EMIM 2018 ControlCenter

Online Program Overview Session: PS-13

To search for a specific ID please enter the hash sign followed by the ID number (e.g. #123).

Hyperpolarization MRI & MPI Technologies

Session chair: David Boas - Bosten, USA; Markus Plaumann - Magdeburg, Germany
Shortcut: PS-13
Date: Thursday, 22 March, 2018, 1:30 PM
Room: Lecture Room 01 | level -1
Session type: Parallel Session


Click on an contribution preview the abstract content.

1:30 PM PS-13-1

Introductory Talk by Kevin Brindle - Cambridge, UK

This talk provides an overview of state-of-the-art research and refers to the following presentations selected from abstract submissions.

1:50 PM PS-13-2

SQUID based magnetic resonance imaging for the investigation of in situ and in vivo hyperpolarization techniques (#85)

K. Buckenmaier1, M. Rudolph1, 2, P. Fehling1, C. Back2, J. Bernarding3, D. Koelle2, R. Kleiner2, K. Scheffler1, M. Plaumann3

1 Max Planck Institute for Biological Cybernetics, High-Field Magnetic Resonance Center, Tuebingen, Baden-Württemberg, Germany
2 University of Tuebingen, Physikalisches Institut and Center for Collective Quantum Phenomena in LISA+, Tuebingen, Baden-Württemberg, Germany
3 Otto-von-Guericke University, Department for Biometrics and Medical Informatics, Magdeburg, Germany


Ultralow-field (ULF) nuclear magnetic resonance (NMR) is a promising spectroscopy method allowing for, e.g., the simultaneous detection of multiple nuclei. To overcome the low signal-to-noise ratio that usually hampers a wider application, we present an alternative approach to prepolarized ULF NMR employing hyperpolarization techniques like signal amplification by reversible exchange (SABRE) or Overhauser dynamic nuclear polarization (ODNP). Both techniques allow continuous hyperpolarization of 1H as well as other MR-active nuclei.


To be able to measure 1H and 19F simultaneously, a superconducting quantum interference device (SQUID)-based ULF NMR/MRI detection unit was constructed (see fig. 1). Due to the very low intrinsic noise level, SQUIDs are superior to conventional Faraday detection coils at ultralow-fields. Additionally, the broad band characteristics of SQUIDs enable them to simultaneously detect the MR signal of different nuclei such as 13C, 19F or 1H. Since SQUIDs detect the MR signal directly, they are an ideal tool for a quantitative investigation of hyperpolarization techniques such as SABRE or ODNP.


Using SABRE we successfully hyperpolarized fluorinated pyridine derivatives and quantitatively characterized the dependency of the magnetization transfer reaction from parahydrogen, which bonds to an iridium complex as well as to the 1H and 19F nuclei of an exchangeable ligand, as a function of hyperpolarization time and magnetic field strength [1]. Spectra (see fig. 2) and images of the samples were acquired.

With ODNP we were able to measure the coupling constant of solutions containing free radicals. Enhancement factors of over 100 were reached in in situ experiments. First proof-of-principle ex vivo images of rats using ODNP enhanced, SQUID based ULF-MRI have been acquired successfully.


We successfully built a SQUID-based ULF NMR/MRI system to quantitatively investigate the hyperpolarization techniques SABRE and ODNP.


[1]    Buckenmaier et al. SQUID-based detection of ultralow-field multinuclear NMR of substances hyperpolarized using signal amplification by reversible exchange. Scientific Reports, 7:13431 (2017).


We thank Hermann Mayer, Tomasz Misztal, Rebekka Bernard and Rolf Pohmann for their support in this project.

Figure 1.
Photo and scheme of the ULF MRI system (a), and scheme of the SQUID based magnetic field detector (b).
Figure 2.
Ultralow-field 19F and 1H MR spectra of hyperpolarized 3-fluoropyridine. Substances and catalysts were dissolved in methanol and measured at 144 µT. The blue lines represent the measured spectra whereas the green lines represent simulated spectra based on high-field determined coupling constants.
Keywords: SABRE, ODNP, SQUID, Hyperpolarization
2:00 PM PS-13-3

Deuteration of hyperpolarized 13C-labelled zymonic acid enables sensitivity-enhanced dynamic MRI of pH (#222)

C. Hundshammer1, 2, 4, S. Düwel1, 2, 4, S. S. Köcher2, 3, M. Gersch2, B. Feuerecker1, C. Scheurer2, A. Haase4, S. J. Glaser2, M. Schwaiger1, F. Schilling2

1 Technical University Munich, Klinikum rechts der Isar, Department of Nuclear Medicine, Munich, Bavaria, Germany
2 Technical University Munich, Department of Chemistry, Garching, Bavaria, Germany
3 Forschungszentrum Jülich, Institute of Energy and Climate Research (IEK–9), Jülich, North Rhine-Westphalia, Germany
4 Technical University Munich, Munich School of Bioengineering, 85748, Bavaria, Germany


Polarization techniques like DNP1 have been developed to overcome signal limitations of MRSI, which triggered the development of hyperpolarized (HP) pH sensors2 that potentially could help to improve precision medicine.3,4 13C-zymonic acid (ZA) is the first chemical shift based HP sensor that has been applied in vivo so far.5,6

We show that deuterated ZA (ZAd) exhibits an up to 39% longer spin lattice relaxation time in vitro, while its in vivo SNR can be increased by up to 46%. For the first time, we further demonstrate that ZAd is capable to sense dynamic pH changes with spatial resolution. 


