Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Aug 1.
Published in final edited form as: Magn Reson Med. 2019 Apr 25;82(2):842–853. doi: 10.1002/mrm.27759

DEVELOPMENT OF A FAST SCAN EPR IMAGING SYSTEM FOR HIGHLY ACCELERATED FREE RADICAL IMAGING

Alexandre Samouilov +,*, Rizwan Ahmad +, James Boslett +, Xiaoping Liu +, Sergey Petryakov ++, Jay L Zweier +,*
PMCID: PMC6510602  NIHMSID: NIHMS1018424  PMID: 31020713

Abstract

Purpose:

In continuous wave (CW) electron paramagnetic resonance imaging (EPRI), the acquisition of high quality images was previously limited by the requisite long acquisition times of each image projection that was typically greater than 1 second. To accelerate the process of image acquisition facilitating greater numbers of projections and higher image resolution, instrumentation was developed to greatly accelerate the magnetic field scan that is used to obtain each EPR image projection.

Methods:

A low inductance solenoidal coil for field scanning was utilized along with a spherical solenoid air core magnet and scans were driven by triangular symmetric waves, allowing forward and reverse spectrum acquisition as rapid as 3.8 ms. The uniform distribution of projections was used to optimize the contribution of projections for 3D image reconstruction.

Results:

Utilizing this fast scan EPR system, high quality EPR images of phantoms and perfused rat hearts were performed using trityl or nanoparticulate LiNcBuO probes with fast scan EPR imaging at L-band, achieving spatial resolutions of up to 250 micrometers in one minute.

Conclusion:

Fast scan EPR imaging can greatly facilitate the efficient and precise mapping of the spatial distribution of free radical and other paramagnetic probes in living systems.

Keywords: electron paramagnetic resonance imaging, continuous wave, fast-scan acquisition, instrument development

INTRODUCTION

The high sensitivity of EPR spectral parameters to many physical and chemical properties such as pH, oxygenation, redox state and thiol content makes EPR-based imaging techniques a valuable tool for in vivo and ex vivo studies (1–5). These techniques have been widely applied to study normal physiology and disease in small animal models such as mice and rats (6). EPR spectroscopy and imaging can provide critical mechanistic information on tissue redox-status and oxygenation and detect free radicals produced in pathological conditions. EPR-related imaging techniques capable of detecting paramagnetic substances in vivo have been developed and continue to be improved. These include methods that enable direct detection of paramagnetic compounds such as continuous wave (CW) EPR (1, 7), time domain pulsed EPR (3), and rapid scan EPR (4, 8) as well as MRI-based indirect techniques such as T1 contrast imaging (9, 10) visualizing a radical-induced enhancement and EPR/ proton MRI hybrid methods that enable visualization of radical-induced Overhauser enhancement (11–13).

In vivo or ex vivo CW electron paramagnetic resonance imaging (EPRI) has been established as a powerful technique for determining the spatial distribution of free radicals and other paramagnetic species in living organs and tissues. EPRI can provide a broad range of important information; however, its application to preclinical models of disease as well as its potential for clinical use is limited by the long acquisition time required to generate high quality images, due to the fact that in conventional EPR systems each projection acquisition magnetic field scan requires durations of at least 1 to 2 seconds. Therefore, it is desirable to develop CW EPR instrumentation optimized for fast scan capability enabling 100-fold acceleration of projection acquisition over commercial systems or 5 times faster than the prior reported developmental systems. The instrumentation developed is capable of acquiring full 3D EPR images with high spatial and temporal resolutions in seconds, and is thus suitable for in vivo applications where physiological conditions may change rapidly. This instrument, along with ongoing developments of new functional paramagnetic probes, will facilitate valuable insights into physiology and pathology including: radical metabolism, redox state, oxygen and thiol content, pH as well as other functional information. With the developments described in this report, EPRI gains the ability to measure and image changes of these functional parameters at temporal resolution of less than a minute.

METHODS

Hardware setup

Overall System Description

A block-diagram of the fast scan EPR system configuration is shown on Figure 1. In order to achieve the fast field sweep necessary for accelerated fast scan EPRI (FS-EPRI) acquisition, the system utilizes a separate water-cooled primary magnetic field electromagnet and fast low inductance sweep coil placed at the center of the primary magnet. Thus, the system utilizes two main electro-magnet sets for the main field and the sweep field as well as additional 3D gradients.

Fig. 1.

Fig. 1.

Open in a new tab

Block diagram of the fast scan EPRI system configuration. The magnet is powered by a pair of serially connected Techron 7780 power supplies capable of providing current up to ~35 A with existing coils. The gradients are powered by three modified Copley 261P switching amplifiers. Sweep coil is powered by a Copley 232 power supply. EPR Microwave bridge is a homemade L-band bridge capable of 1.06 – 1.52 GHz. Lock-in amplifier - Stanford Research Model SR865A DSP. Modulation amplifier - Hafler Model Transnova P3000. DACS, two Data Translation communication boards - Model NI PCIe 6343.

The microwave part of the system consists of the L-band resonator, connected to the homemade microwave bridge; signal processing was done by lock-in amplifier, which also drives the modulation amplifier and modulation coils. The EPR signal acquisition is controlled by a dedicated workstation equipped with two 16-bit A/D boards with maximum sampling rate of 500 kilosamples per second with four independent channels (model PCIe 6343, National Instruments Inc., TX USA). The workstation’s input is interfaced with a lock-in amplifier (SR865A, Stanford Research Systems, CA USA) and the outputs are interfaced with three channel gradient amplifier (261P, Copley Control Inc., OH USA), main field regulated power supply (model 7780, Techron a division of Crown International, Inc, IN USA) and sweep coil amplifier (232 Copley Control Inc., OH USA ). The L-band microwave bridge was similar to that previously described (14, 15), but additionally utilized a laboratory-built cavity stabilized microwave oscillator with maximum power output of 320 mW.

