRB approval was obtained. 2.2. Inclusion and exclusion criteria The study included subjects with an intact cranium and PMI of less than 24 h. It excluded subjects with brain trauma, history of neurological disorder or findings of visible pathology as cause of death on conventional brain MRI. Antemortem controls were free of pathological findings. 2.3. MR imaging Imaging of the in-situ brain was performed on a 1.5 T system (Siemens Megnaton Aera, Erlangen, Germany). Sagittal, axial and coronal T1, axial and coronal T2, axial FLAIR, GRE T2 and SWI were obtained. Axial DTI of the entire brain was done with pulsed gradient, spin-echo and echo-planar imaging: repetition time (TR) 6000, echo time (TE) 97, matrix 132  32, field of view 230 mm  230 mm, contiguous slice thickness 3 mm, 2b values 0 and 1000 s/mm2, acquisition time 3.26 min, pixel size 0.9  0.9 mm2. Diffusion weighting was applied along 30 noncollinear axes. 2.4. Image processing Quantitative analysis of DTI data and DTI maps were generated by the DTI task card using MRWP with a SyngoMR D11 imaging software platform (Siemens Medical Solutions, Erlangen, Germany). Table 1 Subject demographics and clinical characteristics. AM controls (n¼ 10)

PM subjects (n¼ 10)

Age (years) Gender

31(24–39) F¼ 5 M ¼5 Circumstances of death N/A

Clinical cause of death

N/A

37.5(30–62) F¼ 3 M ¼7 Unexplained sudden death Disease Suicide Murder Hypovolemic shock Suffocation Drug/alcohol abuse Pulmonary embolism MI or cardiac arrhythmia Acute pneumonia

3 2 4 1 2 3 1 1 2 1

DTI metrics, including λ1, λ2, λ3, ADC and FA were measured from 25 ROIs for each brain as presented in Fig. 1. From each hemisphere, 12 measuring ROIs were placed in consistent locations. Another single ROI was placed in the vermis for each brain. All measurement ROIs were of uniform size of 3 mm. Gray matter regions included the caudate, globus pallidum, motor cortex, and vermis (1 measurement per brain); the white matter included the midbrain corticospinal tract (CST), pons CST, corpus callosumgenu, body, and splenium, centrum semiovale, frontal white matter and internal capsule (anterior and posterior). 2.6. Data analysis The data were categorized by anatomical region and as either white matter (WM) or gray matter (GM). The DTI values of WM and GM were analyzed separately as is typical in ischemia studies [3]. Because DTI values are temperature-dependent, DTI metrics (λ1, λ2, λ3, and ADC) of the postmortem cases were also temperature-corrected to 38 °C by using a correction factor of 2% per °C according to the equation ADCTc ¼ ADC(100% þ2%)(38 °C-Tscan) used by Scheurer et al. The body core temperature measured (anal probe) at the start of the scan (Tscan) was used for the correction [2]. 2.7. Statistical analysis All variables and derived parameters were tabulated by descriptive statistics. For all statistical tests, a p-value of 5% or less was considered statistically significant. The data were analyzed s using the SAS version 9.1 (SAS Institute, Cary, North Carolina).

3. Results The DTI metrics measured in the different ROIs were categorized in terms of location, type of tissue (white/gray) and study group, and are presented in Table 2. Fig. 2 shows the DTI metrics for white/gray matter for the PM subjects vs. the AM controls, with and without temperature normalization. Most DTI metrics for PM DTI values were significantly different from the DTI metrics of the AM controls. All eigenvalues (λ1, λ2, λ3) decreased significantly in the PM group, independently of the type of tissue analyzed. λ1, λ2, λ3 were equal to 0.96 (Std, 7 0.17), 0.74 (Std, 70.15) and 0.6 (Std, 70.15) for gray matter controls vs. 0.27 (Std, 70.06), 0.20 (Std, 70.04), and 0.15 (Std, 70.03) for gray matter PM, respectively, and 1.52 (Std, 70.25), 0.44 (Std, 7 0.15), and 0.25 (Std, 70.14) for white matter controls vs. 0.33 (Std, 70.08), 0.10 (Std, 70.05), and 0.05 (Std, 70.03) for white matter PM, respectively. For all values (units measured in mm2/s  10  3), p o0.0001 (T-test). The largest relative decrease from control subjects to PM subjects was found for the λ1 of the white matter (1.52 vs. 0.33 mm2/s  10  3). In line with previous reports [9], in our study the ADC values were reduced post-mortem for both the white and gray matter. They dropped from 0.73 (Std, 70.09) mm2/s  10  3 to 0.16 (Std, 7 0.04) mm2/s  10  3 and 0.76 (Std, 70.15) mm2/s  10  3 to 0.21 (Std, 70.07) mm2/s  10  3, for the white and gray matter respectively (po 0.0001, T-test), as listed in Table 2. The FA for the combined gray and white matter in our study (data not shown) did not differ significantly between the PM subject and the AM controls. In both PM subjects and AM controls, the highest FA was measured in the splenium of the corpus callosum with 0.87 (Std, 70.05), and the lowest in frontal WM with

