Optimization of the fluid-attenuated inversion recovery (FLAIR) imaging for use in autopsy imaging of the brain region using synthetic MRI

Abstract

BACKGROUND:

The failure of cerebrospinal fluid (CSF) signal suppression in postmortem fluid-attenuated inversion recovery (FLAIR) of the brain is a problem.

OBJECTIVE:

The present study was to clarify the relationship between the temperature of deceased persons and CSF T ${}_{1}$ , and to optimize the postmortem brain FLAIR imaging method using synthetic MRI.

METHODS:

Forehead temperature was measured in 15 deceased persons. Next, synthetic MRI of the brain was performed, the CSF T ${}_{1}$ was measured, and the optimal TI was calculated. Two types of FLAIR images were obtained with the clinical and optimal TI. The relationship between forehead temperature and the CSF T ${}_{1}$ and optimal TI was evaluated. The optimized FLAIR images were physically and visually evaluated.

RESULTS:

The CSF T ${}_{1}$ and optimal TI were strongly correlated with forehead temperature. Comparing the average SNR and CNR ratios and visual evaluation scores of the two FLAIR images, those captured with the optimal TI showed statistically lower SNR, higher CNR, and higher visual evaluation scores ( $p<$ 0.01).

CONCLUSIONS:

Synthetic MRI enables the quantification of the CSF T ${}_{1}$ resulting from postmortem temperature decreases and calculation of the optimal TI, which could aid in improving the failure of CSF signal suppression and in optimizing postmortem brain FLAIR imaging.

Keywords

Synthetic magnetic resonance imaging autopsy imaging inversion time cerebrospinal fluid fluid-attenuated inversion recovery

1. Introduction

“Autopsy imaging (Ai)” is one of the methods used to determine the cause of death; It involves the use of computed tomography (CT) and magnetic resonance imaging (MRI) to aid in uncovering information regarding the inside of the corpse, which cannot be determined from the outside [1]. In addition to its use in determining the cause of death, “Ai” is sometimes used as a part of histological and pathological studies of the brain [2, 3]. In recent years, several studies have been performed to compare postmortem MRI findings in the brain with histopathological findings, the results of which are being used to elucidate the pathogenesis of unknown neurological diseases and to develop corresponding treatments [4, 5, 6, 7, 8, 9, 10, 11, 12]. Fluid-attenuated inversion recovery (FLAIR) is one of the sequences used in postmortem MRI of the brain. In this imaging method, the signal from the cerebrospinal fluid (CSF) is suppressed by setting an inversion time (TI) that corresponds to the longitudinal relaxation time (T ${}_{1}$ ) of the CSF [13, 14]. This method is highly sensitive to signal changes associated with a wide range of lesions, such as those resulting from periventricular demyelination and inflammatory diseases [15, 16]. However, during postmortem brain FLAIR imaging, we sometimes encounter the failure of CSF signal suppression. Figure 1 presents an example of normal FLAIR images and postmortem brain FLAIR images.

Figure 1.

This figure depicts two FLAIR images. a. The image on the left is an example of a normal FLAIR image. The CSF signal is sufficiently suppressed. b. The image on the right is a postmortem FLAIR image. The CSF signal is not suppressed. Abbreviations: FLAIR, fluid-attenuated inversion recovery; CSF, cerebrospinal fluid.

The cause behind the failure of CSF signal suppression in postmortem brain FLAIR imaging has been clarified in previous studies. The tissue contrast on postmortem MRI shows changes due to the postmortem changes that occur in the corpse, which are different from the findings in a living body [1]. Corpses cease to produce heat because their metabolism is shut down. Therefore, heat dissipation occurs due to physical phenomena (convection, conduction, and dissipation of heat and evaporation of vaporization heat through water evaporation), and the temperature of the corpse decreases by approximately 1 ${}^{\circ}$ C per hour after death [17, 18]. Consequently, the T ${}_{1}$ and transverse relaxation time (T ${}_{2}$ ), which are tissue-specific factors, change along with the variation in correlation time resulting from temperature changes, thereby producing a significant effect on the tissue contrast on MRI [19]. In addition, another factor that changes T ${}_{1}$ and T ${}_{2}$ is strengths of the static magnetic field. However, the change in T ${}_{1}$ due to strengths of the static magnetic field is negligible compared to the effect of the temperature decreases [20]. The body temperature decrease in a deceased person is one of the most important factors that must be considered in postmortem MRI because refrigeration may be necessary to prevent decomposition in clinical practice [1]. The temperature of CSF decreases as the body temperature decreases, and the T ${}_{1}$ of the CSF shortens [21]. As a result, it is not possible to perform postmortem brain FLAIR imaging at the appropriate TI setting resulting in the failure of CSF signal suppression [1, 17, 18, 19, 20, 21, 22]. By measuring the shortened T ${}_{1}$ value of the CSF could allow for the setting of the optimal TI in postmortem FLAIR imaging.