ZA/ZAd synthesis & hyperpolarization (Hypersense, Oxford Instruments) was performed according to Hundshammer et al.6 T1 decay curves were measured at a 3T PET/MR scanner (Siemens, 15° FA, 5s TR).5,6 Chemical shift imaging (CSI) was performed at the PET/MR & at a 7T small animal MR scanner (GE/Agilent). Acquisition parameters, reconstruction schemes & pH calculation procedures are given by Hundshammer et al.For dynamic pH imaging, vitamin C was either added stepwise or as dissolvable vitamin C tablet to 72mL of 80mM Tris buffered solution after 3mL ZAd (cfinal=2mM) has been added & one image was recorded. For in vivo pH imaging, same amounts of ZA & ZAd (+urea) were polarized to saturation & injected into the tail-vein of three rats (n=3) bearing subcutaneous MAT-B-III tumors.6


At 3T, the T1 of C1 & C5 of ZAd are 49±8s & 71±3s respectively & thus 14% & 39% longer than the ones of ZA. In vivo, successive static CSI images demonstrate an SNR gain by deuteration amounting to 43±16% & 46±4% for C1 & C5, respectively (Figure 1). ZA & urea signals were observed in the vena cava & in the tumors. Representative spectra indicate a pH 7.53±0.08 (mean±std) in the vena cava compared to the extracellular tumor pH 7.10±0.05. Values are in agreement with the literature.5  

Stepwise addition of vitamin C with subsequent mixing delivered solutions of uniformly decreasing pH that could accurately be measured with ZAd (Figure 2). Addition of a dissolvable vitamin C tablet steadily decreased the pH of an aqueous solution of initial pH 7.58±0.02 (neutral) to pH≈5.0-5.7 (acidic). Mainly two peak pairs of ZAwere observed, because the CSI slice covered regions of acidic and neutral pH at the same time. The pH values were calculated by weighting peak intensities with respective pH.


Deuteration of zymonic acid prolongs the hyperpolarized signal lifetime, which can be used to non-invasively image pH in vivo. Furthermore we show that this sensor is usable to spatially resolve dynamic pH changes on a time-scale of seconds. In the future, this could help to detect immediate pH changes for instance in cases where proton concentrations are altered by enzymatic reactions,7 in alkaline treatments of acute metabolic acidosis8 & in exercised muscle.9 Additionally, time-resolved imaging of pH could verify targeted drug delivery for localized acidification or basification.


1             Ardenkjaer-Larsen, J. H. et al. Increase in signal-to-noise ratio of > 10,000 times in liquid-state NMR. Proceedings of the National Academy of                             Sciences of the United States of America, 100, 10158-10163, 2003.

2             Hundshammer, C. et al. Imaging of Extracellular pH Using Hyperpolarized Molecules. Israel Journal of Chemistry, doi:10.1002/ijch.201700017, 2017.

3             Collins, F. S. et al. A New Initiative on Precision Medicine. New England Journal of Medicine, 372, 793-795, 2015.

4             Friedman, A. A. et al. Precision medicine for cancer with next-generation functional diagnostics. Nature reviews. Cancer 15, 747-756, 2015.

5             Düwel, S. et al. Imaging of pH in vivo using hyperpolarized 13C-labelled zymonic acid. Nature Communications 8, 15126, doi:10.1038/ncomms1512,                   2017.

6             Hundshammer, C. et al. Deuteration of hyperpolarized 13C-labelled zymonic acid enables sensitivity-enhanced dynamic MRI of pH. Chemphyschem,                   doi:10.1002/cphc.201700779, 2017.

7             Gallagher, F. A. et al. Imaging pH with hyperpolarized 13C. NMR in biomedicine 24, 1006-1015, doi:10.1002/nbm.1742, 2011.

8             Som, A. et al. Monodispersed calcium carbonate nanoparticles modulate local pH & inhibit tumor growth in vivo. Nanoscale 8, 12639-12647, 2016.

9             Kemp, G. et al. Lactate accumulation, proton buffering, and pH change in ischemically exercising muscle. American journal of physiology. Regulatory,                     integrative and comparative physiology 289, R895-901; author reply R904-910, 2005.