N. Berkovitz et al. / Journal of Forensic Radiology and Imaging 3 (2015) 193–199

195

Fig. 1. An illustration of ROI locations on PM brains using DTI for (1) cortico-spinal tracts in pons, (2) cortico-spinal tracts in midbrain, (3) vermis, (4) head of caudate, (5) anterior horn of internal capsule, (6) posterior horn of internal capsule, (7) pallidum, (8) genu of the corpus callosum, (9) splenium of the corpus callosum, (10) body of the corpus callosum, (11) motor cortex, (12) frontal lobe white matter and (13) centrum semiovale.

0.49 (Std, 70.07). These findings are consistent with those reported in previous studies [9]. However when comparing the data for gray and white matter separately, the FA showed a significant increase for gray matter from 0.22 (Std, 70.1) in AM controls to 0.28 (Std, 70.04) in postmortem individuals (p ¼0.003, two tailed T-test), whereas the FA of the white matter was practically the same between these areas measuring 0.72 (Std, 70.14) in controls vs. 0.73 (Std, 7 0.15) in PM (p ¼0.4, two-tailed T-test). The highest increase in FA from AM controls to PM subjects was observed in the caudate increasing from 0.17 (Std, 70.03) to 0.29 (Std, 70.03) (p o.0001, two-tailed T-test). When the MRIs were temperature normalized, there were no changes in trends (Fig. 2). Since the value changes in the normalized results were small in magnitude and did not affect the differences between AM controls and PM subjects, the remaining analyses used the original non-normalized values. The effect size (ES) between groups (PM subjects vs. AM controls) are presented in Table 3. For white matter eigenvalues, longitudinal diffusivity (λ1) was the most affected, with greater longitudinal than transverse ES values at every measured point. For the gray matter, the transverse diffusivity values (λ1, λ2) were the most affected for the caudate and globus pallidum. ES remained high for ADC and low for FA in all brain locations (both white and gray matter) with a high ADC ES for the caudate and vermis (4 and 1.78, respectively). Table 4 lists the correlations between DTI metrics. A high inverse correlation was observed between FA and λ2 and λ3 in white matter (ρ  0.8 to  0.9, po 0.0001), as well as a statistically significant inverse correlation between FA and ADC in the white matter of the PM subjects (ρ  0.2, p¼ 0.06, to  0.5, po 0.0001). In the AM controls’ white matter, the ADC correlated with longitudinal (λ1) diffusivity (ρ 0.43 p o0.0001) and with transverse (λ2 and λ3) diffusivity (ρ 0.19 p¼ 0.06 and 0.33, p ¼0.0007) respectively, p ¼0.06. Similarly, for the PM subjects, ADC showed high correlations with eigenvalues, but was similar in numerical values for all directions (0.71, 0.72, 0.63 for λ1, λ2, λ3, respectively, P o0.0001). For the gray matter, the correlation between ADC and λ1 λ2, λ3 was high for both PM subjects (0.9, 0.9, 0.7, p o0.0001) and the AM controls (0.8, 0.97, 0.9, p o0.0001).

The temperature-normalized DTI metrics of post-mortem cases and associated clinical causes of death (Table 5) did not reveal any significant pattern.