Magnet and Magnet Power Supply

The magnetic field B0 is generated by a water-cooled spherical solenoid resistive magnet, with an axial opening of 29 cm in diameter, capable of generating a magnetic field up to 0.12 T (Fig. 2). The design was originally reported by Hoult et al. (16) and later adapted for EPR experiments (17, 18). The magnet is constructed on two aluminum cast hemispherical formers. The magnet provides homogeneity of better than 3 ppm over a spherical volume of 10 cm. Inductance of serially connected hemispherical coils is 328 mH with active resistance ~ 8.5Ω. Its magnetic field is horizontally oriented along its bore. The magnet was powered by two serially-connected Techron power supply amplifiers Model 7780 capable of providing current up to ~35 A with the existing coils. The magnetic field is current-controlled and its stability relies on the high-precision adjustable reference voltage. The magnet is water-cooled and thermo-stabilized using a heat exchanger.

Fig. 2.

Fig. 2.

Open in a new tab

Schematic representation of the magnetic system setup. A. Main field spherical magnet 85 cm o.d with two 29 cm diameter axial access openings; B. Solenoidal sweep coil; C. X, Y, Z gradient coil set; D. Fiberglass resin tube 14 cm i.d., 110 cm long; E. Water manifold and electrical terminals housing; F. Resonator.

Gradient Coils and Gradient Power Supplies

The custom gradient coil set used for EPRI was constructed by Resonance Research Inc., MA USA. It is a water-cooled set of 3D field cylindrical fast gradients that is interfaced to the gradient power supply amplifiers. The X and Y gradient is a multilayer 2×16 coil set formed from 1.5 mm thick copper plate formed around 1050 mm long fiberglass tube and inner diameter of 138 mm. Z gradient is 66 turn symmetrically distributed solenoid. The gradient coils have low inductance and low resistance (X coil: 1320 µH/143 mΩ; Y coil: 1550 µH/166 mΩ; Z coil: 760 µH/266 mΩ). The gradient coil set is capable of generating a sustained gradient of 250 mT/m along the X, Y, and Z axes and 500 mT/m at 25 % duty cycle with linearity of about ±1 % in a spherical volume of 80 mm and ± 5 % in a spherical volume of 140 mm. Rise time 10–90 % is 1 ms. The gradients are powered by three modified Copley 261P switching amplifiers capable of 62 A per channel, controlled by the computer console (Fig. 1, 2). Cooling of the gradient coils (maximum resistive power dissipation = 2.7 kW) is provided by a water circuit heat exchanger (Varian Medical Systems Inc., CA USA) that also shares the water circuit with the main magnet and sweep coil. Gradient coils are equipped with 8 thermocouple temperature detectors connected to an external scanner interfaced with a safety interlock system that also monitors cooling water pressure.

Sweep Coil and Sweep Power Supply

Sweep of the main magnetic field was provided by a water-cooled cylindrical coil constructed by Resonance Research Inc. It is a water-cooled solenoid resistive magnet formed on a plastic support tube with 230 mm inner diameter and placed concentrically over the gradient setup (Fig. 2). The sweep coil has 396 µH inductance and 84 mΩ resistance. Inhomogeneity of the field over a 50 mm diameter spherical volume is less than 0.01 %. The sweep coil is equipped with 4 thermocouple temperature detectors. The coil is capable of providing field sweep of ± 3 mT with rise time of 1 ms. The coil is driven by a Copley 232 power supply in current amplification mode. Symmetric triangle signal input allowed us to use forward and reverse magnetic field sweep without time loss for back scan. The time delay in the acquisition, caused by a digital circuit in the lock-in amplifier, and delay due to time constant during phase sensitive detection versus sweep signal were digitally compensated for to match the forward and reverse scan signals. Superposition of two EPR spectra of the phantom filled with LiNcBuO suspension (described later) acquired during forward and reverse scan (Fig. 3 panel A) illustrates good overall correlation between signals (Fig. 3, panel C) with a correlation coefficient of 0.9944. The slight mismatch in signal shapes was observed due to the time constant of the lock-in amplifier (Fig. 3, panel B).

Fig. 3.

Fig. 3.

Open in a new tab

A. EPR signals of the conical phantom filled with LiNcBuO suspension acquired with the single forward scan of magnetic field (orange) superimposed on the EPR signal acquired with single reverse scan (blue). Scan width 0.8 mT, scan time 3.8 ms, time constant 10 µs. Modulation frequency 64 kHz, modulation amplitude 80% of probe linewidth of 15.6 µT; B. Difference between above signals (green); C. Correlation plot of above forward and reverse spectra.