A new relaxation time measurement method that has recently emerged is called synthetic MRI. Synthetic MRI uses a pulse sequence technology called the quantification of relaxation times and proton density by multi-echo acquisition of a saturation-recovery using turbo spin echo readout (QRAPMASTER), with the help of multi-delay and multi-echo acquisition, this function collects eight types of data, four types of saturation delays, and two types of echo times (TE) corresponding to each slice [23]. The T ${}_{1}$ relaxation curve is estimated from the four saturation delays, and the T ${}_{1}$ value and proton density (PD) are calculated for each pixel. Similarly, T ${}_{2}$ values are calculated pixel-by-pixel from two different TE images [23]. Subsequently, multiple contrast images are created from the obtained quantitative values under arbitrary conditions. Furthermore, because synthetic MRI measures the quantitative values of each tissue, the T ${}_{1}$ (R ${}_{1}$ ) values, T ${}_{2}$ (R ${}_{2}$ ) values, and PD can be displayed as quantitative maps [24]. By setting a region of interest (ROI) at an arbitrary point on images that are created using dedicated software (“SyMRI”, SyntheticMR, Linköping, Sweden), it is possible to measure the quantitative values of relaxation time and to display a scatter plot [24, 25]. In modern medicine, it is necessary to resolve this problem in order to achieve higher diagnosis and autopsy rates and for the further advancement of histological and pathological research. Therefore, the purpose of the present study was to clarify the relationship between the body temperature of deceased persons and CSF longitudinal relaxation time, and then to optimize the postmortem brain FLAIR imaging method using synthetic MRI.

2. Materials and methods

2.1 Subjects

The subjects were 15 deceased persons who underwent brain autopsies at the research center. The ethics committee of the Institute of Brain and Blood Vessels, Mihara Memorial Hospital, approved the clinical research. The details of the study were fully explained to the bereaved families, and informed consent was obtained. The deceased persons were stored at room temperature if MRI could be obtained immediately after death. If MRI could not be performed immediately after death; the deceased persons were kept refrigerated. The deceased persons were brought into the MRI room without any special pretreatment.

2.2 Body temperature measurement and evaluation

Forehead temperature of the deceased persons was measured. In this study, a simple method for measuring the body surface temperature (forehead) was adopted, using a non-contact thermometer (PT-5LD, OPTEX FA Corporation, Kyoto, Japan). Temperature measurements were recorded before and after MRI scanning, and it was confirmed that the effects of the time elapsed since death and temperature changes owing to the specific absorption rate (SAR) during MRI [26] was small. Subsequently, the relationship between the forehead temperature and T ${}_{1}$ value of the CSF was evaluated. In addition, the forehead temperature measured before MRI scanning was used for all evaluations.

2.3 MR image acquisition

Synthetic MRI of the brain of the deceased persons was performed, and T ${}_{1}$ values of the CSF were measured using a specialized software application, SyMRI. The slice section in which the ventricles could be observed most clearly was selected for the measuring of the T ${}_{1}$ value of the CSF (Fig. 2). Measurement points were set within the left and right ventricles using rectangular ROIs of the same size. The T ${}_{1}$ value of the CSF was defined as the average of the T ${}_{1}$ values measured in the left and right ventricles. The optimal TI was calculated from the measured T ${}_{1}$ value using Eq. (1). Subsequently, two types of FLAIR imaging scans were performed: one using clinical TI (2,700 ms) and the other using optimal TI. The MR system used was a Discovery MR750W 3.0 Tesla scanner (General Electric Healthcare Medical Systems Corporation, Milwaukee, WI, USA), the receiver coil was an 8-channel head coil, and the synthetic MRI application used was the Magnetic Resonance image Compilation (MAGiC) application available onboard the device. The imaging parameters for synthetic MRI and FLAIR imaging are presented in Table 1.

