Published online Mar 18, 2022.
https://doi.org/10.3348/kjr.2021.0477
Hyperoxia-Induced ΔR1: MRI Biomarker of Histological Infarction in Acute Cerebral Stroke
Abstract
Objective
To evaluate whether hyperoxia-induced ΔR1 (hyperO2ΔR1) can accurately identify histological infarction in an acute cerebral stroke model.
Materials and Methods
In 18 rats, MRI parameters, including hyperO2ΔR1, apparent diffusion coefficient (ADC), cerebral blood flow and volume, and 18F-fluorodeoxyglucose uptake on PET were measured 2.5, 4.5, and 6.5 hours after a 60-minutes occlusion of the right middle cerebral artery. Histological examination of the brain was performed immediately following the imaging studies. MRI and PET images were co-registered with digitized histological images. The ipsilateral hemisphere was divided into histological infarct (histological cell death), non-infarct ischemic (no cell death but ADC decrease), and non-ischemic (no cell death or ADC decrease) areas for comparisons of imaging parameters. The levels of hyperO2ΔR1 and ADC were measured voxel-wise from the infarct core to the non-ischemic region. The correlation between areas of hyperO2ΔR1-derived infarction and histological cell death was evaluated.
Results
HyperO2ΔR1 increased only in the infarct area (p ≤ 0.046) compared to the other areas. ADC decreased stepwise from non-ischemic to infarct areas (p = 0.002 at all time points). The other parameters did not show consistent differences among the three areas across the three time points. HyperO2ΔR1 sharply declined from the core to the border of the infarct areas, whereas there was no change within the non-infarct areas. A hyperO2ΔR1 value of 0.04 s-1 was considered the criterion to identify histological infarction. ADC increased gradually from the infarct core to the periphery, without a pronounced difference at the border between the infarct and non-infarct areas. Areas of hyperO2ΔR1 higher than 0.04 s-1 on MRI were strongly positively correlated with histological cell death (r = 0.862; p < 0.001).
Conclusion
HyperO2ΔR1 may be used as an accurate and early (2.5 hours after onset) indicator of histological infarction in acute stroke.
INTRODUCTION
Currently, the early restoration of blood flow to non-infarct ischemic areas is the optimal treatment option to rescue brain tissue at risk of irreversible damage following acute stroke. However, many stroke patients experience neurological deterioration after successful vascular recanalization, even without hemorrhagic transformation [1]. These undesirable results imply that initial imaging examination may underestimate the area of irreversible damage; that is, infarction, or that progressive neuronal damage may occur even after the successful reperfusion of salvageable tissue. Therefore, an accurate imaging parameter is needed to measure the area of infarction to select adequate treatment and monitor treatment outcomes in acute stroke.
Although numerous imaging techniques have been used to identify infarction in acute stroke, they have limitations in accurately measuring the infarct area. The apparent diffusion coefficient (ADC) calculated from diffusion-weighted imaging (DWI) is often renormalized during the aggravation of ischemic injury after recanalization, and perfusion-weighted imaging parameters do not indicate cell viability after vascular recanalization [2, 3, 4].
Hyperoxia-induced ΔR1 (hyperO2ΔR1) is a feasible MRI biomarker of tissue oxygen level [5, 6, 7]. Theoretically, this parameter is based on the oxygen-driven acceleration of longitudinal relaxation (quantified by increased R1); thus, the difference in R1 between hyperoxic and normoxic breathing (i.e., hyperO2ΔR1) indicates the amount of oxygen accumulation during hyperoxic respiration and, therefore, demonstrates the status of oxygen delivery and consumption in tissue. Recently, Suh et al. [7] suggested that the tissue oxygenation status quantified by hyperO2ΔR1 can be used to classify ischemic damage into exacerbating and devastated stages in the late phase (24 hours) following stroke onset. Based on this theoretical background, we hypothesized that hyperO2ΔR1 may be useful in identifying irreversible ischemic damage, such as cell death, in acute stroke.