4. Discussion DTI parameters, eigenvalues (λ1, λ2 and λ3) as well as ADC and FA were measured in the PM brain in-situ to evaluate microstructural tissue deterioration after death. The PM group was compared to AM healthy volunteers. Comparison of PM subjects and AM controls’ brains showed a decrease in diffusivity of the PM brain tissue expressed by ADC, similar to findings reported in Scheurer et al. [2]. Here, however, we report eigenvalues (λ1, λ2 and λ3) of the in-situ PM brains and for a short PMI of less than 24 h. Our analysis shows a decrease in ADC and in all three eigenvalues.

4.1. Trends for the vermis, caudate Similar to the data presented in Scheurer et al. [2], the greatest reduction in ADC in terms of location was in the vermis (1.06 mm  3/s  10  3 to 0.2 mm  3/s  10  3; a 81% reduction). Note that the vermis had the highest initial ADC (1.06 mm  3/s) and expressed structural characteristics unique to the functional vermis, which may be irrelevant to the edematous PM vermis. The vermis also showed high variability; the standard deviation for vermis measurements in live scans (0.11) was only exceeded by the corpus callosum body (0.15). The effect size of the ADC change in the vermis (10.75) was, however, not the highest ES for brain locations. Thus, it cannot be concluded that the vermis is always the locus of the most reduction in ADC due to the high dispersion in initial values. Furthermore, the vermis may be a likely site for CSF measurement contamination. By contrast, the caudate, another gray matter location, exhibited the largest effect size (17.3) in terms of ADC reduction (0.71 mm  3/s  10  3 to 0.19 mm  3/s  10  3; a 73% reduction).

196

Table 2 DTI metrics (mean) per location. All measurements except for Vermis are bi-hemispheral. Control

Post-mortem

λ1 (STD) λ2 (STD) mm2/s  10  3

λ3 (STD)

ADC (STD)

FA (STD)

λ1 (STD) mm2/s  10  3

Gray matter Caudate Pallidum Motor cortex Vermis

0.96 (0.17) 0.84 (0.07) 0.94 (0.12) 0.93 (0.14) 1.21 (0.11)

0.74 (0.15) 0.70 (0.04) 0.71 (0.04) 0.63 (0.06) 1.05 (0.11)

0.58 0.59 0.52 0.47 0.88

(0.15) (0.04) (0.06) (0.08) (0.15)

0.76 (0.15) 0.71 (0.04) 0.72 (0.05) 0.68 (0.05) 1.06 (0.11)

0.22 (0.10) 0.17 (0.03) 0.27n (0.08) 0.32n (0.08) 0.13 (0.03)

0.27 0.25 0.33 0.26 0.23

(0.06) (0.04) (0.05) (0.04) (0.05)

White matter Midbrain corticospinal tract (CST) Pons CST

1.52 (0.25) 1.69 (0.10) 1.31 (0.09)

0.44 (0.15) 0.32 (0.07) 0.56 (0.06)

0.25 (0.14) 0.16 (0.05) 0.31 (0.08)

0.73 (0.09) 0.73 (0.04) 0.70 (0.10)

0.72n (0.14) 0.84 (0.04) 0.60n (0.09)

Corpus callosum Genu Body Splenium Centrum semiovale Frontal White Matter

1.79 (0.12) 1.83 (0.14) 1.69 (0.20) 1.34 (0.11) 1.17 (0.08)

0.35 0.40 0.26 0.47 0.68

(0.07) (0.12) (0.09) (0.09) (0.07)

0.18 (0.06) 0.35 (0.30) 0.15 (0.06) 0.27 (0.08) 0.39 (0.07)

0.77 0.87 0.70 0.68 0.75

0.84n 0.71n 0.87n 0.67n 0.49n

Internal capsule Anterior Posterior

1.38 (0.14) 1.46 (0.08)

0.53 (0.10) 0.41 (0.08)

0.28 (0.11) 0.20 (0.04)

0.72 (0.06) 0.70 (0.03)

n

Indicates p4 0.05 for comparison of PM values vs. controls.