We appreciate the help of Miriam Braeuer, Anna Bartels, Birgit Blechert & Michael Michalik with animal experiments. We acknowledge support from EU Grant No. 294582 (MUMI), BMBF (FKZ 13EZ1114) & DFG (SFB 824). We thank Cambridge Isotope Laboratories Inc. who provided [1-13C]pyruvate for the synthesis of zymonic acid.

Figure 1: In vivo SNR comparison of ZA & ZAd
ZA intensity images of ZAd (A) & ZA (E) overlaid on 1H images. (B, F) pH maps. (C, G) ZAd & ZA spectra of a vena cava voxel (white squares). (DH) Spectra of a tumor voxel (red squares) show two pairs of ZAd or ZA, which indicate a vascular tumor pH~7.5 & an acidic extracellular pH≈7.1 compared to the vena cava spectra showing only one ZAd or ZA peak pair corresponding to blood pH≈7.5.
Figure 2: Spatially resolved imaging of temporal pH changes.
(A, D) Experimental setups. B) Time-resolved pH maps of a beaker filled with 72mL of 80mM Tris at pH7.7 & 3mL co-polarized ZAd (cfinal=2mM) & urea (cfinal=3mM). C) Spectra of the mean of all pixels from B). E) Time-resolved pH maps after addition of a dissolvable vitamin C tablet to a solution as in B) (white arrow). F) Representative spectra of voxels (red squares) from timestep 1 & 5 from E). 
Keywords: hyperpolarization, pH sensors, dynamic MRSI, in vivo, breast cancer
2:10 PM PS-13-4

Hyperpolarized Silicon-29 Nanoparticles for In-Vivo MR Imaging (#43)

G. Kwiatkowski1, P. Wespi1, J. Steinhauser1, M. Ernst2, S. Kozerke1

1 University and ETH Zurich, Institute for Biomedical Engineering, Zurich, Switzerland
2 ETH Zurich, Labolatory of Physical Chemistry, Zurich, Switzerland


A new class of hyperpolarized Mangetic Resonance contrast agents, characterized with a long lifetime were recently proposed which are based on micro/nanoparticles of elemental silicon. Experimental results1,2 revealed a lifetime of the hyperpolarized signal >30min, exceeding those of any other 13C based probes reported so far. The bio-compatibility of silicon and its versatile surface chemistry makes it well suited for in vivo use3. The objective of the present work was to demonstrate the imaging capability of hyperpolarized nanometer size silicon particles in an experimental in-vivo setting.


Silicon nanoparticles: Silicon nanopowder obtained commercially, with average particle size (APS) ~ 55nm was used.

Hyperpolarization of silicon: The nuclear polarization of 29Si nuclei was enhanced by dynamic nuclear polarization (DNP) exploiting endogenous defects between the crystalline silicon core and the oxidized shell as electronic spin pool.

Surface functionalization: The surface of the silicon nanoparticles was functionalized with NHS-dPEG4-(m-PEG12)3-ester4, Heparin5, Dextran6 or Lipid bilayer7.

MRI experiments: After 24h of polarization, the samples were taken out of the polarizer and transferred to a horizontal 9.4 T imaging system.

All animal experiments were performed with adherence to the Swiss Animal Protection law and were approved by the regional veterinary office.


While early reports1,8,9 of hyperpolarized silicon particles demonstrated the feasibility of imaging silicon microparticles (APS ∼2 μm), the relatively large particle size did limit  in-vivo application. The present work shows that functionalized nanoparticles (APS∼55 nm) exhibit comparable relaxation and polarization properties (T1 ~ 30min, maximum achievable polarization 10-12 %), with improved in-vivo compatibility due to their smaller size. For in-vivo imaging, 30 mg of mPEG functionalized silicon nanoparticles were dispersed in 500 μl for an i.g. injection, which corresponded to a concentration of ∼90 mM of 29Si. The injected hyperpolarized silicon could be imaged in the two test animals in vivo with an average SNR/pixel inside the gastrointestinal tract of 12±4 and 9±2. The signal level measured with hyperpolarized silicon-29 was found to be comparable to the signal obtained with a standard concentration of hyperpolarized 1-13C pyruvate (80 mM) used in our laboratory10.


No significant effect of surface functionalization on DNP properties of silicon nanoparticles was found. Overall, good quality images were obtained despite the small amount of material used. In the future, further gains in SNR may be achieved by enriching the nanoparticles with silicon-29. Potential effects on relaxation properties, however, remain to be studied. In the example shown in this work, i.g. injection was used in the animals. In future work we will also investigate other routes of injection to study other organ systems of interest in nanomedicine applications.


1.        Cassidy MC, Chan HR, Ross BD, Bhattacharya PK, Marcus CM. In vivo magnetic resonance imaging of hyperpolarized silicon particles. Nat Nanotechnol. 2013;8(5):363-368. doi:10.1038/nnano.2013.65.