$\displaystyle TI=T_{1}\textit{ Value}\times 0.693$ (1)

Table 1

The imaging parameters for synthetic MRI and FLAIR

	Synthetic MRI [2D]	FLAIR [2D]
Pulse Sequence	MAGiC	Fast Spin Echo
TR [ms]	4020	12000
TE [ms]	1st 21.9, 2nd 98.5	140
Flip Angle [ ${}^{\circ}$ ]	90	160
Band Width [Hz]	22.73	35.71
Slice thickness [mm]	5	5
FOV [mm]	230	230
NEX	1	1
ETL	14	21
Matrix size	320 $\times$ 224	320 $\times$ 224
Acquisition	1	2, 3

Abbreviations: MRI, magnetic resonance imaging; FLAIR, fluid-attenuated inversion recovery; TR, repetition time; TE, echo time; FOV, field of view; NEX, number of excitations; ETL, echo train length.

Figure 2.

Slice section used for measurement and placement of the rectangular ROI. Abbreviations: ROI, region of interest.

2.4 Analysis and evaluation of the FLAIR images

2.4.1 Physical evaluation

A physical evaluation of the two FLAIR images was performed. Signal intensity (SI) and standard deviation (SD) were measured by placing a 15-mm diameter ROI in the left and right ventricles and in the brain parenchyma (Fig. 3). The SI and SD of the CSF were defined as the average of the SI and SD values measured in the left and right ventricles. When measuring SI and SD in the brain parenchyma, the ROIs were placed carefully so that the CSF present in the sulci did not affect the measurement process. FLAIR images of the slice sections in which the ventricles could be seen most clearly were captured for analysis. The signal-to-noise ratio (SNR) of the CSF and contrast-to-noise ratio (CNR) between the CSF and brain parenchyma were calculated using Eqs (2) and (3) presented below. The SNR and CNR of the FLAIR images that were captured using the clinical and optimal TI were compared to evaluate the suppression of the CSF signal. The image analysis software used was ImageJ (National Institutes of Health, Bethesda, MD, USA, http://rsb.info.nih.gov/ij/). where SIa and SIb are the SI of the CSF and brain parenchyma, respectively, and SDa and SDb are the SD of the CSF and brain parenchyma, respectively.

$\displaystyle\textit{SNR}=SI_{a}/SD_{a}$ (2) $\displaystyle\textit{CNR}=({SI_{b}-SI_{a}})/SD_{b}$ (3)

Figure 3.

Placement of the ROI in the measurement of SI and SD. a. CSF present in the ventricles. b. Brain parenchyma. Abbreviations: ROI, region of interest; SI, signal intensity; SD, standard deviation; CSF, cerebrospinal fluid.

2.4.2 Visual evaluation

A visual evaluation was performed to compare the suppression of CSF signals in FLAIR images captured with clinical and optimal TI. A total of 10 observers (male to female ratio 8:2, mean age 36.2 years), including three physicians and seven radiological technologists, participated in the visual evaluation. The ethics committee of the Institute of Brain and Blood Vessels, Mihara Memorial Hospital, approved the observer experiment. The researcher explained the study to the observers and obtained written informed consent. A total of 30 FLAIR images comprised of the two different image types of 15 corpses were randomly displayed one at a time on a medical imaging display monitor (Exa Vision PLUS, Zio Soft Corporation, Tokyo, Japan), with the imaging parameters hidden. The slice section used for the visual evaluation was the one in which the ventricles and sulci were most clearly visible. The suppression of the intraventricular and sulcus CSF signals was scored at the following four levels.

1.
The CSF signals are not completely suppressed
2.
The CSF signals are not suppressed
3.
The CSF signals is suppressed to some extent
4.
The CSF signals is completely suppressed

Figure 4.
Reference images for the visual evaluation.

A reference image was presented to the observers in advance (Fig. 4), and they were provided with a sufficiently detailed explanation on how to evaluate the scores and were allowed to practice. The average and SD of the two FLAIR images were calculated from the scores assigned by 10 observers. We then evaluated whether there was a difference in the average scores of the two FLAIR images captured using the clinical TI and optimal TI.
2.5 Statistical analysis

Pearson’s correlation coefficient test was used to determine the association of the forehead temperature of the deceased persons with the measured T ${}_{1}$ value of the CSF We obtained the coefficient using the least squares method and derived the linear regression between T ${}_{1}$ value and forehead temperature. We then devised an optimization formula to calculate the optimal TI. Wilcoxon’s rank sum test was used to compare the SNR, CNR, and visual evaluation scores of the two FLAIR images. The statistical analysis was performed using the Easy R (EZR) software, and the level of significance was set at 5% [27].