(0.05) (0.04) (0.07) (0.09) (0.02)

(0.05) (0.14) (0.05) (0.07) (0.07)

0.67n (0.11) 0.76n (0.05)

λ2 (STD)

λ3 (STD)

ADC (STD)

FA (STD)

0.20 (0.04) 0.19 (0.02) 0.23 (0.04) 0.20 (0.03) 0.18 (0.05)

0.15 0.14 0.15 0.15 0.15

(0.03) (0.02) (0.04) (0.03) (0.04)

0.21 (0.07) 0.19 (0.02) 0.24 (0.04) 0.20 (0.04) 0.20 (0.05)

0.28 (0.04) 0.29 (0.03) 0.33n (0.07) 0.29n (0.05) 0.21 (0.06)

0.33 (0.08) 0.40 (0.06) 0.37 (0.05)

0.10 (0.05) 0.11 (0.04) 0.13 (0.04)

0.05 (0.03) 0.05 (0.02) 0.07 (0.04)

0.16 (0.04) 0.19 (0.04) 0.20 (0.03)

0.73n (0.15) 0.78 (0.06) 0.62n (0.11)

0.31 0.36 0.37 0.28 0.24

(0.08) (0.15) (0.07) (0.06) (0.05)

0.06 (0.04) 0.09 (0.04) 0.05 (0.03) 0.11 (0.03) 0.13 (0.04)

0.03 0.05 0.03 0.05 0.06

0.13 0.16 0.15 0.15 0.15

0.86n 0.79n 0.89n 0.67n 0.55n

0.30 (0.04) 0.36 (0.07)

0.09 (0.06) 0.13 (0.04)

0.05 (0.04) 0.06 (0.03)

(0.02) (0.03) (0.02) (0.02) (0.03)

(0.04) (0.03) (0.03) (0.03) (0.03)

0.15 (0.04) 0.18 (0.05)

(0.11) (0.11) (0.06) (0.08) (0.17)

0.73n (0.19) 0.72n (0.10)

N. Berkovitz et al. / Journal of Forensic Radiology and Imaging 3 (2015) 193–199

Anatomy

N. Berkovitz et al. / Journal of Forensic Radiology and Imaging 3 (2015) 193–199

Fig. 2. A comparison of DTI metrics between AM control and PM subjects scans, including measured values and normalized values for white matter (a) and gray matter (b).

Table 3 Effect Size for different DTI metrics in white and gray matter. nAll measurements except for the vermis are bi-hemispheral.

White matter effect size Midbrain CST Pons CST Corpus callosum Genu Body Splenium CSO FWM Internal capsule Anterior Posterior Gray matter effect size Caudate Pallidum Motor cortex Vermisn

FA

ADC

λ1

λ2

λ3

1.2 0.2

13.5 7.69

16.1 13.4

3.82 8.6

3.14 4

0.25 0.64 0.36 0 0.5

14.2 7.89 11 8.83 24

14.8 15.5 9.78 12.5 14.3

5.27 3.88 3.5 6 10

3.75 1.82 3 4.4 6.6

0.4 0.53

11.4 13

12 14.67

5.5 4.67

3.07 4

4 0.8 0.46 1.78

17.3 10.7 10.7 10.75

10.7 7.17 7.4 12.25

17 12 9.55 10.88

15 7.4 5.82 7.68

4.2. FA and ADC changes The mechanism of change in diffusion metrics may relate to PM cellular edema and include an increase in the slow diffusion intracellular volume fraction, changes in membrane permeability, shrinkage of the extracellular space, intracellular protein breakdown and reduced energy-dependent transport [4,5]. Here, FA showed an increasing inverse correlation in the PM subjects with transverse diffusivity. In areas of high FA, a basic anisotropic neural