EPR Resonator

The design of a 6-gap loop-gap L-band resonator (1.24 GHz, ID = 18 mm, length = 20 mm), as used here, was similar to the one described in (19) and positioned orthogonally to the magnet bore (Fig. 4–1). This is a 6-gap design with the bore oriented vertically. This design was previously demonstrated to provide good spatial separation of E and B1 fields (19). The dimensions and orientation of the resonator were chosen to accommodate isolated perfused hearts of rats and phantoms of similar size as needed to perform the EPR imaging evaluations and measurements. The resonator body is 3D-printed (PowerSpec 3DPro2 printer, Micro Electronics, Inc., OH USA) from Inland ABS plastic filament manufactured by International Products Sourcing Group (IPSG), OH USA. All elements are enclosed in a mechanically rigid plastic housing (Fig. 4, part 7) and attached to the resonator enclosure for easy positioning in the gradient isocenter of the magnet. Similar to our single loop multi-gap resonator designs, the resonator is magnetically coupled through the fixed coupling loop minimizing the number of moving parts (Fig. 4 part 4). Coupling adjustment was achieved by changing the capacitance of the trimming capacitor (Fig. 4 part 5) placed in the coupling loop circuit through a quarter wavelength transformer. The resonator was enclosed in a copper foil cylindrical shield (Fig. 4 part 2) made of 0.1 mm copper foil. An additional secondary shield (Fig. 4 part 3) was constructed around the entire resonator assembly. The quality factor Q of the shielded empty resonator measured using a RF network analyzer was 2000, and it dropped to ~ 400 with a resonant frequency of 1.26 GHz when loaded with a phantom simulating an isolated heart load. A pair of Helmholtz coils (Fig. 4 part 6), used for field modulation (15 turns of 0.8 mm transformer wire) with diameter 70 mm, is connected in series and provided modulation field up to 4 G. Modulation coils are powered through a FG-P3000D amplifier (Haffler, a division of Rockford Corporation, BC Canada) in non-resonance mode. The setup has provisions for the perfusate supply lines and a tray for collecting the heart effluent.

Fig. 4.

Fig. 4.

Open in a new tab

A 1.2 GHz 6-gap single loop resonator assembly. The coupling capacitor is soldered directly to the feeding cable, through a quarter wavelength transformer. 1. Resonator with inner diameter of 18 mm and 20 mm length; 2. Shield; 3. Secondary shield; 4. Coupling loop circuit; 5. Coupling capacitor; 6. Modulation coil; 7. Housing

Conical Phantom

In order to determine the resolution of the imager with different parameter settings, a phantom consisting of a double cone insert of “hour glass” shape was designed (Fig. 5A). This phantom insert was printed using a 3D printer with ABS plastic filament. In addition, to eliminate porosity caused by extrusion printing and make the phantom water impermeable, the surface of the printed piece was coated with a solution of polystyrene in cyclohexane. The insert was placed into a plastic tube and the surrounding volume (0.6 mL) was filled with 1 mM triaryl methyl (TAM) paramagnetic probe in normal saline or a lithium octa-n-butoxy-substituted naphthalocyanine (LiNcBuO) suspension. A 1% w/w suspension of nanoparticules of the radical probe LiNc-BuO in saline was prepared according to (20) with the LiNcBuO synthesized as previously reported (21). The suspension was stabilized with eight-fold excess of methyl-β-cyclodextrin (average MW 1310) from Sigma-Aldrich Inc., MO USA.

Fig. 5.

Fig. 5.

Open in a new tab

An “hour glass” phantom. Diameter 10 mm, overall height 38 mm. Phantom was inserted in 10 mm plastic tube filled with 1 mM paramagnetic probe TAM. Excluded volume was imaged using different acquisition parameters using resonator for perfused heart experiments. B. Images of the phantom. Please note that length of the image is presented by active volume of the resonator which is about 20 mm in Z direction. a. Central slice in Z-Y plane; b. Central slice in X-Y plane; c. Central slice in X-Z plane.

i. Slices of the 3D image obtained using one 38 ms forward and reverse scan per projection. Number of projections 729. Time constant 100 µs.

ii. Slices of the 3D image obtained using ten 3.8 ms forward and reverse scans per projection. Number of projections 729. Time constant 10 µsec.

iii. Slices of the 3D image obtained using 3.8 ms forward and reverse scan per projection. Number of projections 7569. Time constant 10 µsec.

In all three cases total image acquisition time was approximately 1 min.

Data acquisition parameters: system frequency 1.18 GHz, field-of-view 30×30×30 mm, number of projections 4096, 3D Golden Means sampling (22), sweep width 1.6×10−3 T, max spatial gradient 2.6 × 10−3 T/m (only applied for imaging data), field modulation amplitude 1.6 × 10−5 T at 64 kHz and time per sweep 0.0038 s.

Heart preparation

Sprague-Dawley rats (250–300g) were used in all isolated heart preparations. Briefly, rats were euthanized with pentobarbital (30–50 mg/kg) and heparinized both via IP injection. Once deeply anesthetized, thoracotomy was performed with excision of the heart. The aorta was cannulated and perfused in a retrograde manner with Krebs bicarbonate perfusate consisting of 117 mM NaCl, 24.6 mM NaHCO3, 5.9 mM KCl, l.2 mM MgC12, 2.0 mM CaCl2, 16.7 mM glucose bubbled with 95% O2, and 5% CO2 under constant pressure of 80 mmHg. The temperature of the heart was kept at 37 ± 1 ºC by maintaining perfusate temperature with a heat exchanger and thermostat.

The heart was suspended inside the resonator cavity using a specially designed holder. The inner diameter of the resonator cavity is about 18 mm. The holder was positioned on the top for EPR measurements with the isolated heart positioned at the center of the resonator. After 15 min of control perfusion, the perfusate flow was stopped and 1 mL of 1 % suspension of LiNcBuO nanoparticles in phosphate-buffered saline was perfused through a side arm close to the aortic cannula.