2.        Kwiatkowski G, Jähnig F, Steinhauser J, Wespi P, Ernst M, Kozerke S. Nanometer size silicon particles for hyperpolarized MRI. Sci Rep. 2017;7(July):1-6. doi:10.1038/s41598-017-08709-0.

3.        Park J-H, Gu L, von Maltzahn G, Ruoslahti E, Bhatia SN, Sailor MJ. Biodegradable luminescent porous silicon nanoparticles for in vivo applications. Nat Mater. 2009;8(4):331-336. doi:10.1038/nmat2398.

4.        Aptekar JW, Cassidy MC, Johnson  a. C, et al. Hyperpolarized Long-T1 Silicon Nanoparticles for Magnetic Resonance  Imaging. 2009:1-5. doi:10.1021/nn900996p.

5.        Argyo C, Cauda V, Engelke H, Rädler J, Bein G, Bein T. Heparin-coated colloidal mesoporous silica nanoparticles efficiently bind to antithrombin as an anticoagulant drug-delivery system. Chem - A Eur J. 2012;18(2):428-432. doi:10.1002/chem.201102926.

6.        Schulz A, Woolley R, Tabarin T, McDonagh C. Dextran-coated silica nanoparticles for calcium-sensing. Analyst. 2011;136(8):1722-1727. doi:10.1039/c0an01009j.

7.        Meng H, Wang M, Liu H, et al. Use of a Lipid-Coated Mesoporous Silica Nanoparticle Platform for Synergistic Gemcitabine and Paclitaxel Delivery to Human Pancreatic Cancer in Mice. ACS Nano. 2015;9(4):3540-3557. doi:10.1021/acsnano.5b00510.

8.        Whiting N, Hu J, Shah J V., et al. Real-Time MRI-Guided Catheter Tracking Using Hyperpolarized Silicon Particles. Sci Rep. 2015;5:12842. doi:10.1038/srep12842.

9.        Whiting N, Hu J, Constantinou P, et al. Developing hyperpolarized silicon particles for advanced biomedical imaging applications. Proc SPIE. 2015;9417:941702. doi:10.1117/12.2082252.

10.      Krajewski M, Wespi P, Busch J, et al. A Multisample Dissolution Dynamic Nuclear Polarization System for Serial Injections in Small Animals. Magn Reson Med. 2017;77:904–910. doi:10.1002/mrm.26147.


We would like to thank Daniel Zindel, Laboratory of Physical Chemistry, ETH Zurich for his help with the surface modification of diamond powder. Financial support by the Swiss National Science Foundation (SNF grants 320030_153014, 200021_149707 and 200020_169879) are gratefully acknowledged.

Figure 1.
Anatomical image in coronal (A) and transvers plane (B). Corresponding silicon-29 images recorded at 5 min after sample injection (500 μl of 60 mg/ml) C,D). Overlay of the two images threshholded at 35% threshold level for transparency E,F).
Keywords: hyperpolarization, dnp, MRI, silicon, nanoparticles
2:20 PM PS-13-5

An in vivo metabolic imaging study of myopathy in transgenic mice using hyperpolarized [1-13C]pyruvate generated by ParaHydrogen (#424)

E. Cavallari1, C. Carrera1, M. Sorge1, S. Aime1, F. Reineri1

1 Torino University, Department of molecular biotechnology and health sciences, Torino, Italy


Hyperpolarized [1-13C]pyruvate is an MRI probe that allows to monitor metabolism in vivo, in real time. ParaHydrogen Induced Polarization (PHIP) is a portable, cost effective technique able to generate hyperpolarized molecules in few seconds. The introduction of the Side Arm Hydrogenation (SAH) strategy offered a way to widen the field of PHIP generated systems and to make this approach competitive with the currently applied dissolution-DNP (Dynamic Nuclear Polarization) method.

Here, the first in vivo metabolic study using the PHIP-SAH hyperpolarized [1-13C]pyruvate is reported.


Biocompatible aqueous solution of HP pyruvate was obtained according to the PHIP-SAH procedure. Hydrogenation was carried out in an organic phase in order to eliminate the toxic methanol used in the previously reported proof-of-concept. Hydrolysis was obtained using a diluted aqueous base (NaOH 0.1M) and sodium ascorbate (as scavenger of paramagnetic impurities), physiological pH was reached adding Hepes (144mM). The concentration of the injected pyruvate was 50±5mM (doses of HP pyruvate: 0.3-0.35 mmol/Kg).

4-months-old male (n=3) and 6-months-old male (n=4) LmnaH222P/H222P mice and 6-months-old male WT mice (n=4) were investigated.

13C dynamic studies were performed using series of 13C-MR spectra, with small flip angle pulses and a tailored slice-selective sequence, at 1 Tesla (2M Aspect).


The hydrolysis and phase extraction steps took few tens of seconds, nevertheless, the HP observed on the 13C carboxylate signal (3.9±0.5%) appears already sufficient for carrying out in vivo studies.