197

structure was retained PM. There were more barriers to transverse diffusivity, creating a high correlation with the lowering transverse diffusivity values. Gray matter changes did not exhibit this correlation, as FA parameters in gray matter appeared to involve more complex elements of structural change expressed by the FA increase. Both increased and decreased FA have been described in different temporal periods of brain cell death, particularly in stroke [4,6]. FA is considered predominantly affected by axonal membranes and modulated by myelination [7]. Although DTI can be used to assess gray matter abnormalities, evaluation of the cortical gray matter FA is complex due to the low SNR. This originates from the partial volume effect of cerebrospinal fluid in the sulcus. At the brain surface, FA measurements can also be biased by Eddy currents [8]. Here, for the combined data (white and gray matter, data not shown), FA did not show significant changes, as was also the case in Scheurer et al. [2]. The suggested causes of FA change are based on the outcome of cell death. The lack of FA change in our study and in Scheurer et al. [2] is consistent. The difference between cell death in stroke vs. death is the lack of live tissue responding to the insult. The implications are that the FA changes in stroke, which are sensitive to cell insult, are not direct outcomes of cell death. Rather, these changes are caused by an active reaction by live tissue or retained function in dying cells to local insult. This potentially alters standard interpretations of FA in stroke. FA for gray matter showed a statistically significant increase in PM (p o0.001), particularly in the caudate. An increase in FA for the caudate implies an increase in directionality in gray matter PM. Our data are consistent with focused DTI measurements of the caudate gray matter [9] that found a significant increase in FA with age in both males and females, which was not markedly influenced by the level of noise in the sample. Another study showing an increase in the FA of the caudate and putamen along with a decrease in ADC examined the normal-appearing basal ganglia of multiple sclerosis patients [10]. The authors ruled out gliosis, which would have resulted in more disorganization and reduced anisotropy. A possible source of directionality and FA increase in gray matter might be dendrite deterioration. A regressive neuronal and dendrite elimination develops with age [7]. Dendritic deterioration in the striatum has been demonstrated histologically in dementia with Lewy-Bodies [11]. A more rapid dendrite reduction PM compared to axonal deterioration could explain the increase in directionality expressed in FA (an increase in longitudinal vs. transverse ratio) concurrent with a reduction in ADC. Both increases and decreases in ADC have been described for stroke and have been associated with different temporal periods post-stroke that are not directly related to FA values [4]. Our findings show ADC reduction. One possible interpretation is that the ADC increase seen in certain stroke measurements is also caused by active tissue reaction to the stroke rather than cell death itself. Another possibility is that the ADC increase occurs outside the time frame of 24 h. Neuro-inflammation is a key element in the ischemic cascade. This response appears to contribute to ischemic pathology, and accounts for the popularity of anti-inflammatory strategies. Reperfusion of the occluded vessel, either by compensation of the collateral circulation, or spontaneous or therapeutic recanalization leads to the generation of reactive oxygen species further promoting the inflammatory response [12,13]. Many of these changes do not occur PM as there is no supporting vascular system. Hence FA and ADC changes may be secondary to post-cell death events rather than caused directly by physiological intra and extracellular volume changes as previously described.

198

N. Berkovitz et al. / Journal of Forensic Radiology and Imaging 3 (2015) 193–199

Table 4 Pearson correlation for DTI parameters in PM and normal controls. PM white matter Pearson correlation coefficients FA FA ADC λ1 λ2 λ3

1.000  0.487 0.183n  0.823  0.798

ADC  0.487 1.0000 0.713 0.728 0.6338

AM control white matter Pearson correlation coefficients λ1

λ2 n

λ3

 0.823 0.728 0.196n 1.000 0.845

0.183 0.713 1.000 0.196n 0.111n

 0.798 0.633 0.111n 0.845 1.000

PM gray matter Pearson correlation coefficients FA ADC λ1 λ2 λ3 n

1.000  0.04n 0.213n 0.031n  0.387

 0.044n 1.000 0.852 0.934 0.726

FA

ADC  0.181 1.0000 0.433 0.186n 0.325

1.000  0.181n 0.732  0.894  0.751

n

λ1

λ2

λ3

0.732 0.433 1.000  0.725  0.579

 0.894 0.186n  0.725 1.000 0.599

 0.751 0.325  0.579 0.599 1.000

0.001n 0.800 1.000 0.680 0.497

 0.581 0.972 0.680 1.000 0.918

 0.714 0.898 0.497 0.918 1.000

AM control gray matter Pearson correlation coefficients 0.213n 0.852 1.000 0.824 0.345

0.031n 0.934 0.824 1.000 0.645

 0.387 0.726 0.345nn 0.645 1.000

1.000  0.476 0.001n  0.582  0.715

 0.476 1.000 0.800 0.972 0.898

Indicates p4 0.05.