The acquisition software was then initiated to perform automated projection acquisition for EPR imaging. EPR spectra (zero gradient projections) were acquired before each complete image cycle and were used for deconvolution of projections. The projection data were collected nonstop. In addition, a delay typically of 0.4 ms was introduced between projections. Some of the parameters values used for data acquisition include: system frequency 1.18 GHz, field-of-view 30×30×30 mm, number of projections 4096, 3D Golden Means sampling (22), sweep width 1.6 × 10−3 T, max spatial gradient 2.6 × 10−3 T/m (only applied for imaging data), field modulation amplitude 1.6 × 10−5 T at 64 kHz and time per sweep 0.0038 s. The projection data were deconvoluted and then reconstructed into a 128×128×128 3D image using filtered back projection.

RESULTS

The central field coils and magnetic field sweep coils were initially calibrated using the nitroxide probe tetramethyl-4-oxo-piperidine-N-oxyl (TEMPONE) in such a way that measured g factor and hyperfine splittings of the spectra coincided with known spectral parameters of the probe. A rectangular hollow bar made from an optical plastic cuvette filled with 0.5 mM TAM was used to calibrate the 3D set of gradient coils in a way similar to that previously described (23).

EPR images of the double cone “hour glass” phantom were obtained to optimize image acquisition parameters of the system. The shape of the phantom facilitated assessment of the resolution of the obtained images. The meeting point between cones is about 0.1 mm in diameter. The phantom was positioned vertically along the Y axis of the gradient system. Resolution was assessed by measurement of the distance between cones on the images. In order to optimize the image acquisition routine, a number of 3D images were acquired using different number of projections, number of scans per projection, and different scan time per projection. Time constant of the lock-in amplifier was maintained ~ 1/400 of the projection sweep time, which did not result in any distortion of the EPR spectra. The projection data collection times were fixed to one minute.

Slices from the 3D EPR images (Fig. 5B) demonstrated that 729 projection image acquisition using relatively slow scans of 38 ms resulted in images with resolution limited to ~ 4.5 mm. Details were better resolved when each projection was acquired with 10 times more (twenty) very short 3.8 ms scans per projection with an observed resolution of ~ 1.1 mm. Increasing the number of projections to 7569 with corresponding decrease in number of scans per projection resulted in only minor improvement in image quality with image resolution of ~ 0.6 mm.

From these images, as shown in Fig. 5 all totaling one minute image acquisition time, we see that better resolution was possible using higher number of short scans rather than smaller number of longer scans. When total acquisition time is the same, the signal-to-noise ratio of the spectrum should be similar in both cases. A longer acquisition time with a longer lock-in amplifier time constant and proportionally longer accumulation of the signal, or a shorter acquisition time and more accumulation with a short lock-in amplifier time constant should result in the same SNR. However, this assumption is applicable to white noise. In our experimental environment, the electromagnetic interference is caused by intermittent man-made sources and is not white noise. The observed 0.6 mm resolution is close to the theoretical limit calculated as half of the ratio LW/Gradient (24), which can be achieved using a probe with LW of 2.3 × 10−5 T and gradients 2.67 × 10−2 T/m with scan width of 8 × 10−4 T and FOV of 30×30×30 mm.

Using the above optimized parameters for projection acquisition, we attempted to estimate the minimum time required for image acquisition with desired image quality. For this, images were acquired with different time and number of projections. As seen from Fig. 6, significant progression in image fidelity was observed with increase in time of acquisition from 2 to 14 s. With further increase in time there was only marginal further improvement.

Fig. 6.

Fig. 6.

Open in a new tab

Progression in image quality with acquisition time. Images of the phantom were acquired using one 0.8 mT forward and reverse scan per projection. Scan time 3.8 ms, time constant 10 µsec, 0.4 ms delay between projections. Modulation frequency 64 kHz, modulation amplitude 80% of probe linewidth. Time of image acquisition/number of projections: A - 2 s/256; B - 3.8 s/484; C - 7.2 s/900; D - 14 s/1764; E - 30 s/3721; F- 60 s/7569.

In order to assess the ability of the new fast scan EPRI instrument to perform rapid EPR imaging of a stable paramagnetic probe in a well-defined biological model, we performed experiments by administering stable nanoparticulate suspensions containing 1% w/w of the EPR probe LiNcBuO to isolated rat hearts. After 15 min of normal perfusion, this probe formulation was infused and the heart then subjected to global no-flow ischemia. After infusion, the acquisition of images were initiated for collection of 3D images consisted of 4096 projections. Strong EPR signals were observed enabling the use of a field gradient of 5 × 10−2 T/m, twice that of the phantom studies above. Zero gradient EPR spectra were recorded before each image acquisition to monitor the sharpening of the signal due to the decrease in oxygen concentration, which also were used for the spectral deconvolution in the image reconstruction. Each image acquisition took 64 seconds. A series of images were acquired and an example of compiled 3D images is shown in Fig. 7.

Fig. 7.

Fig. 7.

Open in a new tab

3D images (30×30×30 mm) of ischemic rat heart infused with LiNcBuO suspension: A, B (top) longitudinal slices and C. transverse slice (bottom) showing the internal structure of the heart; D. surface rendering of the full view of the heart with cut-away. Data acquisition parameters: projections 4096 (4 × 3.8 ms scans per projection), total acquisition time 64 s, field sweep 1.6 mT, probe LW 31.3 µT, MA 16 µT at 64 kHz.

To further estimate the spatial resolution achievable with this LiNcBuO probe preparation and the conditions used for the fast scan EPRI, images of the hour glass phantom were obtained with acquisition parameters identical to those used for heart experiments. As seen from Fig. 8, details of less than 250 micrometers can be resolved on the images. This resolution corresponds to the voxel size on the reconstructed 128×128×128 images with 30 mm field of view.

Fig. 8.

Fig. 8.

Open in a new tab

3D images (30×30×30 mm) of hour glass phantom filled with 1 % LiNcBuO suspension: A, B. longitudinal slices and C. transverse slice showing the meeting point of the phantom tips; Data acquisition parameters are similar to parameters used for heart image in fig 7.

DISCUSSION

CW EPR spectroscopy remains the most commonly used and versatile modality for biological EPR measurements. Conventional CW EPR imaging as developed over the last two decades which utilizes typical commercial EPR magnets and field controllers is an inherently “slow” technique with projections typically requiring seconds for each and minutes to even hours for high resolution spatial imaging and spectral-spatial imaging. With traditional iron core magnets with their high inductance resistive coils driven by slow power supplies, scans faster than of mT/s are not possible. The first EPR imagers were based on conventional EPR spectrometers and thus inherited this limitation of slow scan acquisition (1, 25–27). Slow inductive 3D gradient coils required for spatial resolution, also added time to image acquisition (15, 28). However, dynamic processes in vivo, where conditions can change rapidly, such as tissue redox state and oxygenation, or distribution of the probes or presence of organ movements, such as contractile motion of the heart or respiration-based movements, limited the applicability of EPR imaging with low acquisition efficiency. EPRI on conventional instruments that utilize Hall probe magnetic field regulation, inherently requires projection acquisition times of >2 seconds leading to relatively long image acquisition times of minutes. The need for accelerated image acquisition speed for biomedical applications has been clear. Therefore, major efforts were devoted to accelerate the EPR image acquisition. On par with key advances in pulsed EPR imaging (3, 10, 29), Overhauser effect-based MRI (13, 30–32), and development of rapid scan approaches (4, 8, 33, 34), traditional CW with field modulation and phase sensitive detection has undergone major improvements. Pulsed EPR image modality allows functional (oxygen map) images of legs of living mice with matrices 31×31 pixel and submillimeter resolution to be obtained in 264 seconds (3), or for 3D imaging in vivo, the scan time is under 10 minutes and resolution 1.4–2 mm (29). Overhauser effect-based MRI allows increased resolution due to intrinsic advantages of acquiring proton MRI signal. Resolution of ~0.6 mm was achieved in matrix 64×64 and field of view 40×40 mm in 4.2 min (13) or 1.6 mm in 8.4 s using functional nitroxides as spin probes (32). Rapid scan EPR has been shown to provide substantially higher signal to noise than CW for the same amount of data acquisition time (35). In functional imaging modality, a 2D spectral-spatial image of a phantom consisting of two 6 mm diameter tubes filled with 0.5 mM solution of amino-substituted triarylmethyl radical demonstrated resolution on the mm scale when acquired with 18 projections each 43 s long (33).

Important steps were made in optimization of image acquisition software and image reconstruction routines (36–43). Implementation of these approaches has been shown to allow reduction in the number of projections resulting in reduction of total image acquisition time achieving submillimeter resolution in images of isolated rat hearts infused with LiNcBuO suspension and acquired in 264 s(41, 42, 44).

With regard to hardware development, construction of fast 3D gradients even in combination with high inductance main magnet allowed significant reduction of CW EPRI image acquisition time. The gradients were continuously rapidly changed (spun) during projection collection while the main magnetic field is stepped (45). This approach allowed fast acquisitions while avoiding fast changing of the main magnetic field. Over 10-fold accelerated acquisition of image projections were achieved. 2D images with over 200 projections acquired in less than 3s at 300 MHz system (17), and at L-band a 3D image with more than 480 projections was acquired within 58 s which was 4–7-fold faster than the conventional stepped gradient approach (36). This approach, combined with a uniform sampling distribution, was successful in providing acceleration in EPR image data collection without the need for fast sweep of the main magnetic field (42), and also avoided the need to accelerate acquired projection data transfer.

The approach of adding low inductance fast sweeping coils to the main field coil and collecting “rapid” projection data meets several challenges. EPRI on conventional instruments that utilize Hall probe closed-loop magnetic field regulation, inherently requires projection acquisition times of > 2 seconds leading to relatively long image acquisition times of minutes. Implementation of sweep coils dictates direct current control and stabilization of the main field, which is especially challenging with iron core magnets. Even with an air-core magnet setup, strong mutual coupling between the sweep coil and main magnet complicates the task. Sweeping of the magnetic field can cause the main magnet field to shift/oscillate, and the presence of the highly inductive main magnetic field coil may alter the sweep coil performance affecting the linearity of field scanning. Therefore, a feedback system or proper calibration of the driver current/voltage output depending on the scan speed may be required. One can use elaborate software corrections for waveforms and stabilization of the main field similar to that which we previously used for variable field PEDRI (46) or one can rely on powerful high voltage power supplies.

Even with an air core magnet, the implementation of an additional large fast sweep coil powered by a separate power supply is also associated with certain technical difficulties, including the need for the use of sophisticated field driving current amplifiers, stabilization of the current in the main field circuitry overcoming distortions caused by coupling with the sweep field, and the need for proper timing of data acquisition with the magnetic field. Recent developments in fast field scanning CW-EPR reduced acquisition time of single EPR projection to tens of milliseconds (18, 23, 44, 47–49).

The short times required for fast scan sweep of the magnetic field also determines the rate of data transfer which is needed to acquire the given projection. Thus, fast scan projection acquisition requires major acceleration of the acquisition data transfer and also of the data of the control waveforms from the control computer to acquisition electronics. These data transfer issues can also contribute to the slowdown of acquisition speed. The traditional fixed-rate oversampling acquisition scheme and acquisition software allowed a minimum projection acquisition time of 1.3 second (17, 36). Adaptive heterogeneous clocking (AHC) allows a projection acquisition time as short as 5 ms with 1024 data points (44) and as short as 3.8 ms with 512 data points in the present work.

In our EPR instrument, separate from the main field split spherical solenoid magnet, field-sweeping solenoidal coils produce linear field ramp up to 6 mT with ramp speed 6 T/s when powered by Copley 232 power amplifiers, which are capable of delivering 140 V and handling 100 A pulses of current. To avoid main field drift caused by the sweep field ramps, we utilize a symmetrical ramp with acquisition of forward and reverse scans. Oscillations in the field were suppressed by powering the main field with two serially-connected Techron 7780 power supplies with ~ 400 V peak to peak sum output. With these advances, the fast field-scanning EPRI system made it possible to acquire 3D spatial EPR images in a matter of several seconds. Achieved parameters allow imaging of TAM and LiNcBuO probes that have a single-line EPR absorption spectrum. However spatial imaging with field of view 30 mm and gradient strength 5 × 10−2 T/m is confined to the probes with spectral width up to ~4 mT, which allows whole spectrum imaging of 15N nitroxides but limits the utilization of 14N nitroxides to those with narrow linewidth of 0.08 mT or less.

CONCLUSIONS

We have developed an EPR imaging instrument that uses fast three dimensional EPR image acquisition to detect the distribution of paramagnetic materials within biological systems. This instrumentation enables in vivo 3D EPR imaging in biological objects within a time suitable for physiological investigation. Our approach has an advantage over past approaches in that we use a combination of a low inductance solenoid fast sweep coil with an air core solenoid resistive magnet. The resistive magnet system used can be set to any magnetic field up to 0.1 T, allowing choice of EPR frequencies up to 2.8 GHz. The design of the resonator setup holds the sample in place and allows perfusion of isolated organs during acquisition. Our studies show that 3D isolated rat heart images can be obtained in less than one minute with scans as fast as 3.8 ms. We were able to visualize the spatial distribution of a paramagnetic probe within a perfused heart with submillimeter resolution. Thus, the described fast scan EPR imaging instrument can greatly facilitate the efficient and precise mapping of the spatial distribution of paramagnetic molecules in living systems.

ACKNOWLEDGMENTS

Special thanks to Arkadiy Iosilevich, laboratory machinist for constructing the resonator assembly and phantoms used, and Boris Epel, Research Professor, Department of Radiation and Cellular Oncology at University of Chicago, for critical help in troubleshooting the acquisition software.

This work was supported by NIH grants EB014542, EB004900, HL135648 and HL131941.

References

  • 1.Berliner JL, Fujii H. Magnetic resonance imaging of biological specimens by electron paramagnetic resonance of nitroxide spin labels. Science 1985;227(4686):517–9. [DOI] [PubMed] [Google Scholar]
  • 2.He G, Samouilov A, Kuppusamy P, Zweier JL. In vivo imaging of free radicals: applications from mouse to man. Mol Cell Biochem 2002;234–235(1–2):359–67. [PubMed] [Google Scholar]
  • 3.Kishimoto S, Matsumoto KI, Saito K, Enomoto A, Matsumoto S, Mitchell JB, Devasahayam N, Krishna MC. Pulsed Electron Paramagnetic Resonance Imaging: Applications in the Studies of Tumor Physiology. Antioxid Redox Signal 2018;28(15):1378–93. 10.1089/ars.2017.7391. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Eaton SS, Shi Y, Woodcock L, Buchanan LA, McPeak J, Quine RW, Rinard GA, Epel B, Halpern HJ, Eaton GR. Rapid-scan EPR imaging. J Magn Reson 2017;280:140–8. 10.1016/j.jmr.2017.02.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Gorodetsky AA, Kirilyuk IA, Khramtsov VV, Komarov DA. Functional electron paramagnetic resonance imaging of ischemic rat heart: Monitoring of tissue oxygenation and pH. Magn Reson Med 2016;76(1):350–8. 10.1002/mrm.25867. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Subramanian S, Krishna MC. Dancing with the Electrons: Time-Domain and Cw in Vivo Epr Imaging. Magn Reson Insights 2008;2:43–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Amida T, Nakaoka R, Komarov DA, Yamamoto K, Inanami O, Matsumoto S, Hirata H. A 750-MHz Electronically Tunable Resonator Using Microstrip Line Couplers for Electron Paramagnetic Resonance Imaging of a Mouse Tumor-Bearing Leg. IEEE Trans Biomed Eng 2018;65(5):1124–32. 10.1109/TBME.2017.2743232. [DOI] [PubMed] [Google Scholar]
  • 8.Czechowski T, Chlewicki W, Baranowski M, Jurga K, Szczepanik P, Szulc P, Tadyszak K, Kedzia P, Szostak M, Malinowski P, Wosinski S, Prukala W, Jurga J. Two-dimensional EPR imaging with the rapid scan and rotated magnetic field gradient. J Magn Reson 2014;248:126–30. 10.1016/j.jmr.2014.09.022. [DOI] [PubMed] [Google Scholar]
  • 9.Winter JD, Akens MK, Cheng HL. Quantitative MRI assessment of VX2 tumour oxygenation changes in response to hyperoxia and hypercapnia. Phys Med Biol 2011;56(5):1225–42. 10.1088/0031-9155/56/5/001. [DOI] [PubMed] [Google Scholar]
  • 10.Matsumoto KI, Mitchell JB, Krishna MC. Comparative studies with EPR and MRI on the in vivo tissue redox status estimation using redox-sensitive nitroxyl probes: influence of the choice of the region of interest. Free Radic Res 2018;52(2):248–55. 10.1080/10715762.2018.1427235. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Lurie DJ. Commentary: electron spin resonance imaging studies of biological systems. Br J Radiol 1996;69(827):983–4. [DOI] [PubMed] [Google Scholar]
  • 12.Efimova OV, Sun Z, Petryakov S, Kesselring E, Caia GL, Johnson D, Zweier JL, Khramtsov VV, Samouilov A. Variable radio frequency proton-electron double-resonance imaging: application to pH mapping of aqueous samples. J Magn Reson 2011;209(2):227–32. 10.1016/j.jmr.2011.01.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Kishimoto S, Krishna MC, Khramtsov VV, Utsumi H, Lurie DJ. In Vivo Application of Proton-Electron Double-Resonance Imaging. Antioxid Redox Signal 2018;28(15):1345–64. 10.1089/ars.2017.7341. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Zweier JL, Kuppusamy P. Electron paramagnetic resonance measurements of free radicals in the intact beating heart: a technique for detection and characterization of free radicals in whole biological tissues. Proc Natl Acad Sci U S A 1988;85(15):5703–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Kuppusamy P, Chzhan M, Zweier JL. Development and optimization of three-dimensional spatial EPR imaging for biological organs and tissues. J Magn Reson B 1995;106(2):122–30. [DOI] [PubMed] [Google Scholar]
  • 16.Hoult DI, Goldstein S, Caponiti J. Electromagnet for Nuclear Magnetic-Resonance Imaging. Rev Sci Instrum 1981;52(9):1342–51. 10.1063/1.1136769. [DOI] [Google Scholar]
  • 17.Deng Y, He G, Petryakov S, Kuppusamy P, Zweier JL. Fast EPR imaging at 300 MHz using spinning magnetic field gradients. J Magn Reson 2004;168(2):220–7. 10.1016/j.jmr.2004.02.012. [DOI] [PubMed] [Google Scholar]
  • 18.Ahmad R, Samouilov A, Zweier JL. Accelerated dynamic EPR imaging using fast acquisition and compressive recovery. J Magn Reson 2016;273:105–12. 10.1016/j.jmr.2016.10.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Petryakov S, Samouilov A, Kesselring E, Kuppusamy P, Zweier JL, editors. L-band Resonator for Imaging of Whole Mouse 11th In Vivo EPR Spectroscopy and Imaging Conference EPR 2005; 2005; Columbus, Ohio. [Google Scholar]
  • 20.Kuppusamy P, Pandian RP, Parinandi LP, Zweier JL, inventorsNANOPARTICULATE PROBE FOR IN VIVO MONITORING OF TISSUE OXYGENATION patent 20140170073 2014.
  • 21.Pandian RP, Parinandi NL, Ilangovan G, Zweier JL, Kuppusamy P. Novel particulate spin probe for targeted determination of oxygen in cells and tissues. Free Radic Biol Med 2003;35(9):1138–48. [DOI] [PubMed] [Google Scholar]
  • 22.Chan RW, Ramsay EA, Cunningham CH, Plewes DB. Temporal Stability of Adaptive 3D Radial MRI Using Multidimensional Golden Means. Magnetic Resonance in Medicine 2009;61(2):354–63. 10.1002/mrm.21837. [DOI] [PubMed] [Google Scholar]
  • 23.Samouilov A, Caia GL, Kesselring E, Petryakov S, Wasowicz T, Zweier JL. Development of a hybrid EPR/NMR coimaging system. Magn Reson Med 2007;58(1):156–66. 10.1002/mrm.21205. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Hoch MJR, Ewert U. Resolution in EPR imaging. In: Eaton GR, Eaton SS, Ohno K, editors. EPR imaging and in vivo EPR: Boca Raton: CRC Press; 1991. p. 153–9. [Google Scholar]
  • 25.Fujii H, Berliner LJ. One- and two-dimensional EPR imaging studies on phantoms and plant specimens. Magn Reson Med 1985;2(3):275–82. [DOI] [PubMed] [Google Scholar]
  • 26.Eaton SS, Maltempo MM, Stemp EDA, Eaton GR. 3-dimensional Electron-Paramagnetic-Res imaging with one spectral and 2 spatial dimensions. Chemical Physics Letters 1987;142(6):567–9. [Google Scholar]