The pyruvate-lactate exchange rate of the 13C hyperpolarized label was obtained, using the model free approach based on the ratio of the total area under the curve of these metabolites.

The results clearly show a marked decrease of lactate/pyruvate ratio in heart muscle of the LmnaH222P/H222P with respect to WT mice. The reduced pyruvate/lactate exchange rate is a reporter of the general cells metabolic activity and might be due either to lower activity of the transporters (MCT) or to an altered cytosolic redox state.

No difference on kidney metabolism was observed, this observation supports the view that the metabolic result obtained in the heart reports on differences in the metabolism of the cardiac tissue and not from an impairment in the distribution of HP metabolites in the blood pool.


This study shows that in vivo metabolic investigations based on the administration of [1-13C]pyruvate obtained from the PHIP-SAH procedure is possible, also at this relatively low magnetic field strength (1 Tesla).

By comparing Pyruvate-Lactate 13C label exchange rate in healthy and genetically modified Lmna mice, it was found that the metabolic dysfunction occurring in the cardiac muscle of the transgenic (LmnaH222P/H222P) mice can be detected well before the disease can be assessed by echographic investigations.


Reineri F.et al.Nat.Commun.2015;6:5858

Cavallari E.et al.J.Phys.Chem.B,2015,Vol. 119,10053-10041

Arimura, T. et al. Hum. Mol. Genet., 2005; 14, 155–169


We gratefully acknowledge AIRC (2015 TRIDEO call) for financial support.

Figure 1

Lactate/pyruvate ratios obtained from the 13C-MR spectra centered on the heart (red symbols) of mutant Lmna mice and WT mice. The L/P ratio in the heart of Lmna is significantly lower than in WT mice, already at 4 months of age and becomes more significant on 6 months old Lmna mice. When the 13C-MR spectra are centered on the kidneys (blue symbols) there is not any significant difference.

Figure 2

a) Stacked plot of 13C-MR spectra at 1T from a dynamic study on the heart of a WT mouse; spectra were acquired every 2’’ from a 12 mm thick axial slice centered on the heart. The lactate peak is easily identifiable (182 ppm) (13Cpyruvate signals 170 ppm).

b) Dynamic curves of metabolite levels in same slice, the metabolite level was obtained from the integral of the metabolite peak.

Keywords: Hyperpolarization, Pyruvate metabolism, Heart dysfunction
2:30 PM PS-13-6

Eight-Fold Resolution Improvement in Magnetic Particle Imaging with Pulsed Waveforms For Large Core Size Magnetic Nanoparticles (#478)

Z. W. Tay1, D. Hensley2, P. W. Goodwill2, B. Zheng1, S. M. Conolly1

1 University of California, Berkeley, Bioengineering, Berkeley, California, United States of America
2 Magnetic Insight, Inc, Alameda, California, United States of America


Magnetic Particle Imaging (MPI) is an emerging tracer modality with high sensitivity [1-2] and zero ionizing radiation. Applications include stem cell tracking [3], tumor imaging [4], stroke detection / angiography [5] and gut bleed imaging [6]. One challenge is improving the spatial resolution of MPI (currently ~ 1.5 mm w/o deconvolution). While Langevin theory predicts a cubic improvement in MPI resolution with increasing magnetic nanoparticle (tracer) core size, relaxation blurring has been a barrier[7]. Here, we use pulsed waveforms to circumvent relaxation and achieve Langevin resolution.


Imagion Biosystems PrecisionMRX® SPIO nanoparticles (Imagion Biosystems, Albuquerque, NM, USA) with carboxylic acid coated outer shell and varying core diameters were used. All single cores are single crystalline magnetite (Fe3O4) as confirmed in [8]. Fig2 used VivotraxTM (Magnetic Insight, Alameda, CA, USA) for small-core size images. 1D point spread functions and resolution measurements were obtained on the Berkeley arbitrary waveform relaxometer (1D scanner) described in [9]. 2D resolution simulations were performed using custom matlab code and assuming a Debye relaxation model. 2D images were obtained on a modified version of the abovementioned relaxometer with a permanent magnet field-free-line configuration with gradient strength of 3.5 x 3.5 T/m for the x and z axes. 


Pulsed MPI uses square waveforms to wait for MPI tracer magnetization relaxation to complete before scanning the next voxel (Fig 1b). In contrast, standard MPI does not wait and causes MPI signal to blur between voxels. As a result, Pulsed MPI is able to circumvent relaxation-induced blurring that worsens with larger core sizes to uncover the cubic resolution improvement with core size predicted by the Langevin Model. Experimental 1D measurements (Fig 1c) show achieved resolution closely approaches the Langevin Model. This is verified by 2D simulations (Fig 1d) and 2D imaging data (Fig 2). Standard MPI has been limited to smaller core sizes (< 25 nm) due to increased relaxation-induced blurring at larger core sizes. With Pulsed MPI, we can unlock the potential of larger core sizes, achieving 0.6 mm resolution with our current setup. By improving gradient to 7 T/m and increasing core size from 27.4 nm to 40 nm, ~ 0.1 mm resolution should be achievable (before deconvolution).