Table 5 A comparison of DTI metrics (mean of all brain ROI) between PM individuals by clinical cause of death. Clinical Cause of Death (n)

ADC (STD) 2

mm /s  10 Hypovolemic shock (2) Suffocation (3) Drug/alcohol abuse (1 ) Pulmonary embolism (1 ) MI or cardiac arrhythmia (2) Acute pneumonia (1)

0.24 0.28 0.28 0.27 0.23 0.24

λ1 (STD)

λ2 (STD)

λ3 (STD)

0.44 0.51 0.48 0.47 0.41 0.39

0.18 0.21 0.24 0.18 0.17 0.19

0.10 0.12 0.14 0.12 0.11 0.13

3

(0.05) (0.08) (0.07) (0.07) (0.07) (0.08)

(0.11) (0. 12) (0.12) (0.11) (0.11) (0.06)

(0.08) (0.10) (0.10) (0.10) (0.10) (0.12)

(0.06) (0.08) (0.09) (0.08) (0.08) (0.10)

4.3. Relationships between FA and ADC

4.5. Limitations and future work

The inverse correlation between FA and ADC increased in white matter but not in gray matter. A possible explanation is that PM (at low ADC), the basic retained neural structure underlying the FA becomes the dominant consistent factor, creating a correlation between the FA and the low ADC, not necessarily a causative relationship. The FA of gray matter expresses more complicated changes possibly due to dendritic death. FA and ADC in stroke do not correlate directly [12]. Again, the different findings for FA in PM may be due to a lack of post ischemic cascade.

Ideally the study groups would have been age and sex matched. Considering the significance of DTI changes published in the single comparable study [2] with a smaller control group, we considered the AM control group in our study statistically valid. All cases were scanned in the early (o24 h) PM period. Though this is a more focused time period than the previous study, we do not present a correlation of DTI values per hour after death. To directly plot the changes of DTI over the short period would require a differently structured study. Post-mortem DTI in situ poses different problems compared with imaging the living. Effective fiber tracking may be compromised by pneumencephalon, SAH, galeal hematoma and blood around the head in the body bag [14]. Temperature and decay (post-mortem interval and ambient circumstances) in addition to intracranial pathologies significantly degrade the DTI quality and have a major effect on fiber tracking results and diffusion-weighted imaging [14]. We tried to minimize the influence of these limitations by focusing the ROI in areas without gas and decomposition using T1, T2 and FLAIR images. In addition, the PM scans were performed early ( o24 h) to decrease the existence of these limitations. We were unable to duplicate a direct correlation between cause of death and DTI derived metrics. This may be because the circumstances of death were of limited variability and limited group size in our study. Had we been able to show such a correlation, a forensic determination of cause of death could have been advanced. Further research could examine the similarity of PM DTI to stroke DTI values. If the similarity is high, it may provide another

4.4. Significance The study adds to a growing body of basic science data regarding DTI values. There is a dearth of PM DTI neurological studies in the ‘in situ’ radiological setting. We aimed to duplicate and expand on this scant previous work. A better understanding of PM changes and a comparison with live known pathological studies could provide a greater understanding of cell death and cell injury DTI effects. The differences in DTI values between PM cell death and post ischemic event cell death imply that the ischemic cascade in stroke has a greater impact on DTI values than previously described. Rather than measuring the cellular damage, FA and ADC may thus be better measures of cellular reaction. The eigenvalue changes shed light on the PM changes occurring in the early (o24 h) PM period. We showed the differences in situ between the gray and white matter in the PM period. These findings highlight the different micro-structural bases of the DTI values between different brain structures and different brain tissues.

N. Berkovitz et al. / Journal of Forensic Radiology and Imaging 3 (2015) 193–199

source of time of death definition. Conversely, the differences between the conditions may provide information regarding changes involving the penumbra and reversibility of insult or cell death which does not exist in the PM setting.

5. Conclusion In the 24 hour PM interval, quantifiable diffusion changes occur. These differences between PM and AM pathologies need be better understood with regard to the underlying mechanisms to enable a better interpretation of DTI data in both AM and PM cases. Both longitudinal diffusivity and transverse diffusivity are reduced PM and ADC is reduced. ADC was not seen to increase, unlike in some published stroke measurements. The lack of FA changes in white matter PM implies that FA changes in stroke are due to the ischemic cascade rather than direct cell death. Gray matter shows the greatest ADC reduction while gray matter, particularly the caudate, showed an increase in FA, similar to results in a number of both degenerative and inflammatory pathologies. This implies a common cause for FA increase in gray matter. Future studies of DTI parameters could provide information on the microstructural changes in the early PM period as compared to AM pathologies.