Of the various methods used to evaluate the severity of tissue impairment in acute stroke, morphological cell death by light microscopy is most frequently used to identify the area of infarction. Therefore, a meaningful correlation between morphological cell death and hyperO2ΔR1 may indicate that this MRI-driven parameter can be used to identify irreversible ischemic damage in acute stroke. In this study, we evaluated whether hyperO2ΔR1 can accurately identify histological infarction in an acute cerebral stroke model.
MATERIALS AND METHODS
All experimental protocols, including the animal care and experimental procedures in this study, were approved by the Institutional Animal Care and Use Committee of Asan Medical Center.
Transient Stroke Model
The right middle cerebral arteries (MCAs) of adult male Wistar rats (weight, 280–300 g) were occluded for 60 minutes using the intraluminal filament technique, as previously described [8]. The details are summarized in the Supplement.
MRI, 18F-FDG PET, and Histology
MRI and PET were performed at three time points (2.5 hours, 4.5 hours, and 6.5 hours) following the initiation of transient stroke. The first image acquisition time of 2.5 hours was assigned for the stabilization of animal physiology, transfer, and fixation to the bed of imaging machines after surgery. Then, the time window between imaging experiments was arbitrarily assigned as 2 hours The histological examinations were performed immediately after imaging.
The MRI parameters were acquired using a 7T Bruker pre-clinical MRI scanner (PharmaScan 70/16, Brucker BioSpin GmbH), which included T2-weighted images (T2WIs), ADC images from DWI, hyperO2ΔR1 from R1 mapping, and cerebral blood flow (CBF) and volume (CBV) from dynamic susceptibility contrast imaging. The MRI acquisition parameters are listed in Table 1. 18F-FDG PET was obtained before MRI using a nanoScanPET/MRI system (1T, Mediso), which provided the standard uptake value. Further details are provided in the Supplement.
Table 1
MRI Acquisition Parameters
Light microscopy histological examination was performed on the mid-coronal sections of the brain. Each brain section was stained with hematoxylin and eosin (H&E) and 0.25% cresyl violet for Nissl staining [9].
Based on the histologic findings and acceptable MRI and PET quality, this study included 18 rats with stroke, including six rats at each time point, which demonstrated histological infarction in more than one-third of the ipsilateral hemisphere.
Calculation of HyperO2ΔR1
To measure hyperO2ΔR1, normoxic (20% oxygen) and hyperoxic (100% oxygen) gases were infused through a respiratory mask without mechanical ventilation. The gas flow was maintained at a constant rate (1 L/min at 1 atm) during the experiment. Hyperoxia-induced ΔR1 was calculated according to the difference in R1 between hyperoxic and normoxic breathing. Further details are provided in the Supplement.
Co-Registration of Images to the Histological Map
Using T2WI as a reference frame, all parameter maps were co-registered and interpolated to achieve the same voxel geometry. For image-histology correlation, slices of MRI and PET images were selected and registered to the corresponding digitized histological map.
Image Analysis
The image analysis process is shown in Figure 1. The ipsilateral hemisphere was categorized into three areas based on the conjunction of the incorporated histological and ADC information. In this approach, the ADC value was used to identify the occurrence of ischemia, whereas histologic findings indicated irreversible severe damage (i.e., cell death). The histological findings were used to determine the viability of the given cells. Cells were identified as dead according to the presence of cell body shrinkage, darkly stained pyknotic nuclei, intensely stained red eosinophilic cytoplasm in H&E staining, and loss of Nissl substance in Nissl staining [10, 11]. An ipsilateral-to-contralateral ratio of ADC < 0.95 indicated ischemic cell damage, as described previously [7]. These criteria of cell death or damage on ADC were then used to classify the ipsilateral hemisphere into three areas: histological infarct (cell death), non-infarct ischemic (cell damage without cell death), and non-ischemic (no cell death or damage). Regions-of-interest (ROIs) corresponding to the three areas were drawn on MRI and PET images co-registered to the histological map. The values of each imaging parameter were then compared among the three ROIs.
Fig. 1
Image analysis.