While Pulsed MPI needs significant MPI hardware changes, unlocking cubic resolution gains with tracer core size can dramatically improve resolution (8-fold for 25 nm to 50 nm). Using large core sizes also gives increased sensitivity to the microenvironment for binding contrast and in vivo viscosity measurements [10]. Lastly, the resolution gains from pulsed MPI can be traded-off for lower field gradients to make clinical translation of MPI hardware easier and to enable faster scanning in situations where scan speed is limited by gradient vs. SAR and magnetostimulation safety limits [11]. 


  1. Gleich B, Weizenecker J. Tomographic imaging using the nonlinear response of magnetic particles. Nature. 2005 Jun 30;435(7046):1214–1217.
  2. Graeser M, Knopp T, et al. Towards Picogram Detection of Superparamagnetic Iron-Oxide Particles Using a Gradiometric Receive Coil. Sci Rep. 2017 Jul 31;7(1):6872. 
  3. Them K, Salamon J, et al. Increasing the sensitivity for stem cell monitoring in system-function based magnetic particle imaging. Phys Med Biol. 2016 May 7;61(9):3279–3290. 
  4. Yu EY, Bishop M, et al. Magnetic Particle Imaging: A Novel in vivo Imaging Platform for Cancer Detection. Nano Lett [Internet]. 2017 Feb 16; 
  5. Ludewig P, Gdaniec N, et al. Magnetic Particle Imaging for Real-Time Perfusion Imaging in Acute Stroke. ACS Nano. 2017 Oct 24;11(10):10480–10488. 
  6. Yu EY, Chandrasekharan P, et al. Magnetic Particle Imaging for Highly Sensitive, Quantitative, and Safe in Vivo Gut Bleed Detection in a Murine Model. ACS Nano [Internet]. 2017 Nov 30; Available from: http://dx.doi.org/10.1021/acsnano.7b04844 PMID: 29165995
  7. Tay ZW, Hensley D, et al. The relaxation wall: experimental limits to improving MPI spatial resolution by increasing nanoparticle core size. Biomed Phys Eng Express [Internet]. IOP Publishing; 2017 Apr 3 
  8. Vreeland E C et al. 2015 Enhanced nanoparticle size control by extending lamers mechanism Chem. Mater. 27 6059–66
  9. Tay ZW, Goodwill PW, et al. A High-Throughput, Arbitrary-Waveform, MPI Spectrometer and Relaxometer for Comprehensive Magnetic Particle Optimization and Characterization. Sci Rep. Nature Publishing Group; 2016 Sep 30;6:34180.
  10. Fidler F, Steinke M, et al. Stem Cell Vitality Assessment Using Magnetic Particle Spectroscopy. IEEE Trans Magn. ieeexplore.ieee.org; 2015 Feb;51(2):1–4.
  11. Saritas EU, Goodwill PW, et al. Magnetostimulation Limits in Magnetic Particle Imaging. IEEE Trans Med Imaging. 2013;32(9):1600–1610.


We would like to acknowledge NIH funding and the A*STAR NSS-PhD and the Siebel Scholars fellowship (ZW Tay).

Pulsed MPI Drive Waveforms achieves Cubic Resolution Improvement with Tracer Core Size
a) Photo of 2D Arbitrary Waveform Relaxometer used.  b) Square (pulsed) drive waveforms waits for tracer relaxation to complete before scanning the next voxel in contrast to standard (sine) continuous scanning.  c) Pulsed MPI achieves cubic resolution improvement with core size (Langevin model) while sine MPI stops improving at 25 nm.   d) 2D simulation shows improved resolution with Pulsed MPI
Experimental 2D MPI Images show Eight-Fold Resolution Improvement with Pulsed MPI Drive Waveforms
Pulsed MPI circumvents relaxation blurring to uncover the good resolution native to large core sizes. 0.6 mm spacing is just resolved for a 27.4 nm core tracer. This is 8-fold better than Standard MPI on the same tracer that has ~ 5.0 mm resolution. Standard MPI typically uses small core sizes with poorer native resolution but no relaxation blurring. Pulsed MPI is 5-fold better than std MPI setup.
Keywords: Magnetic Particle Imaging, High Resolution, Magnetic Nanoparticles
2:40 PM PS-13-7

Noninvasive Detection and Quantification of Gastrointestinal Bleeding with Magnetic Particle Imaging (#346)