Conflict of interest NB, MA, PG, MV, and ST have no conflicts of interest to disclose.

References [1] T. Sigal, M. Shmuel, D. Mark, H. Gil, A. Anat, Diffusion tensor imaging of corpus callosum integrity in multiple sclerosis: correlation with disease variables, J.

199

Neuroimaging 22 (1) (2012) 33–37. [2] E. Scheurer, K.O. Lovblad, R. Kreis, S.E. Maier, C. Boesch, R. Dirnhofer, K. Yen, Forensic application of postmortem diffusion-weighted and diffusion tensor MR imaging of the human brain in situ, Am. J. Neuroradiol. 32 (8) (2011) 1518–1524. [3] A.G. Sorensen, O. Wu, W.A. Copen, T.L. Davis, R.G. Gonzalez, W.J. Koroshetz, T. G. Reese, B.R. Rosen, V.J. Wedeen, R.M. Weisskoff, Human acute cerebral ischemia: detection of changes in water diffusion anisotropy by using MR imaging, Radiology 212 (3) (1999) 785–792. [4] Y. Ozsunar, P.E. Grant, T.A. Huisman, P.W. Schaefer, O. Wu, A.G. Sorensen, W. J. Koroshetz, R.G. Gonzalez, Evolution of water diffusion and anisotropy in hyperacute stroke: significant correlation between fractional anisotropy and T2, Am. J. Neuroradiol. 25 (5) (2004) 699–705. [5] D. Le Bihan, J.F. Mangin, C. Poupon, C.A. Clark, S. Pappata, N. Molko, H. Chabriat, Diffusion tensor imaging: concepts and applications, J. Magn. Reson. Imaging 13 (4) (2001) 534–546. [6] Q. Yang, B.M. Tress, P.A. Barber, P.M. Desmond, D.G. Darby, R.P. Gerraty, T. Li, S. M. Davis, Serial study of apparent diffusion coefficient and anisotropy in patients with acute stroke, Stroke 30 (11) (1999) 2382–2390. [7] C. Beaulieu, The basis of anisotropic water diffusion in the nervous system-a technical review, NMR Biomed. 15 (7–8) (2002) 435–455. [8] S.K. Lee, D.I. Kim, J. Kim, D.J. Kim, H.D. Kim, D.S. Kim, S. Mori, Diffusion-tensor MR imaging and fiber tractography: a new method of describing aberrant fiber connections in developmental CNS anomalies, Radiographics 25 (1) (2005) 53–65 (discussion pp. 66–58). [9] K.M. Hasan, C. Halphen, M.D. Boska, P.A. Narayana, Diffusion tensor metrics, T2 relaxation, and volumetry of the naturally aging human caudate nuclei in healthy young and middle-aged adults: possible implications for the neurobiology of human brain aging and disease, Magn. Reson. Med. 59 (1) (2008) 7–13. [10] O. Ciccarelli, D.J. Werring, C.A. Wheeler-Kingshott, G.J. Barker, G.J. Parker, A. J. Thompson, D.H. Miller, Investigation of MS normal-appearing brain using diffusion tensor MRI with clinical correlations, Neurology 56 (7) (2001) 926–933. [11] S. Zaja-Milatovic, C.D. Keene, K.S. Montine, J.B. Leverenz, D. Tsuang, T. J. Montine, Selective dendritic degeneration of medium spiny neurons in dementia with Lewy bodies, Neurology 66 (10) (2006) 1591–1593. [12] Q. Wang, X.N. Tang, M.A. Yenari, The inflammatory response in stroke, J. Neuroimmunol. 184 (1–2) (2007) 53–68. [13] A.G. Ceulemans, T. Zgavc, R. Kooijman, S. Hachimi-Idrissi, S. Sarre, Y. Michotte, The dual role of the neuroinflammatory response after ischemic stroke: modulatory effects of hypothermia, J. Neuroinflamm. 7 (2010) 74. [14] Flach, et al., Deep into the fibers! Postmortem diffusion tensor imaging in forensic radiology, Am. J. Forensic Med. Pathol. 36 (3) (2015) 153–161.