3. Results

3.1 Relationship between forehead temperature and CSF T ${}_{1}$ value and optimal TI setting

Table 2
The details regarding the clinical data

Case	Time elapsed after death (hrs)	Clinical diagnosis	Forehead temperature ( ${}^{\circ}$ C)	Cerebrospinal fluid T1 value (ms)		Average T1 value (ms)	Inversion time (ms)
				Right side	Left side
1	14.4	Basal ganglia degeneration	22	3094	3081	3088	2144
2	6.5	Progressive supranuclear palsy	22	3270	3101	3186	2266
3	9.3	Creutzfeldt-Jakob disease	21	2945	2874	2910	2041
4	1.3	Amyotrophic lateral sclerosis	32	3650	3628	3639	2529
5	20.1	Supercentenarian	15	2570	2529	2550	1781
6	4.8	Supercentenarian	25	3215	3308	3262	2228
7	12.2	Creutzfeldt-Jakob disease	11	2130	2145	2138	1476
8	14.2	Supercentenarian	20	2887	2840	2864	2001
9	1.6	Pneumonia	30	3632	3496	3564	2517
10	6.9	Prion disease	10	2021	2146	2084	1401
11	4.9	Parkinson’s disease	23	3275	3129	3202	2270
12	3.2	Alzheimer’s disease	25	3280	3152	3216	2273
13	1.8	Parkinson’s disease	35	3806	3892	3849	2638
14	10.4	Basal ganglia degeneration	18	2716	2714	2715	1882
15	18.5	Progressive supranuclear palsy	14	2519	2469	2494	1746

Figure 5.

The relationship between the T ${}_{1}$ value of the CSF and forehead temperature. The T ${}_{1}$ value of the CSF is smaller in cases in which forehead temperatures are lower and is higher in cases involving higher forehead temperatures. Abbreviations: T ${}_{1}$ , longitudinal relaxation time; CSF, cerebrospinal fluid.

The details regarding the clinical data of the subjects and T ${}_{1}$ value and optimal TI of the CSF are presented in Table 2. The age of the subjects ranged from 49–113 years (average: 88.7, male: female $=$ 9: 6), and the time elapsed since death ranged from 1.3–20.1 h (average: 8.7 h). The T ${}_{1}$ values of the CSF ranged from 2,084 to 3,849 ms (2,984 $\pm$ 503 ms). Figure 5 shows the relationship between the T ${}_{1}$ value of the CSF and forehead temperature. There was a very strong correlation between the T ${}_{1}$ value of the CSF and forehead temperature ( $r=$ 0.95, $p<$ 0.01). We determined the regression formula for the T ${}_{1}$ value as Eq. (4). Then we devised the optimization Eq. (5) to calculate the optimal TI.

$\displaystyle T_{1}=67.9\times\textit{Forehead temperature}+1466$ (4) $\displaystyle TI=T_{1}\times 0.693=({67.9\times\textit{Forehead temperature}+1% 466})\times 0.693$ (5)

3.2 Physical evaluation of FLAIR images

The cases corresponding to the two types of FLAIR images (forehead temperatures of 10 ${}^{\circ}$ C, 15 ${}^{\circ}$ C, 20 ${}^{\circ}$ C, 25 ${}^{\circ}$ C, 30 ${}^{\circ}$ C, and 35 ${}^{\circ}$ C) are presented in Fig. 6, and the SNR and CNR of each case are compared in Figs 7 and 8. The FLAIR images obtained using the clinical TI setting exhibited the suppression of the CSF signal when the temperature was close to that of the living body, and failure of CSF signal suppression occurred as the temperature decreased. On the other hand, the FLAIR images obtained using the optimal TI setting resulted in the suppression of CSF signals in all cases. The SNR values were lower for the FLAIR images obtained with the optimal TI setting than for those obtained with the clinical TI setting in all the cases. The CNR values of the FLAIR images obtained with the optimal TI setting were higher than those of the images obtained with the clinical TI setting in all cases. Comparing the SNR and CNR of the two types of FLAIR images, the difference in the SNR and CNR values was found to increase as the temperature decreased. A comparison of the average calculated SNR and CNR values of the two types of FLAIR images in all the cases is shown in Figs 9 and 10. The average values of SNR and CNR were statistically significant ( $p<$ 0.01), and the FLAIR images obtained with the optimal TI resulted in a lower SNR and a higher CNR.

Figure 6.