E. Y. Yu1, P. Chandrasekharan1, R. Berzon1, Z. W. Tay1, X. Y. Zhou1, A. P. Khandhar2, R. M. Ferguson2, S. J. Kemp2, B. Zheng1, P. W. Goodwill3, M. F. Wendland1, K. M. Krishnan2, 4, S. Behr5, J. Carter6, S. M. Conolly1, 7

1 University of California, Berkeley, Department of Bioengineering, Berkeley, California, United States of America
2 Lodespin Labs, LLC, Seattle, Washington, United States of America
3 Magnetic Insight, Inc., Alameda, California, United States of America
4 University of Washington, Department of Material Science and Engineering, Seattle, Washington, United States of America
5 University of California, San Francisco, Department of Radiology and Biomedical Imaging, San Francisco, California, United States of America
6 University of California San Francisco Medical Center, San Francisco, California, United States of America
7 University of California, Berkeley, Department of Electrical Engineering and Computer Science, Berkeley, California, United States of America


Colon polyps are a known cause of gastrointestinal bleeding (GIB) [1]. 99mTc-RBC scintigraphy is the most sensitive imaging technique for GIB detection. However, drawbacks include radioactive dose, poor spatial resolution (~4 mm), bulky radio-pharmaceutical agent preparation and long scan times, which can delay surgical interventions. We present the use of Magnetic Particle Imaging (MPI) [2-5] to detect GIB using long-circulating superparamagnetic iron-oxide (SPIO) tracer as the vascular tracer agent in a mouse model of Familial Adenomatous Polyposis.


Five 12 week old C57BLK6/Apcmin/+ mice [6] with a genetic mutation in FAP (JAX® Laboratory) (Hct=0.21-0.30) [7] were used in this study. Three Wild-type C57BL/6 mice were used as control. SPIOs (LodeSpin-017, 5 mg Fe/kg in 100 μL) and heparin were injected through a lateral tail vein catheter for each animal. MPI was performed with a custom-built vertical bore field-free line (FFL) MPI scanner with a gradient strength of 6.3 T/m. Twenty-one dynamic projection MPI scans were acquired with respiratory gating over 130 minutes. Region-Of-Interest based compartment fitting was performed after converting the MPI signal to iron concentration for flow quantification.


MPI images of Apcmin/+ and wild type mice at 2 time points post-SPIO injection are shown in Fig. 1. MPI signal was observed to accumulate in the gut lumen of Apcmin/+ mice, whereas the wild type showed signal throughout the vasculature, typical of LS-017 (a blood pool tracer). In Fig. 2 (a), the MPI image at the first time point was digitally subtracted from all images in time course to capture positive tracer accumulation. The GIB is visualized with extraordinary contrast in the Apcmin/+ mice, whereas minimal positive accumulation is observed in the control mice.

The rate of tracer accumulation of the gut lumen was between 1 – 5 µL/min. A representative two-compartment model fitting result is shown in Fig. 2 (b). GIB was detected using MPI at a sensitivity rivaling that of 99mTc-RBC scintigraphy at clinically relevant doses, but by using a blood-pool, non-radioactive SPIO tracer [8]. Additional sensitivity improvement is possible through tracer and receive hardware optimization [9,10].


In this work, we have demonstrated highly sensitive detection of GI bleeding in a murine model using MPI. Although there is still a long path to clinical translation for MPI tracers and imager, MPI is a clinically translatable imaging modality with superb contrast, sensitivity, linear quantitation and safety. We believe that in the future, MPI could complement the current clinical workflow for cases of occult or obscure GI bleeding. This may significantly improve the accuracy of diagnosis and ultimately reduce cost and improve outcome.


[1] Carrera, V. G., Villaverde, A. F., Campos, A. C. & López, S. V. Acute lower gastrointestinal bleeding from a polyp. Ann. Gastroenterol. Hepatol. 27, 263 2014.

[2] B. Gleich and J. Weizenecker. Tomographic imaging using the nonlinear response of magnetic particles. (2005). Nature, 435(7046):1217-1217, 2005. 

[3] T. Knopp and T. M. Buzug. Magnetic Particle Imaging: An Introduction to Imaging Principles and Scanner Instrumentation. Springer, Berlin/Heidelberg, 2012. [4] Goodwill, P. W. & Conolly, S. M. The X-space formulation of the magnetic particle imaging process: 1-D signal, resolution, bandwidth, SNR, SAR, and magnetostimulation. IEEE Trans. Med. Imaging 29, 1851–1859, 2010. 

[5]  Goodwill, P. W., Konkle, J. J., Zheng, B., Saritas, E. U. & Conolly, S. M. Projection x-space magnetic particle imaging. IEEE Trans. Med. Imaging 31, 1076–1085, 2012.

[6]  Wei, H., Shang, J., Keohane, C., Wang, M., Li, Q., Ni, W., O’Neill, K. & Chintala, M. A novel approach to assess the spontaneous gastrointestinal bleeding risk of antithrombotic agents using Apc(min/+) mice. Thromb. Haemost. 111, 1121–1132, 2014.