A. The ipsilateral hemisphere was categorized into infarct, ischemic, and non-ischemic areas based on histological findings and ADC values. Areas with histologic findings indicated cell death, including cell body shrinkage, darkly stained pyknotic nuclei in H&E stain (× 20), and a loss of Nissl substance in Nissl staining (× 20), were considered infarct. The imaging parameters were then compared in these three areas. B. Voxel-wise line analysis of hyperO2ΔR1 and ADC values from the center of the infarct (outlined by blue in H&E stain [× 20] and Nissl stain [× 20], representing histologic cell death) to the non-ischemic area. Here, areas with ipsilateral-to-contralateral ratios of ADC < 0.95 are outlined by a black dashed line, and the area between the blue (i.e., cell death) and black dashed lines define non-infarct ischemia. On MRI, a line is drawn from the infarct center to the periventricular non-ischemic region on images co-registered to the histologic map. Every line is drawn to cross the three areas with identical proportions: 60% of the total length of each line crossed the infarct area, 20% crossed the non-infarct ischemic area, and the remaining 20% crossed the non-ischemic area. ADC = apparent diffusion coefficient, CBF = cerebral blood flow, FDG = fluorodeoxyglucose, H = hours, H&E = hematoxylin and eosin, hyperO2ΔR1 = hyperoxia-induced ΔR1
The delineation of the border between infarct and non-infarct areas requires setting a cutoff value to differentiate these areas. Accordingly, we performed a voxel-scaled line study to continuously measure hyperO2ΔR1 and ADC. In each rat, we drew a line from the infarct center to the periventricular non-ischemic region on images co-registered with the histologic map. Every line was drawn to proportionally identically cross the three areas: 60% of the total length of each line crossed the infarct area, 20% crossed the non-infarct ischemic area, and the remaining 20% crossed the non-ischemic area. This approach allowed the integrated analysis of the changes in hyperO2ΔR1 and ADC from the infarct core to the non-ischemic area in all rats. In addition, the correlations between the image parameters measured on the lines were also analyzed.
Finally, to validate whether the hyperO2ΔR1 could be used to locate the infarct area, we assessed the spatial correlation between hyperO2ΔR1 and histological features. The area of infarction indicated by hyperO2ΔR1 was measured and its ratio in the area of the ipsilateral hemisphere was calculated. The area ratio of cell death lesions in the ipsilateral hemisphere was also estimated on the histological map.
Statistical Analysis
The Friedman test was used to compare each value across the three areas. In cases of statistical significance, Wilcoxon signed-rank tests were performed for post-hoc analysis. Spearman correlation was used to evaluate the correlations between the areas of hyperO2ΔR1-derived infarction and histological cell death. Statistical significance was set at p < 0.05.
RESULTS
MRI and PET Parameters according to Histology at Each Time Point
Figure 2 and Table 2 show the values of imaging parameters measured in the histological infarct, non-infarct ischemic, and non-ischemic areas at the three time points (2.5, 4.5, and 6.5 hours). Compared to the other parameters, hyperO2ΔR1 showed a distinctive pattern in terms of discrimination of the infarct area, with increased values occurring exclusively in the infarct area (p ≤ 0.046) compared to the other areas (Supplementary Table 1). This pattern was consistently observed at all three time points. In contrast, the ADC values demonstrated a pattern of stepwise decrease from the non-ischemic to infarct areas at all time points (p = 0.002 at all time points).
Fig. 2
Comparisons of hyperO2ΔR1 and ADC in infarct, ischemic, and non-ischemic areas.
HyperO2ΔR1 increases only in the infarct area at all three time points. In contrast, the ADC values demonstrate a pattern of stepwise decrease from the non-ischemic to infarct areas at all time points. *Statistically significant difference between infarct and ischemia, †Statistically significant difference between infarct and non-ischemia, ‡Statistically significant difference between ischemia and non-ischemia. ADC = apparent diffusion coefficient, hyperO2ΔR1 = hyperoxia-induced ΔR1
Table 2
MRI and PET Parameters in Infarct, Ischemia, and Non-Ischemia Areas
18F-FDG uptake, CBF, and CBV did not show consistent differences among the three areas across the three time points (Supplementary Fig. 1). 18F-FDG uptake was higher in the infarct area than in the other areas at 2.5 hours, but not at 4.5 and 6.5 hours. Likewise, differences in CBF and CBV were observed only at 2.5 hours (p = 0.009 and 0.011, respectively), when those values were lower in the infarct area compared to the other areas. The CBF value was lower in the infarct area at 4.5 hours (p = 0.042) compared to the non-ischemic area, although the difference was not significant compared to the ischemic area. The lack of significant between-area differences in the CBF and CBV levels at 4.5 and 6.5 hours indicated that the blood supply was restored and evenly distributed across all areas after the transient occlusion of the MCA. Given the inconsistent patterns or differences across the time points, the parameters of 18F-FDG, CBF, and CBV were inadequate for use as reproducible biomarkers for assessment of the degree of ischemic injury.
Voxel-Wise Line Analysis of HyperO2ΔR1 and ADC
Based on the ROI experiments, hyperO2ΔR1 and ADC were chosen as imaging parameters for further investigation to discriminate areas of irreversible and reversible ischemic damage.
hHyperO2ΔR1 and ADC showed different distributions in voxels arranged on lines drawn from the infarct core to the non-ischemic area (Fig. 1B). HyperO2ΔR1 showed a sharp decline in the voxel values moving from the core to the border of the infarct areas, whereas there was no change within the non-infarct areas (Fig. 3A). HyperO2ΔR1 was > 0.04 s-1 in most of the voxels included in the infarct area and < 0.04 s-1 in the voxels in the non-infarct areas. We observed no notable difference in hyperO2ΔR1 between the non-infarct ischemic and non-ischemic areas. In contrast, ADC showed a pattern of gradual increase from the infarct core to the periphery without a pronounced difference at the border between the infarct and non-infarct ischemic areas (Fig. 3B). Thus, it was difficult to assign a clear ADC cutoff value to clearly define the border between the infarct and non-infarct areas.
Fig. 3
Voxel-wise line analysis of hyperO2ΔR1 and ADC values from the center of the infarct to the non-ischemic area.
A. Scatterplot demonstrating a clear transition in hyperO2ΔR1 values at the border of the infarct core and non-infarct ischemia area at a cutoff of 0.4 s-1. B. Monotonic increase in ADC values from the infarct core to the non-ischemic area, with no clear boundary between areas. C. Scatterplot demonstrating the negative correlation between hyperO2ΔR1 and ADC values below the hyperO2ΔR1 cutoff of 0.04 s-1, and a plateaued distribution of ADC values above 0.04 s-1. ADC = apparent diffusion coefficient, hyperO2ΔR1 = hyperoxia-induced ΔR1
The correlation between hyperO2ΔR1 and ADC showed a dichotomized pattern, as shown in Figure 3C. In the voxels with hyperO2ΔR1 < 0.04 s-1, most of which were identified in the non-infarcted and non-ischemic areas, hyperO2ΔR1 and ADC showed a strong negative correlation. In contrast, the voxels with hyperO2ΔR1 > 0.04 s-1, which were observed mostly in the infarct area, showed plateaued ADC values, regardless of the hyperO2ΔR1 values. This dichotomized distribution was consistently observed at all time points after stroke onset. The negative correlation between the two parameters demonstrated that cell edema gradually progressed because of continual energy failure, while the plateaued correlation reflects the fact that brain tissue with hyperO2ΔR1 > 0.04 > s-1 was maintained under the highest degree of cellular edema or restricted cytoplasmic movement. Therefore, we presume that a hyperO2ΔR1 of 0.04 s-1 could indicate complete membrane failure of ischemic cells in acute stroke.
Correlation between HyperO2ΔR1 Area and Histological Infarct Area
Areas of hyperO2ΔR1 > 0.04 s-1 were strongly positively correlated with those of histological cell death (r = 0.862; p < 0.001) (Fig. 4). On visual inspection, the areas of hyperO2ΔR1 > 0.04 s-1 spatially corresponded with the histologically determined areas of cell death. Thus, hyperO2ΔR1 was an accurate indicator of the location of the area of histological infarct.
Fig. 4
Strong positive correlation between the histologically defined infarct area and the area with hyperO2ΔR1 greater than 0.04 s-1 (r = 0.862).
hyperO2ΔR1 = hyperoxia-induced ΔR1
DISCUSSION
In our study, hyperO2ΔR1 values increased at various time points only in the areas with histological cell death and differed notably at the borders between cell death and other areas. These results demonstrated that hyperO2ΔR1 can consistently identify the cell death area in any time window of acute stroke. Our findings subsequently showed that areas with hyperO2ΔR1 values > 0.04 s-1 were very similar to areas showing cell death on histological analysis. The results of the cross-sectional study validated that a hyperO2ΔR1 value > 0.04 s-1 is a reliable indicator of histological cell death; that is, irreversible ischemic damage, in stroke. A longitudinal study is needed to track the fate of an area with hyperO2ΔR1 values > 0.04 s-1 in each rat to more accurately demonstrate the ability of this criterion to differentiate irreversible from reversible ischemic damage.
The hyperO2ΔR1 value is determined by the balance between oxygen delivery and consumption in the tissue. In cells under hypoxic damage in acute stroke, cellular oxygen utilization for energy production remains hypoactivated, even after the blood supply is restored. This condition leads to an accumulation of unconsumed oxygen during hyperoxic respiration, which consequently increases hyperO2ΔR1 (i.e., the difference in R1 between hyperoxic and normoxic respiration). Therefore, increased hyperO2ΔR1 in the transient ischemic brain is an MRI phenotype reflecting impaired oxygen metabolism.
As mitochondria consume oxygen for energy production, hyperO2ΔR1 quantification of unconsumed oxygen indicates the degree of mitochondrial dysfunction. Various harmful processes in mitochondria, such as altered biogenetics, disorganized structural architecture, and aberrant biochemical dynamics, play a pivotal role in neuronal death in stroke [12]. Mitochondrial damage leads not only to the depletion of energy production but also to an overproduction of reactive oxygen species. In addition to the collapse of electrophysiological homeostasis due to a lack of energy, reactive oxygen species induce pathological processes that cause neuronal death, such as caspase activation, the release of inflammatory mediators, and oxidative damage to the cell [13, 14]. Therefore, our criteria of a hyperO2ΔR1 > 0.04 s-1, which showed a close correlation between histological cell death and failure of renormalization, may indicate irreversible neuronal death triggered by mitochondrial dysfunction in stroke.
T2WI, DWI, and perfusion imaging have been widely used in both preclinical and clinical trials to measure areas with reversible or irreversible damage [15, 16, 17]. However, the origins of the signals on T2WI (water content), DWI (diffusivity of water protons), and perfusion imaging (delivery of contrast material) do not provide information on cell viability, and accurate cutoff values to distinguish reversible and irreversible damage have not been identified for these MRI methods. Furthermore, perfusion imaging has a significant limitation in terms of reliability, as perfusion-related parameters vary according to the calculation algorithm and the location used to extract the arterial input function [18, 19].
In response to these limitations, we attempted to systemically validate the utility of hyperO2ΔR1 for identifying regions with irreversible damage. We found that 1) measurements at various time points showed the reliability of hyperO2ΔR1 for identifying irreversible damage, 2) voxel-wise analysis allowed the establishment of criteria to delineate the border between infarct and non-infarct areas, and 3) the areas with histological cell death and hyperO2ΔR1 > 0.04 s-1 were similar. These findings validate the efficacy and reliability of a hyperO2ΔR1 value > 0.04 s-1 to estimate areas with irreversible damage.
The results of the present study demonstrated that the ADC may not be an adequate marker to identify irreversible ischemic damage; moreover, because of the gradual progression from non-ischemia to infarct, it was difficult to assign a cutoff ADC value to delineate the infarct area. Despite these limitations, as the ADC is particularly useful for assessing ischemic damage; thus, we recommend the combined implementation of ADC and hyperO2ΔR1 (decreased ADC and hyperO2ΔR1 < 0.04 s-1) to provide more accurate and reliable information for identifying non-infarct ischemia. Using these criteria, improvement from non-infarct ischemia to non-ischemia can be identified, as was observed for the spontaneous improvement shown in this study.
Several potential concerns and challenges must be addressed to apply hyperO2ΔR1 in clinical settings. First, application of administration of 100% O2 for several minutes may raise a concern regarding oxygen toxicity. Actually, hyperbaric oxygen therapy in longer than 60 minutes has risks of harmful effects [20, 21]. However, the use of hyperO2ΔR1 only for several minutes is unlikely to produce any harmful effects. Nevertheless, to completely remove such a risk, a maximum measurement time to guarantee safety must be established. The inevitable extension of examination time because of our recommended acquisition of multiple parameters (hyperO2ΔR1 in conjunction with ADC or CBF) is another important point of concern, as it is critical to minimize the scanning time in patients with acute stroke. Thus, we propose the use of ultrafast scanning techniques. Using the Look-Locker technique, R1 maps of the entire human brain can be acquired within 2 minutes, with a spatial resolution of 1.1 × 1.1 × 4 mm3 [22]. Another faster MRI sequence is the ultrafast low-angle RARE sequence, which has a temporal resolution that allows the measurement of R1 in th-e mouse brain within 6 seconds [23]. Moreover, as the R1 relaxation time can be influenced by the magnetic field strength, the hyperO2ΔR1 > 0.04 s-1 determined in our study on a 9.4T MRI system cannot be generally applied to 1.5T or 3T MRI systems. For example, the cutoff of 0.05 s-1 used in the previous study was determined under a 4.7T MRI system; thus, further refinement is needed for application in the clinical setting.
Several methods are used for the histological characterization of acute stroke, including 2,3,5-Triphenyltetrazolium chloride (TTC), H&E, Nissl, and caspase staining. TTC is one of the most widely used methods. This study used methods requiring glass slides, including H&E and Nissl staining, because they allow easier and more orthogonal photographic images that are required for co-registration with MRI images. Moreover, H&E and Nissl staining provide a wider time window for histological analysis compared to TTC [10].
In summary, the results of our study showed that MRI-driven hyperO2ΔR1 above 0.04 s-1 delineated areas with histological cell death 2.5 hours after stroke onset. In conjunction with other traditional MRI parameters, the use of hyperO2ΔR1 to identify areas with irreversible ischemic damage could help in determining adequate treatment and monitoring treatment effects in acute stroke.
Supplement
The Supplement is available with this article at https://doi.org/10.3348/kjr.2021.0477.
Click here to view.(192K, pdf)
Conflicts of Interest:The authors have no potential conflicts of interest to disclose.
Author Contributions:
Conceptualization: all authors.
Data curation: all authors.
Formal analysis: all authors.
Funding acquisition: all authors.
Investigation: all authors.
Methodology: all authors.
Project administration: all authors.
Resources: all authors.
Software: all authors.
Supervision: all authors.
Validation: all authors.
Visualization: all authors.
Writing—original draft: all authors.
Writing—review & editing: all authors.
Funding Statement:This work was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2017R1A2B3007567) and through the Korea Health Industry Development Institute (KHIDI) funded by the Ministry of Health & Welfare, Republic of Korea (HI14C1090).
Availability of Data and Material
The datasets generated or analyzed during the study are available from the corresponding author on reasonable request.
References
-
Li F, Liu KF, Silva MD, Meng X, Gerriets T, Helmer KG, et al. Acute postischemic renormalization of the apparent diffusion coefficient of water is not associated with reversal of astrocytic swelling and neuronal shrinkage in rats. AJNR Am J Neuroradiol 2002;23:180–188.
-
-
O’Connor JP, Boult JK, Jamin Y, Babur M, Finegan KG, Williams KJ, et al. Oxygen-enhanced mri accurately identifies, quantifies, and maps tumor hypoxia in preclinical cancer models. Cancer Res 2016;76:787–795.
-
-
Butcher KS, Parsons MW, Davis S, Donnan G. PWI/DWI mismatch: better definition required. Stroke 2003;34:e215–e216.
-
MeSH Terms
Figures
1 / 4
Tables
1 / 2
Funding Information
-
National Research Foundation of Korea
2017R1A2B3007567
-
Korea Health Industry Development Institute
HI14C1090