The two types of FLAIR images captured in all the cases (forehead temperature: 10 ${}^{\circ}$ C, 15 ${}^{\circ}$ C, 20 ${}^{\circ}$ C, 25 ${}^{\circ}$ C, 30 ${}^{\circ}$ C, and 35 ${}^{\circ}$ C). The upper half of the figure depicts the FLAIR images obtained with the clinical TI setting showing the suppression of the CSF signal when the temperature was close to the temperature of the living body, and the failure of CSF signal suppression occurred as the temperature decreased. The bottom half of the figure depicts the FLAIR images obtained with the optimal TI setting resulting in the suppression of CSF signals in all cases. Abbreviations: FLAIR, fluid-attenuated inversion recovery; CSF, cerebrospinal fluid; TI, inversion time.

Figure 7.

Comparison of the SNR in each case (forehead temperature: 10 ${}^{\circ}$ C, 15 ${}^{\circ}$ C, 20 ${}^{\circ}$ C, 25 ${}^{\circ}$ C, 30 ${}^{\circ}$ C, and 35 ${}^{\circ}$ C). The SNR values were lower for the FLAIR images obtained with the optimal TI setting than for those obtained with the clinical TI setting in all the cases. Abbreviations: SNR, signal-to-noise ratio; FLAIR, fluid-attenuated inversion recovery; CSF, cerebrospinal fluid.

Table 3

The difference between the average and standard deviation of the visual evaluation scores (forehead temperature: 10 ${}^{\circ}$ C, 15 ${}^{\circ}$ C, 20 ${}^{\circ}$ C, 25 ${}^{\circ}$ C, 30 ${}^{\circ}$ C, and 35 ${}^{\circ}$ C) of the two FLAIR images in each case

Forehead temperature [ ${}^{\circ}$ C]	10	15	20	25	30	35
Clinical TI	1.0 $\pm$ 0.0	1.0 $\pm$ 0.0	1.7 $\pm$ 0.5	2.2 $\pm$ 0.4	2.7 $\pm$ 0.5	3.8 $\pm$ 0.4
Optimal TI	3.9 $\pm$ 0.3	4.0 $\pm$ 0.0	4.0 $\pm$ 0.0	4.0 $\pm$ 0.0	4.0 $\pm$ 0.0	4.0 $\pm$ 0.0
Differences	2.9	3.0	2.3	1.8	1.3	0.2
Wilcoxon rank sum test	$p<$ 0.01	$p<$ 0.01	$p<$ 0.01	$p<$ 0.01	$p<$ 0.01	n.s.

n.s.: not significant. Abbreviations: FLAIR, fluid-attenuated inversion recovery.

Figure 8.

Comparison of CNR in each case (forehead temperature: 10 ${}^{\circ}$ C, 15 ${}^{\circ}$ C, 20 ${}^{\circ}$ C, 25 ${}^{\circ}$ C, 30 ${}^{\circ}$ C, and 35 ${}^{\circ}$ C). The CNR values were higher for the FLAIR images obtained with the optimal TI setting than for those obtained with the clinical TI setting in all the cases. Abbreviations: CNR, contrast-to-noise ratio; FLAIR, fluid-attenuated inversion recovery; TI, inversion time.

Figure 9.

Comparison of the average SNR values of the two different FLAIR images in all the cases. Abbreviations: SNR, signal-to-noise ratio; FLAIR, fluid-attenuated inversion recovery.

Figure 10.

Comparison of the average CNR values of the two different FLAIR images in all the cases. Abbreviations: CNR, contrast-to-noise ratio FLAIR, fluid-attenuated inversion recovery.

Figure 11.

Comparison of the average scores (forehead temperature: 10 ${}^{\circ}$ C, 15 ${}^{\circ}$ C, 20 ${}^{\circ}$ C, 25 ${}^{\circ}$ C, 30 ${}^{\circ}$ C, and 35 ${}^{\circ}$ C) of the two FLAIR images in each case. The average score of the FLAIR images captured using the clinical TI setting reduced as the forehead temperature decreased. The average score of the FLAIR images captured using the optimal TI setting was three or higher in all cases. Abbreviations: FLAIR, fluid-attenuated inversion recovery; TI, inversion time.

Figure 12.

Comparison of the average scores of the two different FLAIR images in all cases. Abbreviations: FLAIR, fluid-attenuated inversion recovery.