[7] Graça, B. M., Freire, P. A., Brito, J. B., Ilharco, J. M., Carvalheiro, V. M. & Caseiro-Alves, F. Gastroenterologic and radiologic approach to obscure gastrointestinal bleeding: how, why, and when? Radiographics 30, 235–252, 2010. 

[8] Yu, E. Y., Chandrasekharan, P., Berzon, R., Tay, Z., Zhou, X. Y., Khandhar, A. P., Ferguson, R. M., Kemp, S. J., Zheng, B., Goodwill, P. W., Wendland, M. F., Krishnan, K. M., Behr, S., Carter, J., and Conolly, S. M. Magnetic Particle Imaging for Highly Sensitive, Quantitative, and Safe in vivo Gut Bleed Detection in a Murine Model. ACS Nano, 2017. 

[9] Ferguson, R. M., Minard, K. R., Khandhar, A. P. & Krishnan, K. M. Optimizing magnetite nanoparticles for mass sensitivity in magnetic particle imaging. Med. Phys. 38, 1619–1626, 2011. 

[10] Zheng, B., Goodwill, P. W., Dixit, N., Xiao, D., Zhang, W., Gunel, B., Lu, K., Scott, G. C. & Conolly, S. M. Optimal Broadband Noise Matching to Inductive Sensors: Application to Magnetic Particle Imaging. IEEE Trans. Biomed. Circuits Syst, 2017. 


The authors would like to acknowledge funding support from NIH 5R01EB019458-03, NIH 5R24MH106053-03, UC Discovery Grant 29623, W. M. Keck Foundation Grant 009323, and NSF GRFP for this work. Additionally, work at Lodespin Labs and University of Washington was supported by NIH 1R41EB013520-01 and NIH 2R42EB013520-02A1.

Dynamic projection MPI captures whole body tracer bio-distirbution.

Figure 1 Representative MPI images with CT overlay of an ApcMin/+ (left) and Wild-Type (right) mouse over time. MPI clearly captures dynamics of tracer extravasation into the gut in the ApcMin/+ mouse, whereas no tracer extravasation into the gut is seen in the Wild-Type mouse.

MPI quantitatively captures positive tracer accumulation over time.

(a) Representative subtracted MPI images of an ApcMin/+ mouse (left) and a wild-type mouse (right) over time with CT overlay. The GI bleed is visualized with extraordinary contrast in the ApcMin/+ mouse, while for the wild-type mouse, the tracer stayed in the blood pool throughout the study, hence minimal subtracted signal. (b) Two compartment model fit results for ApcMin/+ and wild-type mice.

2:50 PM PS-13-8

Retrospectively gated cardiac PET-MR imaging in rodents using MRI-based cardiac motion information. (#549)

W. Gsell1, 2, A. Nauerth3, C. Molinos4, C. Correcher4, A. J. Gonzalez5, S. Sven3, T. Greeb3, R. Polo4, B. Holvoet2, 6, C. M. Deroose2, 6, U. Himmelreich1, 2, M. Heidenreich3

1 University of Leuven, Imaging and Pathology, Biomedical MRI, Leuven, Belgium
2 University Of Leuven, MoSAIC, Leuven, Belgium
3 Bruker BioSpin, Preclinical Imaging MRI, Ettlingen, Germany
4 Bruker BioSpin, Preclinical Imaging NMI, Valencia, Spain
5 Institute for Instrumentation in Molecular Imaging, i3M-CSIC, Valencia, Spain
6 University of Leuven, Imaging and Pathology, Nuclear Medicine, Leuven, Belgium


Preclinical PET inserts for MRI scanners enable simultaneous assessment of cardiac function, tissue integrity and molecular pathways. The acquisition of cardiac PET-MRI in rodents remains challenging due to the short cardiac cycle, implying the use of rapid imaging with ECG gating. Strong gradients and fast gradient switching can interfere with the ECG signal, rendering the conventional ECG gating very difficult if not achievable. The aim of this work was to test a method based on an MRI self-gating technique (IntraGate)1 to provide the gating information for retrospective PET reconstruction.


PET-MRI data were acquired on a BioSpec 70/30 MRI system equipped with a SiPM based PET insert (Bruker Biospin), using a quadrature volume coil (diameter of 86 mm). Wistar rats und B6 mice were used. Animals were not fasted to ensure good FDG uptake. Anaesthesia was induced and maintained through inhalation of 2% Isoflurane carried by 100% oxygen. 18F-FDG (46.1 ± 9.7 MBq) was administered intravenously through the tail vein. For MRI, we used a self-gated gradient echo sequence (igFLASH) with the following parameters: TE: 3,585 ms, TR: 10 ms, flip angle: 15 degrees, matrix: 128x128, FOV: 60x60 mm, 10 contiguous slices of 1.5 mm thickness, scan time: 53:20 min. