Magnetic Particle Imaging for Real-Time Perfusion Imaging in Acute

Department of Radiology and Radiological Science, Johns Hopkins University School of Medicine, Baltimore, Maryland 21205, United States. □ Departmen...
0 downloads 0 Views 2MB Size
Download PDF
Magnetic Particle Imaging for Real-Time Perfusion Imaging in Acute Stroke Peter Ludewig,*,†,○ Nadine Gdaniec,‡,⊥,○ Jan Sedlacik,§ Nils D. Forkert,# Patryk Szwargulski,‡,⊥ Matthias Graeser,‡,⊥ Gerhard Adam,∥ Michael G. Kaul,∥ Kannan M. Krishnan,¶,△ R. Matthew Ferguson,¶ Amit P. Khandhar,¶ Piotr Walczak,▼,□ Jens Fiehler,§ Götz Thomalla,† Christian Gerloff,† Tobias Knopp,‡,⊥,○ and Tim Magnus†,○ †

Department of Neurology, ‡Section for Biomedical Imaging, §Department for Neuroradiological Diagnosis and Intervention, and ∥ Department of Diagnostic and Interventional Radiology and Nuclear Medicine, University Medical Center Hamburg-Eppendorf, 20246 Hamburg, Germany ⊥ Institute for Biomedical Imaging, Hamburg University of Technology, 21071 Hamburg, Germany # Department of Radiology and Hotchkiss Brain Institute, University of Calgary, Calgary, AB T2N 1N4, Canada ¶ LodeSpin Laboratories LLC, Seattle, Washington 98103, United States △ Materials Science and Engineering Department, University of Washington, Seattle, Washington 98195, United States ▼ Department of Radiology and Radiological Science, Johns Hopkins University School of Medicine, Baltimore, Maryland 21205, United States □ Department of Neurology and Neurosurgery, University of Warmia and Mazury, Olsztyn, Poland S Supporting Information *

ABSTRACT: The fast and accurate assessment of cerebral perfusion is fundamental for the diagnosis and successful treatment of stroke patients. Magnetic particle imaging (MPI) is a new radiation-free tomographic imaging method with a superior temporal resolution, compared to other conventional imaging methods. In addition, MPI scanners can be built as prehospital mobile devices, which require less complex infrastructure than computed tomography (CT) and magnetic resonance imaging (MRI). With these advantages, MPI could accelerate the stroke diagnosis and treatment, thereby improving outcomes. Our objective was to investigate the capabilities of MPI to detect perfusion deficits in a murine model of ischemic stroke. Cerebral ischemia was induced by inserting of a microfilament in the internal carotid artery in C57BL/6 mice, thereby blocking the blood flow into the medial cerebral artery. After the injection of a contrast agent (superparamagnetic iron oxide nanoparticles) specifically tailored for MPI, cerebral perfusion and vascular anatomy were assessed by the MPI scanner within seconds. To validate and compare our MPI data, we performed perfusion imaging with a small animal MRI scanner. MPI detected the perfusion deficits in the ischemic brain, which were comparable to those with MRI but in real-time. For the first time, we showed that MPI could be used as a diagnostic tool for relevant diseases in vivo, such as an ischemic stroke. Due to its shorter image acquisition times and increased temporal resolution compared to that of MRI or CT, we expect that MPI offers the potential to improve stroke imaging and treatment. KEYWORDS: magnetic particle imaging, ischemic stroke, stroke imaging, nanoparticles, nanomedicine, cerebrovascular disease, animal models of human disease, basic science research

W

of the most important pathophysiologic parameters in stroke diagnosis. The assessment of the cerebral perfusion allows discrimination between the infarcted stroke core and the

ith a prevalence of 17,000,000 strokes per year, cerebral ischemia is one of the leading causes of death and sustained disability worldwide.1 The management of stroke patients has become increasingly complex, but time remains a critical factor for their successful treatment. This is underlined by the fact that 2 million neurons die every minute after the onset of a stroke.2 Cerebral perfusion is one © 2017 American Chemical Society

Received: August 15, 2017 Accepted: September 30, 2017 Published: October 4, 2017 10480