3.3 Visual evaluation of FLAIR images

The average and SD of the visual evaluation scores (forehead temperature: 10 ${}^{\circ}$ C, 15 ${}^{\circ}$ C, 20 ${}^{\circ}$ C, 25 ${}^{\circ}$ C, 30 ${}^{\circ}$ C, and 35 ${}^{\circ}$ C) of the two FLAIR images in each case and the differences between the average and SD of the visual evaluation scores are presented in Table 3 and Fig. 11, respectively. The average score of the FLAIR images captured using the clinical TI reduced as the forehead temperature decreased. The average score of the FLAIR images captured using the optimal TI was three or higher in all cases. In addition, the difference in the average scores between the two types of FLAIR images increased as the temperature decreased, and the statistical analysis revealed a significant difference in all the cases when the temperature of the deceased person was less than 35 ${}^{\circ}$ C ( $p<$ 0.01). A comparison of the average scores of the two FLAIR images in all the cases is presented in Fig. 12. The average visual assessment scores of the two FLAIR images in all the cases were lower on using the clinical TI setting and higher on using the optimal TI setting, and the statistical analysis revealed a significant difference ( $p<$ 0.01).

4. Discussion

Table 2 shows that the T ${}_{1}$ value of the CSF measured in the deceased subjects ranged from 2,084 to 3,849 ms (2,984 $\pm$ 503 ms), which is shorter than the T ${}_{1}$ value of the CSF in a living body. This is because deceased persons are cooler than the living body due to the temperature decrease that occurs after death, as reported in previous studies [1, 17, 18]. Figure 5 shows that the T ${}_{1}$ value of the CSF at 35 ${}^{\circ}$ C, which is close to the temperature of the living body, is approximately 3,900 ms. Said value decreases by approximately 500 ms for each 10 ${}^{\circ}$ C decrease in temperature. This suggests that the T ${}_{1}$ value of the CSF is greatly affected by temperature changes. Therefore, it is very important to calculate the optimal TI corresponding to the temperature decrease in the CSF. Since the T ${}_{1}$ values of the CSF exhibited a strong correlation with the forehead temperature of the deceased persons, we devised the regression equation for the T ${}_{1}$ values and the optimization formula to calculate the optimal TI. By using Eqs (4) and (5), it is possible to estimate the approximate T ${}_{1}$ and optimal TI values of the CSF using only simple body surface temperature measurements. However, because the tissue relaxation time varies with different strengths of the static magnetic field, this is the only indication for the 3 Tesla MRI used in this study. In addition, since the measured relaxation time differs slightly from one device to another regardless of whether the strength of the static magnetic field is the same, the performance of each device needs to be verified. The differences in T ${}_{1}$ values between different devices is negligible compared to the difference resulting from temperature decreases. Therefore, we consider that the difference in the type of equipment does not need to be taken into account to a large extent if the strength of the static magnetic field is the same for 3 Tesla MRI.

As shown in Figs 7 and 8, on comparing the SNR and CNR of the two FLAIR images in each case (forehead temperature: 10 ${}^{\circ}$ C, 15 ${}^{\circ}$ C, 20 ${}^{\circ}$ C, 25 ${}^{\circ}$ C, 30 ${}^{\circ}$ C, and 35 ${}^{\circ}$ C), the CSF signal on the FLAIR images captured with the clinical TI setting was suppressed in cases involving temperatures of 30 ${}^{\circ}$ C or more, and the differences in the SNR and CNR values were small. This finding suggests that the effect of the temperature decrease due to the time elapsed since death is less pronounced in cases involving temperatures of above 30 ${}^{\circ}$ C. However, as the body temperature of the deceased person decreased (below 30 ${}^{\circ}$ C), the difference in the SNR and CNR values between the two types of FLAIR images increased. As seen in Fig. 9, a significant difference between the average SNR values of the two types of FLAIR images ( $p<$ 0.01) was found. The CNR results also exhibited a similar trend as that of the SNR results, as shown in Fig. 1. This suggests that the CSF signal was appropriately suppressed on using the optimal TI setting even in cases in which the body temperature of the deceased person was different. Table 3 and Fig. 12 show that the average score of the FLAIR images captured with the optimal TI setting was more than three in all the cases, and the difference between the average scores of the images captured with the clinical TI and optimal TI settings tended to increase as the body temperature of the deceased person decreased. This finding indicates that the failure of CSF signal suppression becomes more pronounced as the body temperature decreases. In case a short time period elapsed since death, FLAIR imaging using the clinical TI setting does not significantly affect CSF signal suppression. However, in case a long time period elapsed since death or if the body were stored under refrigeration, a significant decrease in body surface temperature would occur. Therefore, it is possible to obtain FLAIR images with suppressed CSF signals effectively and rapidly by using the method proposed in this study. In other words, it was considered possible to optimize the postmortem brain FLAIR imaging method by measuring the shortened T ${}_{1}$ value of the CSF resulting from body temperature decreases after death using the synthetic MRI function and by calculating the optimal TI. Unlike the results of an autopsy, those of “Ai” are not affected by the procedure or by the condition of the deceased person, and useful clinical findings can be preserved semi-permanently by optimizing the imaging parameters. Thus, it is a useful modality that can be used to elucidate unknown pathologies. Postmortem brain FLAIR imaging is a highly sensitive method for detecting signal changes caused by a wide range of lesions, such as those resulting from periventricular demyelination and inflammatory diseases, owing to its ability to suppress the CSF signal and to produce images with high T ${}_{2}$ contrast [13, 14]. As an excellent diagnostic imaging method, its use is important in comparing imaging findings with histopathological findings in studies of brain tissue [15, 16].

In this study, we focused on the use of the synthetic MRI function for quickly and easily measuring relaxation times at arbitrary regions and thereby optimized the postmortem brain FLAIR imaging method. As a result, we found that synthetic MRI enabled us to measure the T ${}_{1}$ value of the CSF and to calculate the optimal TI effectively. However, in this study, only the use of synthetic MRI was investigated as a relaxation time measurement method, so it is impossible to perform this method in facilities that do not have synthetic MRI. Therefore, in this study, we devised the optimization formula to calculate the optimal TI. However, since we did not evaluate FLAIR images using the optimization formula in this study, the formula is not reliable. In the future, we consider that it is necessary to evaluate postmortem brain FLAIR images obtained using the optimization formula.

In this study, the forehead temperature of the deceased subjects was assumed to be the temperature of the CSF. The core or rectal temperature measured at the time of autopsy is usually adopted as the temperature of the deceased person. However, this was not done in our research facilities. In this study, we measured the forehead temperature using a non-contact thermometer, which is an easy method. Although the correlation between forehead temperature and the T ${}_{1}$ value of the CSF was found to be very strong, it is necessary to choose a highly accurate method for measuring the temperature of the CSF in order to propose a reliable optimization formula.

5. Conclusion

Our results revealed that postmortem brain FLAIR imaging can be optimized by using synthetic MRI. These results suggest that FLAIR images could potentially contribute to the elucidation of the pathogenesis of unknown neurological diseases and to the development of novel treatment methods. In the future, MRI is expected to play a central role in “Ai” in order to achieve higher diagnosis and autopsy success rates and for the further advancement of histological and pathological research.

Footnotes

Acknowledgments

The authors would like to express their gratitude to the families of the deceased for their consent to be involved in this study.

Conflict of interest

The authors declare that they have no conflicts of interest.

References

Kobayashi

Shiotani

Kaga

, et al. Characteristic signal intensity changes on postmortem magnetic resonance imaging of the brain. Jpn J Radiol 2010; 28: 8-14.

Okuda

Shiotani

Sakamoto

Kobayashi

. Background and current status of postmortem imaging in Japan: Short history of Autopsy imaging (Ai). Forensic Science International 2013; 225: 3-8.

Shiotani

Yamazaki

Kikuchi

, et al. Postmortem magnetic resonance imaging (PMMRI) demonstration of reversible injury phase myocardium in a case of sudden death from acute coronary plaque change. Radiat Med 2005; 23: 563-565.

Scola

Conte

Palumbo

, et al. High resolution post-mortem MRI of non-fixed in situ foetal brain in the second trimester of gestation; Normal foetal brain development. Eur Radiol 2018; 28: 363-371.

Dawe

Leurgans

, et al. Postmortem MRI; A novel window into the neurobiology of late life cognitive decline. Neurobiol Aging 2016; 45: 169-177.

Niwa

Shindo

, et al. Comparative Analysis of Cortical Microinfarcts and Microbleeds using 3.0-Tesla Postmortem Magnetic Resonance Images and Histopathology. J Alzheimers Dis 2017; 59: 951-959.

Popescu

Klaver

Versteeg

, et al. Postmortem validation of MRI cortical volume measurements in MS. Hum Brain Mapp 2016; 37: 2223-2233.

Calabrese

Hickey

Hulette

, et al. Postmortem diffusion MRI of the human brainstem and thalamus for deep brain stimulator electrode localization. Hum Brain Mapp 2015; 36: 3167-3178.

Jonkman

Soriano

Amor

, et al. Can MS lesion stages be distinguished with MRI? A postmortem MRI and histopathology study. J Neurol 2015; 262: 1074-1080.

10.

Kanawaku

Someya

Kobayashi

, et al. High resolution 3D-MRI of postmortem brain specimens fixed by formalin and gadoteridol. Leg Med (Tokyo) 2014; 16: 218-221.