  • 27.Halpern HJ, Spencer DP, Polen JV, Bowman MK, Nelson AC, Dowey EM, Teicher BA. Imaging radio frequency electron spin resonance spectometer with high resolution and sensitivity for in vivo mesurements. Rev Sci Instrum 1989;60:1040:50. [Google Scholar]
  • 28.Kuppusamy P, Chzhan M, Vij K, Shteynbuk M, Lefer DJ, Giannella E, Zweier JL. Three-dimensional spectral-spatial EPR imaging of free radicals in the heart: a technique for imaging tissue metabolism and oxygenation. Proc Natl Acad Sci U S A 1994;91(8):3388–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Matsumoto KI, Kishimoto S, Devasahayam N, Chandramouli GVR, Ogawa Y, Matsumoto S, Krishna MC, Subramanian S. EPR-based oximetric imaging: a combination of single point-based spatial encoding and T1 weighting. Magn Reson Med 2018. 10.1002/mrm.27182. [DOI] [PMC free article] [PubMed]
  • 30.Lurie DJ, Li H, Petryakov S, Zweier JL. Development of a PEDRI free-radical imager using 0.38 T clinical MRI system. Magnetic Resonance in Medicine 2002;47(1):181–6. [DOI] [PubMed] [Google Scholar]
  • 31.Lurie DJ, Davies GR, Foster MA, Hutchison JM. Field-cycled PEDRI imaging of free radicals with detection at 450 mT. Magn Reson Imaging 2005;23(2):175–81. [DOI] [PubMed] [Google Scholar]
  • 32.Samouilov A, Efimova OV, Bobko AA, Sun Z, Petryakov S, Eubank TD, Trofimov DG, Kirilyuk IA, Grigor’ev IA, Takahashi W, Zweier JL, Khramtsov VV. In vivo proton-electron double-resonance imaging of extracellular tumor pH using an advanced nitroxide probe. Anal Chem 2014;86(2):1045–52. 10.1021/ac402230h. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Elajaili HB, Biller JR, Tseitlin M, Dhimitruka I, Khramtsov VV, Eaton SS, Eaton GR. Electron spin relaxation times and rapid scan EPR imaging of pH-sensitive amino-substituted trityl radicals. Magn Reson Chem 2015;53(4):280–4. 10.1002/mrc.4193. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Subramanian S, Koscielniak JW, Devasahayam N, Pursley RH, Pohida TJ, Krishna MC. A new strategy for fast radiofrequency CW EPR imaging: direct detection with rapid scan and rotating gradients. J Magn Reson 2007;186(2):212–9. 10.1016/j.jmr.2007.01.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Biller JR, Tseitlin M, Mitchell DG, Yu Z, Buchanan LA, Elajaili H, Rosen GM, Kao JP, Eaton SS, Eaton GR. Improved sensitivity for imaging spin trapped hydroxyl radical at 250 MHz. Chemphyschem 2015;16(3):528–31. 10.1002/cphc.201402835. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Deng Y, Petryakov S, He G, Kesselring E, Kuppusamy P, Zweier JL. Fast 3D spatial EPR imaging using spiral magnetic field gradient. J Magn Reson 2007;185(2):283–90. 10.1016/j.jmr.2007.01.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Ahmad R, Vikram DS, Clymer B, Potter LC, Deng Y, Srinivasan P, Zweier JL, Kuppusamy P. Uniform distribution of projection data for improved reconstruction quality of 4D EPR imaging. J Magn Reson 2007;187(2):277–87. 10.1016/j.jmr.2007.05.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Ahmad R, Deng Y, Vikram DS, Clymer B, Srinivasan P, Zweier JL, Kuppusamy P. Quasi Monte Carlo-based isotropic distribution of gradient directions for improved reconstruction quality of 3D EPR imaging. J Magn Reson 2007;184(2):236–45. 10.1016/j.jmr.2006.10.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Ahmad R, Clymer B, Deng Y, He G, Vikram D, Kuppusamy P, Zweier JL. Optimization of data acquisition for EPR imaging. J Magn Reson 2006;179(2):263–72. 10.1016/j.jmr.2005.12.013. [DOI] [PubMed] [Google Scholar]
  • 40.Kuppusamy Periannan, Zweier JL. A Forward-Substraction Procedure for Removing Hyperfine Artifacts in Electron Paramagnetic Resonance Imaging. Magnetic Resonance in Medicine 1996;35:316–22. [DOI] [PubMed] [Google Scholar]
  • 41.Johnson DH, Ahmad R, He G, Samouilov A, Zweier JL. Compressed sensing of spatial electron paramagnetic resonance imaging. Magn Reson Med 2014;72(3):893–901. 10.1002/mrm.24966. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Johnson DH, Ahmad R, Liu Y, Chen Z, Samouilov A, Zweier JL. Uniform spinning sampling gradient electron paramagnetic resonance imaging. Magn Reson Med 2014;71(2):893–900. 10.1002/mrm.24712. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Deng Y, Kuppusamy P, Zweier JL. Progressive EPR imaging with adaptive projection acquisition. J Magn Reson 2005;174(2):177–87. 10.1016/j.jmr.2005.01.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Chen Z, Reyes LA, Johnson DH, Velayutham M, Yang C, Samouilov A, Zweier JL. Fast gated EPR imaging of the beating heart: spatiotemporally resolved 3D imaging of free-radical distribution during the cardiac cycle. Magnetic Resonance in Medicine 2013;69(2):594–601. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Ohno K, Watanabe M. Electron paramagnetic resonance imaging using magnetic-field-gradient spinning. J Magn Reson 2000;143(2):274–9. 10.1006/jmre.1999.2005. [DOI] [PubMed] [Google Scholar]
  • 46.Shet K, Caia GL, Kesselring E, Samouilov A, Petryakov S, Lurie DJ, Zweier JL. A novel variable field system for field-cycled dynamic nuclear polarization spectroscopy. J Magn Reson 2010;205(2):202–8. 10.1016/j.jmr.2010.04.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Komarov DA, Hirata H. Fast backprojection-based reconstruction of spectral-spatial EPR images from projections with the constant sweep of a magnetic field. J Magn Reson 2017;281:44–50. 10.1016/j.jmr.2017.05.005. [DOI] [PubMed] [Google Scholar]
  • 48.Sato-Akaba H, Fujii H, Hirata H. Improvement of temporal resolution for three-dimensional continuous-wave electron paramagnetic resonance imaging. Rev Sci Instrum 2008;79(12):123701. [DOI] [PubMed] [Google Scholar]
  • 49.Sato-Akaba H, Kuwahara Y, Fujii H, Hirata H. Half-life mapping of nitroxyl radicals with three-dimensional electron paramagnetic resonance imaging at an interval of 3.6 seconds. Anal Chem 2009;81(17):7501–6. [DOI] [PubMed] [Google Scholar]

RESOURCES