DOI: 10.1021/acsnano.7b05784 ACS Nano 2017, 11, 10480−10488

Article

www.acsnano.org

Cite This: ACS Nano 2017, 11, 10480-10488

Article

ACS Nano

RESULTS AND DISCUSSION MPI Measurement of Cerebral Perfusion in Healthy C57BL/6 Mice. In the first step, we analyzed the cerebral perfusion in healthy mice (Experimental workflow: Figure 1). After one single bolus of the MPI tracer LS-008 (LodeSpin Laboratories LLC, Seattle, Washington, USA), 3D real-time MPI data sets (n = 3) were obtained during the passage of the tracer through the brain (Figure 2a and Video V1). For MRI (n = 3), only 2D(+t) data sets could be acquired after a single bolus of gadolinium to achieve the highest possible temporal resolution of 177 ms (Figure 2a). The high temporal resolution in MRI came at the expense of a reduced spatial resolution and a small field of view (FoV), which covered only a few brain slices, whereas MPI showed the vascular anatomy of the neck and perfusion of the entire brain parenchyma (Figure 2a). As this was the first attempt to measure cerebral perfusion with MPI, we validated the data with dynamic susceptibility contrast (DSC) MRI perfusion imaging, which is an established technique. The concentration− time curves of MPI and MRI were nearly identical for corresponding locations and showed a similar progression (Figure 2b; for averaged data from all animals, see Figure S1). Using these data, parametric maps of relative cerebral blood volume (rCBV), flow (rCBF), relative time to peak (rTTP), and relative mean transit time (rMTT) were calculated (Figure 2c). To verify the results obtained by our MATLAB script, MPI data were processed with the perfusion analysis tool AnToNIa (see Figure S2).13 Additionally, the high resolution and, at the same time, the larger FoV of the MPI data set allowed tracking of the contrast agent bolus in the different vascular compartments from the heart, common carotid artery (CCA), and brain parenchyma to the jugular veins (Figure 2d). From the modulated concentration− time curves, a heart rate of 334.41 ± 12.22 bpm (n = 6) was extracted by Fourier analysis, which was similar to the heart rates measured during anesthesia (364.33 ± 21.20 bpm, measured with PhysioSuite, Kent Scientific, Torrington, CT, USA). MPI Detects Perfusion Deficits in Ischemic Stroke. Based on the results in healthy mice, we performed middle cerebral artery occlusion (MCAO) in C57BL/6 mice and analyzed the cerebral perfusion during stroke with MPI. The MPI signal decreased in the ischemic hemisphere and allowed precise detection of the ischemic stroke of a few cubic millimeters with a similar efficiency to the 7 T animal MRI but with a considerably higher temporal resolution (Figure 3b and Video V2). Comparison with the DSC-MRI perfusion revealed similar results. The CBF dropped about 89.15 ± 1.33% (MPI, n = 3) versus 91.52 ± 2.76% (MRI, n = 3) in the ischemic hemispheres. This was in agreement with the data of the laser Doppler during stroke induction, which showed a reduction of 87.32 ± 4.13% in the CBF. The decrease of the CBV in the ischemic hemispheres was 93.44 ± 5.01% (MPI) versus 90.14 ± 2.89% (MRI) compared to the contralateral hemispheres. Consistently, the TTP prolonged by 85.85 ± 22.67% in MPI versus 89.26 ± 25.19% in MRI, and the MTT was extended by 22.89 ± 13.52% in MPI versus 36.72 ± 12.43% in MRI (Figure 3c). Similar results were obtained with the perfusion software AnToNIa (see Figure S2). Additionally, we could obtain anatomical information on the vasculature with a single bolus of the MPI tracer. We could clearly see the occlusion of the common carotid artery (CCA; Figure 3b, red arrow and asterisk), as shown by the time of flight angiography obtained by MRI (Figure 3a). With the high temporal resolution of MPI, we could even differentiate between the arterial and venous vessels (Figure 3e).

salvageable ischemic tissue, the so-called penumbra, which is necessary to identify acute stroke patients that could benefit from reperfusion therapies. Perfusion can be measured noninvasively with tomographic imaging, which is an integral part of the daily clinical routine and stroke diagnosis. However, there are inherent layers of treatment delays, such as the increased time for neuroimaging, due to missing prehospital imaging solutions.3 Most of these techniques, like computed tomography (CT) or magnetic resonance imaging (MRI), were invented in the second half of the last century. Since then, these techniques have been improved, but substantially new imaging technologies have not been developed. One disadvantage of CT is the inevitable radiation exposure. The drawbacks of MRI are susceptibility artifacts (e.g., hemorrhage, calcification, metal, air, and bone),4 as well as relatively low temporal and spatial resolution of most standard MRI scanners for clinical routines, which range from 1 s to 1 h and from 0.5 to 5 mm. In some cases, the acquired data can be insufficient for precise quantification of the cerebral perfusion. This can result in suboptimal patient treatment5 and poor success rates of stroke therapy, with 50% of the treated patients remaining disabled for life.6 Therefore, innovative concepts to combat ischemic brain diseases are urgently needed. The introduction of nanotechnologies in medicine, especially nanoscalestructured materials like superparamagnetic iron oxide nanoparticles (SPIOs), holds great promise for fast and advanced diagnostics. Between 2001 and 2005, Gleich and Weizenecker developed a new imaging modality called magnetic particle imaging (MPI).7 MPI is a new radiation-free tomographic imaging method that provides fast, background-free, sensitive, directly quantifiable, four-dimensional information about the spatial distribution of superparamagnetic iron oxide particles at a high temporal resolution of more than 46 volumes per second, combined with a high spatial resolution of 1 to 3 mm.8 MPI is also the first biomedical imaging technique that truly depends on nanoscale material properties, in particular, the magnetic relaxation dynamics of SPIOs subject to alternating fields in a true biological optimized environment.9 Due to the high temporal resolution, the observation of perfusion in real-time becomes possible.10 Diseases for which the rapid assessment of the vasculature and perfusion are mandatory for the treatment could substantially benefit from MPI. Due to the lack of appropriate MPI scanners with sufficient sensitivity, at a high temporal resolution, these theoretical advantages have not yet been proven in vivo. In this study, we worked with a preclinical MPI scanner (developed by Philips Medizin System, and manufactured by Bruker, Karlsruhe-Ettlingen, Germany), which uses a field free point (FFP) for spatial encoding. The FFP is moved rapidly along a 3D Lissajous sampling trajectory by applying three orthogonal excitation fields with slightly different frequencies of around 25 kHz and a field strength of up to 14 mT. In contrast to the FFP and field free line (FFL) MPI scanners of recent in vivo studies,11 which all had a single excitation channel, the three excitation channels of our scanner allow for a much better temporal resolution of at least 46 volumes per second. This is essential for capturing the dynamic inflow of a contrast agent bolus and assessing the cerebral perfusion real-time. With this first commercial, preclinical MPI scanner for in vivo experiments, and tracers truly optimized for MPI physics,12 we can now provide the first MPI data in a murine stroke model. As the success of stroke treatment is time-dependent, we address whether MPI is sensitive enough to detect ischemic lesions and is suitable for the usage as a future clinical imaging modality. 10481

DOI: 10.1021/acsnano.7b05784 ACS Nano 2017, 11, 10480−10488

Article

ACS Nano

Figure 1. Experimental setup and workflow of stroke-MPI in mice. In the first step (a), anatomical data were acquired with a 7 T small animal MRI with 3D-T1- and T2-weighted sequences, and time of flight (TOF) angiography for vascular anatomy. Diffusion-weighted (DWI) sequences and calculated apparent diffusion coefficient (ADC) maps were used for stroke detection. Additionally, dynamic susceptibility contrast perfusion imaging was performed to validate the MPI perfusion data. In the next step, the mice were transferred into the MPI scanner (b, lower panel shows the interior structure of the scanner, Bruker Biospin). The head position was marked with Resovist filled markers (fiducials, c), which could be detected in MPI and MRI and were used for the registration of the MPI and MRI data sets (d).

Histological Assessment of Tracer Extravasation. Gadolinium chelates or SPIOs do not cross the healthy bloodbrain barrier. During the time course of a stroke, the blood-brain barrier breaks down multiple times. On the one hand, extravasation of the tracer into the brain parenchyma is not wanted due to potential side effects of the tracer deposition in the central nervous system. On the other hand, tracer extravasation can also potentially contaminate the perfusion signal, resulting in an underestimation of the cerebral blood volume.14 After 3 h of reperfusion, we did not detect any deposition of LS-008 in the brain parenchyma of healthy and stroke mice (Figure 4). Discussion. For the first time, we showed that MPI might be a useful imaging method for the assessment of cerebral perfusion in an acute stroke model and potentially a new clinical imaging modality. Since the invention and the first phantom experiments in 2005,7 MPI has quickly evolved to a versatile tomographic imaging technique. More sophisticated MPI scanners are introduced every year,15−18 and with these scanners and new reconstruction methods, promising results have been achieved.19 Recently, the superior temporal resolution of MPI could visualize the heartbeat in real-time,20 and the high sensitivity for SPIO detection of MPI has been used to monitor cell migration in vivo.21 However, the explicit strength of MPI is ultrafast, four-dimensional, vascular imaging; therefore, we decided to use ischemic stroke to demonstrate the capabilities of MPI. We were able to detect small ischemic strokes of a few cubic millimeters by MPI, with comparable results to a small animal MRI. This was possible, even though our MPI scanner has a large bore diameter and receive coils, which makes the detection of small objects

harder than in the 7 T animal MRI. Although this is a drawback for small animal experiments, an upscaled version of our scanner with the same coil geometry could be used for the imaging of human-sized objects with a similar spatial resolution. However, maintaining the temporal resolution in a high-resolution imaging protocol with a large FoV will be challenging in human scanners. This problem can be circumvented by applying multipatch techniques, which move multiple small imaging FoVs through a larger target volume.22 An MPI coil setup capable of generating the magnetic fields necessary for human imaging was presented recently,23 showing the feasibility of the clinical translation of our study. Our results give a first hint of the versatile capabilities of MPI and its possible advantages over MRI and CT. Within seconds and with a single MPI tracer bolus, we could acquire important information on perfusion and stroke, vascular anatomy, and vessel occlusion of the common carotid artery, as well as differentiate the arterial and venous vessel and estimate the heart rate. In contrast, several different, time-consuming sequences are necessary for an MRI to obtain the same data, for example, DSC-MRI for the assessment of the perfusion and TOF or contrast-enhanced angiography for the imaging of the vasculature. A 64-channel CT scanner requires two shots of contrast agent for angiography and perfusion, with a dose length product of 4300 mGy*cm, which is four times higher than the annual radiation dose, compared to the zero radiation of an MPI scanner. The additional acquisition time for CT perfusion and CT angiography can be up to 15 min.24 Additionally, the accuracy and reliability of CT and MRI perfusion have recently been critically discussed.25 Defining the viable tissue through 10482

DOI: 10.1021/acsnano.7b05784 ACS Nano 2017, 11, 10480−10488

Article

ACS Nano

Figure 2. Assessment of cerebral perfusion with MPI. MPI measures cerebral perfusion in a comparable way to MRI. After the injection of 20 μL of LS-008 (MPI) or 10 μL if gadolinium (MRI), MPI and MRI scans were acquired dynamically, resulting in 3D(+t)-MPI and 2D(+t)-MRI perfusion data sets. (a) Contrast agent bolus passing through the brain in slices of the automatically fused 3D MPI/MRI data at several time points ((a) upper row, sagittal sections; middle row, coronal sections; lower row, transversal sections; lower panel, coronal 2D-MRI perfusion data set; see Video V1). The signal information from these data sets was converted into a concentration−time curve for each voxel of the imaged volume ((b) MPI, red crosses; MRI, black squares). Curve progressions were similar in both techniques, showing that MPI achieves similar results as MRI, but with a higher temporal resolution. Perfusion parameter maps were calculated from the concentration−time curves with our MATLAB software ((c) rCBF, relative cerebral blood flow; rCBV, relative cerebral blood volume; rTTP, relative time to peak; rMTT, relative mean transit time; upper row, transversal sections through the parameter maps calculated from MPI; middle row, coronal sections through the parameter maps calculated from MPI; lower row, MRI coronal sections through the parameters maps calculated from MRI). Due to the high temporal resolution and 3D data sets of MPI, the SPIO bolus could be tracked and differentiated in different vascular compartments (d), which was not possible based on the MRI data ((a) colored arrows indicate four regions of different temporal signal evolution; red, heart; orange, common carotid artery; black, brain parenchyma; blue, external jugular vein). Heartbeats could be calculated from the signal by Fourier analysis ((d) red arrow).

branches of the medial cerebral artery.28,29 Nevertheless, the sensitivity of MPI needs to be further evaluated in different stroke entities, such as lacunar or multiple small thromboembolic strokes. There are further critical questions that need addressing in the future. One disadvantage of MPI is the missing morphological information, which, therefore, limits an exact definition of an infarct core. Particularly, in patients with unknown stroke onset, a combination of CT or MRI with MPI will be necessary to identify treatable patients.30 However, the first MPI/MRI hybrid systems have been developed to solve this problem.31,32 These hybrid machines can also bypass the time-consuming coregistration step of MRI and MPI data. Aside from standard stroke imaging, using MPI as a point-of-care device for prehospital or bedside imaging could be an optimal field of application. Another strength of MPI is its efficient structure, which requires simple infrastructure for MPI scanner operations. The technology behind MPI allows the construction of portable MPI scanners. The first prototypes of such small, single-sided mobile MPI scanners have already been developed.17 Nowadays, patients on stroke units undergo neurological examinations periodically only every few hours, resulting in a medical care gap between these exams. Portable MPI scanners could be deployed on stroke units or intensive care units to monitor the cerebral