11.

Kilsdonk

Jonkman

Klaver

, et al. Increased cortical grey matter lesion detection in multiple sclerosis with 7T; MRI; A post-mortem verification study. Brain 2016; 139: 1472-1481.

12.

Van Veluw

Zwanenburg

Rozemuller

, et al. The spectrum of MR detectable cortical microinfarcts; A classification study with 7-tesla postmortem MRI and histopathology. J Cereb Blood Flow Metab 2015; 35: 676-683.

13.

De Coene

Hajnal

Gatehouse

, et al. MR of the brain using fluid-attenuated inversion recovery (FLAIR) pulse sequences. AJNR 1992; 13: 1555-1564.

14.

Hajnal

Bryant

, Larry Kasuboski, et al. Use of Fluid-Attenuated Inversion Recovery (FLAIR) Pulse Sequences in MRI of the Brain. J Comput Assist Tomogr 1992; 16: 841-844.

15.

white

Hajnal

Young

Bydder

. Use of fluid-attenuated inversion recovery pulse sequences for imaging in the spinal cord. Magn Reson Med 1992; 28: 153-162.

16.

Stevenson

Gawne-Cain

Barker

Thompson

Miller

. Imaging of the spinal cord and brain in multiple sclerosis: A comparative study between fast flair and fast spin echo. J Neurol 1997; 224: 119-124.

17.

Abe

Kobayashi

Shiotani

, et al. Optimization of Inversion Time for Postmortem Fluid-Attenuated Inversion Recovery (FLAIR) MR Imaging at 1.5T: Temperature-based Suppression of Cerebrospinal Fluid. Magn Reson Med Sci 2015; 4: 251-255.

18.

Kobayashi

Isobe

Shiotani

, et al. Postmortem magnetic resonance imaging dealing with low temperature objects. Magn Reson Med Sci 2010; 9: 101-108.

19.

Bloembergen

Purcell

Pound

. Relaxation effects in nuclear magnetic resonance absorption. Phys Rev 1948; 73: 679-721.

20.

de Bazelaire

Duhamel

Rofsky

, et al. MR imaging relaxation times of abdominal and pelvic tissues measured in vivo at 3.0. T: preliminary results. Radiology 2004; 230: 652-659

21.

Tofts

Jackson

Tozer

, et al. Imaging cadavers: Cold FLAIR and noninvasive brain thermometry using CSF diffusion. Magn Reson Med 2008; 59: 190-195.

22.

Ruder

Hatch

Siegenthaler

, et al. The influence of body temperature on image contrast in postmortem MRI. Eur J Radiol 2012; 81: 1366-1370.

23.

Warntjes

Leinhard

West

, et al. Rapid magnetic resonance quantification on the brain: Optimization for clinical usage. Magn Reson Med 2008; 60(2): 320-329.

24.

Krauss

Gunnarsson

Andersson

Thunberg

, et al. Accuracy and reproducibility of a quantitative magnetic resonance imaging method for concurrent measurements of tissue relaxation times and proton density. Magn Reson Imaging 2015; 33: 584-591.

25.

Hagiwara

Warntjes

Hori

, et al. SyMRI of the Brain Rapid Quantification of Relaxation Rates and Proton Density, With Synthetic MRI, Automatic Brain Segmentation, and Myelin Measurement. Invest Radiol 2017; 52: 647-657.

26.

Sukhoon

Andrew

Thomas

BuSik

Christopher

. Experimental and numerical assessment of MRI-induced temperature change and SAR distributions in phantoms and in vivo . Magn Reson Med 2010; 63: 218-223.

27.

Kanda

. Investigation of the freely available easy-to-use software ‘EZR’ for medical statistics. Bone Marrow Transplantation 2013; 48: 452-458.

Optimization of the fluid-attenuated inversion recovery (FLAIR) imaging for use in autopsy imaging of the brain region using synthetic MRI

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSIONS:

Keywords

1. Introduction

2.1 Subjects

2.2 Body temperature measurement and evaluation

2.3 MR image acquisition

2.4.1 Physical evaluation

3. Results

3.1 Relationship between forehead temperature and CSF T 1 value and optimal TI setting

Table 2 The details regarding the clinical data

4. Discussion

5. Conclusion

Footnotes

Acknowledgments

Conflict of interest

References

3.1 Relationship between forehead temperature and CSF T ${}_{1}$ value and optimal TI setting

Table 2
The details regarding the clinical data