perfusion imaging is necessary to select optimal candidates for reperfusion therapy and has been identified as a priority by the American Heart and Stroke Association.26 However, the low temporal resolution, >1 s, in standard CT and MRI perfusion imaging can lead to variations of the penumbra of up to 100 mL depending on the quantification methods.27 This variation might deprive patients of an optimal stroke treatment.25,27 Besides the short image acquisition, the high temporal resolution of 46 frames per second in MPI (or even more than 1000 frames per second in 2D(+t) data sets) might help to identify the penumbra more precisely and to improve the prediction which patients will benefit from stroke treatment. The stroke model used in this study mimics a malignant middle cerebral artery infarction with a massive and constant drop of cerebral perfusion throughout the whole hemisphere and no definable penumbra by imaging. Further studies in other stroke models are necessary to estimate the capabilities of MPI to characterize the penumbra. In terms of translating our data to humans, we believe that a spatial resolution of 3 × 3 × 1.5 mm3 of our preclinical MPI scanner, which can also be achieved by human-sized MPI scanners,23 is sufficient to detect very small infarct volumes in patients (e.g., 0.08 mL in our MCAO model versus 54 mL in typical supratentorial infarcts).2 Additionally, the murine vessels occluded in the MCAO model range from 0.15 to 0.2 mm, which is still smaller than M3 or M4 10483

DOI: 10.1021/acsnano.7b05784 ACS Nano 2017, 11, 10480−10488

Article

ACS Nano

Figure 3. MPI stroke imaging. MPI detects reduced cerebral perfusion comparable to MRI after induction of stroke. The stroke area was assessed with different MRI sequences ((a) from left to right: diffusion-weighted imaging, which shows the ischemic stroke in the left hemisphere; T2-weighted sequences; TOF angiography showed disruption of the CCA/ICA, DSC-MRI, and ASL perfusion, which show reduced perfusion in the stroke hemisphere). After injection of 20 μL of LS-008 (MPI) or 10 μL of gadolinium (MRI), MPI/MRI scans were acquired dynamically, resulting in 3D(+t)-MPI and 2D(+t)-MRI perfusion data sets. (b) Contrast agent bolus passing through the brain in slices of the automatically fused 3D MPI/MRI data at several time points (see Video V2). The ischemic hemisphere could be easily detected in MPI ((b) red hash mark, ischemic hemisphere; upper row, sagittal sections through the right, nonoccluded CCA/ICA, red arrow; second row, sagittal sections through the left, occluded CCA/ICA, red asterisk; middle row, coronal sections; lower row, transversal sections). MPI was able to differentiate the anterior and posterior cerebral circulation via the arteria basilaris, which was not occluded in the MCAO model and showed unimpaired perfusion (b, blue arrow). MRI and MPI signals were plotted over time for certain selected regions of interests ((c) filled black circles, MRI signal ischemic hemisphere; filled black squares, MRI signal healthy hemisphere; red dotted line, MPI signal ischemic hemisphere; red crosses, MPI signal healthy hemisphere). The concentration−time curves of the MPI and MRI showed similar progression and reduced wash-out of the contrast agents into the ischemic hemisphere. Calculated perfusion parameter maps (d) of the MPI and MRI curves showed a similar reduction of the rCBF and rCBV or a delay in rTTP and rMTT. Overlaying the MPI with the TTP parameter map enabled the differentiation of arterial and venous vessels (e).

However, this might change over time due to increased blood-brain barrier breakdown over the following days. Other studies did not observe any extravasation of SPIOs if the hydrodynamic size was greater than 50 nm.34,35 Resovist, which is a clinically approved SPIO contrast agent, has already been used in humans with a similar safety profile, as compared to MRI contrast agents, and can even be applied in patients with renal failure. In contrast to gadolinium-based contrast agents, no brain parenchyma deposition has been reported for Resovist so far.36 To circumvent side effects, SPIOs can be encapsulated in red blood cells, minimizing the extent of injected iron oxide and increasing the halftime of the circulating SPIOs.37 In summary, we showed that MPI could be a useful tool for the rapid assessment of cerebral vasculature and perfusion in stroke, with comparable results to established imaging techniques. MPI provides great innovation potential to influence stroke imaging and treatment in the future.

perfusion of critically ill patients continuously. The worsening of patients attributed to vessel occlusion, bleeding, or vasospasm after a subarachnoid hemorrhage could be detected much faster with corresponding treatment. In this setting, the anatomical information is not as relevant as in the acute setting. For example, real-time perfusion monitoring of stroke patients could be done without any anatomical sequences, since the healthy hemisphere is used as a reference. To translate the monitoring capabilities of MPI to human applications, determining the safety limits of the magnetic fields for MPI scanners will be a vital step to avoid nerve excitations. First studies about the magnetostimulation safety limits for whole-body, human-sized MPI scanners have been performed, but further investigations into smaller brain scanners will be necessary.33 Another safety issue is the question of SPIO pharmacodynamics and nanotoxicology in stroke. We did not observe any extravasation of the 65 nm sized LS-008 in the first hours after stroke. 10484

DOI: 10.1021/acsnano.7b05784 ACS Nano 2017, 11, 10480−10488

Article

ACS Nano

Figure 4. No SPIO extravasation within the first hours. The deposition of SPIOs after induction was histologically assessed after the induction of an ischemic stroke. LS-008 circulated 3 h after reperfusion. Every 100 μm, brain sections were stained with an iron stain kit (HT20-1KT, SigmaAldrich, MO, USA). No SPIOs were detected in the healthy cortex (a, scale bar 50 μm), the striatum (c, scale bar 50 μm), the ischemic cortex (b, scale bar 50 μm), the ischemic striatum (d, scale bar 50 μm), or the liver (e, scale bar 50 μm). Spleen tissue (f, scale bar 50 μm) served as a positive control and showed iron deposition (blue). was inserted in the left internal carotid artery (ICA) to occlude the CCA, ICA, and the origin of the left middle cerebral artery (MCA). The operation time per animal did not exceed 15 min. The vital parameters of the mice were continuously monitored (blood oxygen saturation, heartbeats per minute, breathing frequency) with the PhysioSuite (Kent Scientific Corporation, USA). The occlusion of the MCA was confirmed with a laser Doppler monitor (moorVMS-LDF; Moor Instruments, UK). The anesthesia during MRI and MPI was maintained with 1.5% isoflurane in 100% O2. The vital parameters were monitored using an animal support unit (Minerve, Esternay, France). For the tracer injection, a catheter (inner tube diameter 0.28 mm, Portex, Smiths Medical International Ltd., USA) was placed into the tail vein of the mouse. After the image acquisition, the stroke filaments were removed to allow cerebral reperfusion for 3 h. MPI and MRI Measurements. The basic principle of MPI is based on the visualization of SPIOs capitalizing the nonlinear magnetization of the tracer and using external magnetic fields for excitation and spatial encoding.39 As a tracer-based imaging technique, MPI does not provide morphological data. Therefore, to obtain anatomical information about the murine head and neck, a 7-T small animal MRI was used (Figure 1a, Bruker ClinScan, Billerica, MA, USA) using a dedicated transmit/receive

METHODS Induction of Ischemic Stroke. All animal experiments were approved by local animal care committees (Behörde für Lebensmittelsicherheit and Veterinärwesen Hamburg, Nr. 42/14, 70/14, 16/41). We conducted the experiments following the “Guide for the Care and Use of Laboratory Animals” published by the U.S. National Institutes of Health (NIH Publication No. 83-123, revised 1996) and performed all procedures in accordance with the ARRIVE guidelines (http://www.nc3rs.org/ARRIVE). C57BL/6 mice (stroke animals, n = 3; healthy animals, n = 3) were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). The mice were kept under a 12 h light/dark cycle, at a constant temperature (22 ± 2 °C) and with food and water ad libitum. Acute ischemic stroke, mimicking a malignant middle cerebral artery infarction, was induced in 12-week-old male mice as described previously.38 The mice were anesthetized with 1.5% isoflurane (Abbott, Wiesbaden, Germany) in 100% O2 and an intraperitoneal injection of 0.05 mg/kg body weight buprenorphine in saline. After a midline skin incision in the neck, the proximal common carotid artery (CCA) and the external carotid artery (ECA) were ligated without the disruption of the venous vessels. A standardized siliconecoated 6.0 nylon filament (60-23-910; Doccol Corp., Redlands, USA) 10485

DOI: 10.1021/acsnano.7b05784 ACS Nano 2017, 11, 10480−10488

Article

ACS Nano (Bruker, Billerica, MA, USA) mouse body volume coil and a 3D gradient echo method with the following measurement parameters: TE = 0.84 ms, TR = 10 ms, flip angle = 10°, matrix 192 × 192 × 192 with elliptical k-space sampling, field of view = 32 × 32 × 32.64 mm3, 2 averages, readout bandwidth = 700 Hz/pixel and strong asymmetric echo readout. The scan time was 9 min and 40 s. MRI perfusion and angiographic measurements were performed using a dedicated receive only 2 × 2 channel mouse head coil array in combination with a rat body transmit volume coil. Arterial vessels were measured using the TOF magnetic resonance angiography with the following scanning parameters: TE = 3.06 ms, TR = 17 ms, flip angle = 30°, matrix = 256 × 256 × 24 with 16.7% slab oversampling and elliptical k-space sampling, field of view = 20 × 20 mm2, slice thickness 130 μm, 151 Hz/pixel readout bandwidth and moderate asymmetric echo readout, a parallel acceleration factor of 2 with 19 inline acquired reference lines and a scan time of 3 min:44s. To validate the results of the MPI perfusion measurement, we performed a dynamic susceptibility contrast (DSC)- MRI using echoplanar imaging (EPI) pulse sequences, which enables a relatively high temporal resolution of 177 ms in our animal 7 T MRI. After the injection of 10 μL gadolinium chelate (Gadovist Bayer Schering Pharma AG, Germany), images were acquired dynamically during the passage of the contrast agents through the brain. The tracer was injected manually into the tail vein within 1 s. Sequence parameters were as follows: TE = 8.9 ms, TR = 177 ms, flip angle = 30°, matrix = 64 × 48 due to partial Fourier factor = 6/8, the field of view = 20 × 20 mm2, nine slices, 400 μm slice thickness, 2004 Hz/pixel readout bandwidth, parallel acceleration factor of 2 with 19 in-line acquired reference lines, and 150 repetitions resulting in a scan time of 26.55 s. The head position of the animal was marked with fiducials filled with Resovist (Ferucarbotran, Bayer Schering Pharma AG, Germany) for image co-registration of the MRI and MPI data during postprocessing.40 The fiducials were further used within the experiment planning phase for placing the mouse head into the center of the MPI device (Figure 1). The placement was performed by mechanically adjusting the mouse position under the guidance of a custom online image reconstruction and visualization framework developed in.41 After animal positioning, a 4D MPI acquisition was started, and a 20 μL SPIO bolus (LS-008, LodeSpin Laboratories, USA) with a concentration of 46 mmol(Fe)/L was manually injected via the mouse’s tail vein within 1 s. MPI scans (4D data with 21.5 ms temporal resolution, a selection field gradient strength of 2.5 T/m in one direction and 0.75 T/m in the remaining orthogonal directions, and a drive field (DF) amplitude of 14 mT) were acquired dynamically with 100% duty cycle while administering the contrast agent bolus. The scan parameters result in a DF-field of view of 37.33 × 37.33 × 18.66 mm3. For reconstruction of the distribution of magnetic nanoparticles, a system matrix is acquired on a 50 × 50 × 25 mm3 field of view using a Delta sample filled with 4 μL LS-008. Each reconstructed 3D volume, in turn, consists of 25 × 25 × 25 voxels with a voxel resolution of 2 × 2 × 1 mm3. The physical resolution of the MPI data is decoupled from the voxel resolution and depends on the signal-to-noise ratio (SNR) of the measurement signal.42 The shown MPI data had a physical resolution of about 3 × 3 × 1.5 mm3. For more information about MPI, see supplemental methods. MPI Image Reconstruction and Postprocessing. Image reconstruction was performed offline using an iterative reconstruction framework introduced by Knopp et al. (Figure 1; for a detailed description of the image reconstruction, see supplemental methods).43 Image postprocessing was performed using an in-house-developed software developed in MATLAB (The Mathworks, USA). Signal information was converted into a contrast medium concentration−time curve for each voxel in both data sets. From these data, parametric maps of relative cerebral blood volume (rCBV), flow (rCBF), relative time to peak (rTTP), and relative mean transit time (rMTT) were calculated (see supplemental methods for image postprocessing). As no valid arterial input function could be defined intracerebrally in the MRI data due to the small animal size and low spatial resolution, the perfusion maps were calculated in a relative manner. More precisely, rCBV was calculated by the integral of the contrast bolus time curve, whereas rMTT was defined as the first moment of the curve and rCBF was

calculated according to the central volume theorem (rCBF = rCBV/rMTT).44 For the presentation of the rTTP and rMTT parameter maps, the brightness of the color of a voxel of the parameter maps was modulated by the signal of the voxel in the maximum intensity projection (MIP) over time. We used the software FSL viewer for combining the parameter maps with the MIP in time. The color of a voxel with high signal in the temporal MIP is depicted with the maximum brightness of the respective color. On the other hand, the brightness of the color is reduced for voxels where less or no contrast agent went, for example, low or no signal in the temporal MIP. The results obtained with the MATLAB script were confirmed with a sophisticated perfusion software, ANTONIA, which uses a deconvolution-based perfusion analysis method (see Figure S2).13 Histological Analysis. Histological analyses were performed as described.38 Briefly, mice were deeply anesthetized with isoflurane followed by perfusion through the left ventricle using 10 mL of phosphate-buffered saline and 50 mL of cold 4% paraformaldehyde (PFA). The brains were postfixed in 4% PFA overnight at 4 °C and cryoprotected in 30% sucrose (w/v) in PBS for 3 days. After snap freezing the brains embedded in Tissue-Tek OCT compound (Sakura Finetek Europe B.V, Flemingweg, Netherlands) in precooled isopentane, brains were cut into 6 μm coronal sections using a Leica CM3050 cryotome. Sections were stained with an iron stain kit according to the manufacturer’s instructions (iron stain kit, HT20-1KT, Sigma-Aldrich, MO, USA) and examined under a Leica DM5000B microscope.

ASSOCIATED CONTENT S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsnano.7b05784. Averaged concentration−time curves from all animals; calculation of the perfusion parameters with the software tool AnToNIa; schematic representation of the MCAO model; equations used for MPI image reconstruction and the calculation of the perfusion parameter maps (PDF) Video V1: Real-time MPI of the cerebral perfusion in healthy mice, showing the tracer bolus through the murine brain with the full temporal resolution of 47 fps (MPG) Video V2: Real-time MPI of the cerebral perfusion in acute stroke, showing the tracer bolus through the murine brain after induction of MCAO with the full temporal resolution of 47 fps (MPG)

AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. Phone: +49 40741018800. ORCID

Peter Ludewig: 0000-0001-9025-6402 Author Contributions ○

P.L., N.G., T.K., and T.M. contributed equally to this work.

Notes

The authors declare no competing financial interest.

ACKNOWLEDGMENTS This work was supported by the “Forschungszentrums Medizintechnik Hamburg” (FMTHH, granted to P.L. and N.G.), the “Werner Otto Stiftung” (P.L.), the “Schilling Professur” (T.M.), and the ERANET Grant “NeuroSurv” (T.M.). Work at the University of Washington and LodeSpin Laboratories was supported by NIH 1R41EB013520-01 and NIH 2R42EB01352002A1. K.M.K. acknowledges the Alexander von Humboldt Forschungspreis (2016). 10486

DOI: 10.1021/acsnano.7b05784 ACS Nano 2017, 11, 10480−10488

Article

ACS Nano

(15) Mason, E. E.; Cooley, C. Z.; Cauley, S. F.; Griswold, M. A.; Conolly, S. M.; Wald, L. L. Design analysis of an MPI human functional brain scanner. Int. J. Magn Part Imaging 2017, 3, 1703008. (16) Vogel, P.; Lother, S.; Ruckert, M. A.; Kullmann, W. H.; Jakob, P. M.; Fidler, F.; Behr, V. C. MRI Meets MPI: a bimodal MPI-MRI tomograph. IEEE Trans Med. Imaging 2014, 33 (10), 1954−9. (17) Grafe, K.; von Gladiss, A.; Bringout, G.; Ahlborg, M.; Buzug, T. M. 2D Images Recorded With a Single-Sided Magnetic Particle Imaging Scanner. IEEE Trans Med. Imaging 2016, 35 (4), 1056−65. (18) Goodwill, P. W.; Konkle, J. J.; Zheng, B.; Saritas, E. U.; Conolly, S. M. Projection x-space magnetic particle imaging. IEEE Trans Med. Imaging 2012, 31 (5), 1076−85. (19) Panagiotopoulos, N.; Duschka, R. L.; Ahlborg, M.; Bringout, G.; Debbeler, C.; Graeser, M.; Kaethner, C.; Ludtke-Buzug, K.; Medimagh, H.; Stelzner, J.; Buzug, T. M.; Barkhausen, J.; Vogt, F. M.; Haegele, J. Magnetic particle imaging: current developments and future directions. Int. J. Nanomed. 2015, 10, 3097−114. (20) Weizenecker, J.; Gleich, B.; Rahmer, J.; Dahnke, H.; Borgert, J. Three-dimensional real-time in vivo magnetic particle imaging. Phys. Med. Biol. 2009, 54 (5), L1−L10. (21) Zheng, B.; von See, M. P.; Yu, E.; Gunel, B.; Lu, K.; Vazin, T.; Schaffer, D. V.; Goodwill, P. W.; Conolly, S. M. Quantitative Magnetic Particle Imaging Monitors the Transplantation, Biodistribution, and Clearance of Stem Cells In Vivo. Theranostics 2016, 6 (3), 291−301. (22) Knopp, T.; Them, K.; Kaul, M.; Gdaniec, N. Joint reconstruction of non-overlapping magnetic particle imaging focus-field data. Phys. Med. Biol. 2015, 60 (8), L15−21. (23) Rahmer, J.; Stehning, C.; Gleich, B. Spatially selective remote magnetic actuation of identical helical micromachines. Science Robotics 2017, 2, eaal2845. (24) Srinivasan, A.; Goyal, M.; Lum, C.; Nguyen, T.; Miller, W. Processing and interpretation times of CT angiogram and CT perfusion in stroke. Can. J. Neurol. Sci. 2005, 32 (4), 483−6. (25) Huang, X.; Kalladka, D.; Cheripelli, B. K.; Moreton, F. C.; Muir, K. W. The Impact of CT Perfusion Threshold on Predicted Viable and Nonviable Tissue Volumes in Acute Ischemic Stroke. J. Neuroimaging 2017, DOI: 10.1111/jon.12442. (26) Latchaw, R. E.; Yonas, H.; Hunter, G. J.; Yuh, W. T.; Ueda, T.; Sorensen, A. G.; Sunshine, J. L.; Biller, J.; Wechsler, L.; Higashida, R.; Hademenos, G. Council on Cardiovascular Radiology of the American Heart, A., Guidelines and recommendations for perfusion imaging in cerebral ischemia: A scientific statement for healthcare professionals by the writing group on perfusion imaging, from the Council on Cardiovascular Radiology of the American Heart Association. Stroke 2003, 34 (4), 1084−104. (27) Forkert, N. D.; Kaesemann, P.; Treszl, A.; Siemonsen, S.; Cheng, B.; Handels, H.; Fiehler, J.; Thomalla, G. Comparison of 10 TTP and Tmax estimation techniques for MR perfusion-diffusion mismatch quantification in acute stroke. AJNR Am. J. Neuroradiol 2013, 34 (9), 1697−703. (28) Cilliers, K.; Page, B. J. Anatomy of the Middle Cerebral Artery: Cortical Branches, Branching Pattern and Anomalies. Turk Neurosurg 2017, 27 (5), 671−681. (29) Schambach, S. J.; Bag, S.; Steil, V.; Isaza, C.; Schilling, L.; Groden, C.; Brockmann, M. A. Ultrafast high-resolution in vivo volume-CTA of mice cerebral vessels. Stroke 2009, 40 (4), 1444−50. (30) Thomalla, G.; Fiebach, J. B.; Ostergaard, L.; Pedraza, S.; Thijs, V.; Nighoghossian, N.; Roy, P.; Muir, K. W.; Ebinger, M.; Cheng, B.; Galinovic, I.; Cho, T. H.; Puig, J.; Boutitie, F.; Simonsen, C. Z.; Endres, M.; Fiehler, J.; Gerloff, C. A multicenter, randomized, double-blind, placebo-controlled trial to test efficacy and safety of magnetic resonance imaging-based thrombolysis in wake-up stroke (WAKE-UP). Int. J. Stroke 2014, 9 (6), 829−836. (31) Vogel, P.; Lother, S.; Rückert, M. A.; Kullmann, W. H.; Jakob, P. M.; Fidler, F.; Behr, V. C. MRI Meets MPI: A Bimodal MPI-MRI Tomograph. IEEE Transactions on Medical Imaging 2014, 33 (10), 1954−1959. (32) Franke, J.; Heinen, U.; Lehr, H.; Weber, A.; Jaspard, F.; Ruhm, W.; Heidenreich, M.; Schulz, V. System Characterization of a Highly

ABBREVIATIONS MPI, magnetic particle imaging; CT, computed tomography; MRI, magnetic resonance imaging; SPIO, superparamagnetic iron oxide; CCA, common carotid artery; ICA, internal carotid artery; MCA, medial cerebral artery; ECA, external carotid artery; CBF, cerebral blood flow; CBV, cerebral blood volume; TTP, time to peak; MTT, mean transit time; TOF, time of flight; ASL, arterial spin labeling; DF-FoV, drive field-field of view; DSC, dynamic susceptibility contrast; FFP, field free point; FFL, field free line REFERENCES (1) Feigin, V. L.; Forouzanfar, M. H.; Krishnamurthi, R.; Mensah, G. A.; Connor, M.; Bennett, D. A.; Moran, A. E.; Sacco, R. L.; Anderson, L.; Truelsen, T.; O’Donnell, M.; Venketasubramanian, N.; Barker-Collo, S.; Lawes, C. M.; Wang, W.; Shinohara, Y.; Witt, E.; Ezzati, M.; Naghavi, M.; Murray, C. Global and regional burden of stroke during 1990−2010: findings from the Global Burden of Disease Study 2010. Lancet 2014, 383 (9913), 245−255. (2) Saver, J. L. Time is brain–quantified. Stroke 2006, 37 (1), 263−6. (3) Lin, M. P.; Sanossian, N.; Liebeskind, D. S. Imaging of prehospital stroke therapeutics. Expert Rev. Cardiovasc. Ther. 2015, 13 (9), 1001− 15. (4) Essig, M.; Nguyen, T. B.; Shiroishi, M. S.; Saake, M.; Provenzale, J. M.; Enterline, D. S.; Anzalone, N.; Dorfler, A.; Rovira, A.; Wintermark, M.; Law, M. Perfusion MRI: the five most frequently asked clinical questions. AJR, Am. J. Roentgenol. 2013, 201 (3), W495−510. (5) Molad, J. A.; Findler, M.; Auriel, E. Computed Tomography Perfusion-Based Decision Making for Acute Ischemic Stroke-Missing the Mismatch. J. Stroke Cerebrovasc Dis 2017, 26 (5), e78−e79. (6) Hacke, W.; Kaste, M.; Bluhmki, E.; Brozman, M.; Davalos, A.; Guidetti, D.; Larrue, V.; Lees, K. R.; Medeghri, Z.; Machnig, T.; Schneider, D.; von Kummer, R.; Wahlgren, N.; Toni, D.; et al. Thrombolysis with alteplase 3 to 4.5 h after acute ischemic stroke. N. Engl. J. Med. 2008, 359 (13), 1317−1329. (7) Gleich, B.; Weizenecker, J. Tomographic imaging using the nonlinear response of magnetic particles. Nature 2005, 435 (7046), 1214−7. (8) Weizenecker, J.; Gleich, B.; Rahmer, J.; Dahnke, H.; Borgert, J. Three-dimensional real-time in vivo magnetic particle imaging. Phys. Med. Biol. 2009, 54 (5), L1. (9) Ferguson, R. M.; Khandhar, A. P.; Kemp, S. J.; Arami, H.; Saritas, E. U.; Croft, L. R.; Konkle, J.; Goodwill, P. W.; Halkola, A.; Rahmer, J.; Borgert, J.; Conolly, S. M.; Krishnan, K. M. Magnetic particle imaging with tailored iron oxide nanoparticle tracers. IEEE Trans Med. Imaging 2015, 34 (5), 1077−84. (10) Kaul, M. G.; Mummert, T.; Jung, C.; Salamon, J.; Khandhar, A. P.; Ferguson, R. M.; Kemp, S. J.; Ittrich, H.; Krishnan, K. M.; Adam, G.; Knopp, T. In vitro and in vivo comparison of a tailored magnetic particle imaging blood pool tracer with Resovist. Phys. Med. Biol. 2017, 62 (9), 3454−3469. (11) Zheng, B.; Vazin, T.; Goodwill, P. W.; Conway, A.; Verma, A.; Ulku Saritas, E.; Schaffer, D.; Conolly, S. M. Magnetic Particle Imaging tracks the long-term fate of in vivo neural cell implants with high image contrast. Sci. Rep. 2015, 5, 14055. (12) Khandhar, A. P.; Keselman, P.; Kemp, S. J.; Ferguson, R. M.; Goodwill, P. W.; Conolly, S. M.; Krishnan, K. M. Evaluation of PEGcoated iron oxide nanoparticles as blood pool tracers for preclinical magnetic particle imaging. Nanoscale 2017, 9 (3), 1299−1306. (13) Forkert, N. D.; Cheng, B.; Kemmling, A.; Thomalla, G.; Fiehler, J. ANTONIA perfusion and stroke. A software tool for the multi-purpose analysis of MR perfusion-weighted datasets and quantitative ischemic stroke assessment. Methods Inf. Med. 2014, 53 (6), 469−81. (14) Jahng, G. H.; Li, K. L.; Ostergaard, L.; Calamante, F. Perfusion magnetic resonance imaging: a comprehensive update on principles and techniques. Korean J. Radiol 2014, 15 (5), 554−77. 10487

DOI: 10.1021/acsnano.7b05784 ACS Nano 2017, 11, 10480−10488

Article

ACS Nano Integrated Preclinical Hybrid MPI-MRI Scanner. IEEE Transactions on Medical Imaging 2016, 35 (9), 1993−2004. (33) Saritas, E. U.; Goodwill, P. W.; Zhang, G. Z.; Conolly, S. M. Magnetostimulation limits in magnetic particle imaging. IEEE Trans Med. Imaging 2013, 32 (9), 1600−10. (34) Shilo, M.; Sharon, A.; Baranes, K.; Motiei, M.; Lellouche, J. P.; Popovtzer, R. The effect of nanoparticle size on the probability to cross the blood-brain barrier: an in-vitro endothelial cell model. J. Nanobiotechnol. 2015, 13, 19. (35) Barua, S.; Mitragotri, S. Challenges associated with Penetration of Nanoparticles across Cell and Tissue Barriers: A Review of Current Status and Future Prospects. Nano Today 2014, 9 (2), 223−243. (36) Kanda, T.; Ishii, K.; Kawaguchi, H.; Kitajima, K.; Takenaka, D. High signal intensity in the dentate nucleus and globus pallidus on unenhanced T1-weighted MR images: relationship with increasing cumulative dose of a gadolinium-based contrast material. Radiology 2014, 270 (3), 834−41. (37) Markov, D. E.; Boeve, H.; Gleich, B.; Borgert, J.; Antonelli, A.; Sfara, C.; Magnani, M. Human erythrocytes as nanoparticle carriers for magnetic particle imaging. Phys. Med. Biol. 2010, 55 (21), 6461−73. (38) Ludewig, P.; Sedlacik, J.; Gelderblom, M.; Bernreuther, C.; Korkusuz, Y.; Wagener, C.; Gerloff, C.; Fiehler, J.; Magnus, T.; Horst, A. K. Carcinoembryonic antigen-related cell adhesion molecule 1 inhibits MMP-9-mediated blood-brain-barrier breakdown in a mouse model for ischemic stroke. Circ. Res. 2013, 113 (8), 1013−22. (39) Knopp, T.; Gdaniec, N.; Moddel, M. Magnetic particle imaging: from proof of principle to preclinical applications. Phys. Med. Biol. 2017, 62, R124. (40) Werner, F.; Jung, C.; Hofmann, M.; Werner, R.; Salamon, J.; Säring, D.; Kaul, M. G.; Them, K.; Weber, O. M.; Mummert, T.; Adam, G.; Ittrich, H.; Knopp, T. Geometry planning and image registration in magnetic particle imaging using bimodal fiducial markers. Med. Phys. 2016, 43 (6), 2884−2893. (41) Knopp, T.; Hofmann, M. Online reconstruction of 3D magnetic particle imaging data. Phys. Med. Biol. 2016, 61 (11), N257. (42) Graeser, M.; Knopp, T.; Szwargulski, P.; Friedrich, T.; von Gladiss, A.; Kaul, M.; Krishnan, K. M.; Ittrich, H.; Adam, G.; Buzug, T. M. Towards Picogram Detection of Superparamagnetic Iron-Oxide Particles Using a Gradiometric Receive Coil. Sci. Rep. 2017, 7 (1), 6872. (43) Knopp, T.; Rahmer, J.; Sattel, T. F.; Biederer, S.; Weizenecker, J.; Gleich, B.; Borgert, J.; Buzug, T. M. Weighted iterative reconstruction for magnetic particle imaging. Phys. Med. Biol. 2010, 55 (6), 1577. (44) Kennan, R. P.; Jäger, H. R. T2- and T2*-W DCE-MRI: Blood Perfusion and Volume Estimation Using Bolus Tracking. Quantitative MRI of the Brain; John Wiley & Sons, Ltd.: New York, 2004; pp 365− 412.

10488

DOI: 10.1021/acsnano.7b05784 ACS Nano 2017, 11, 10480−10488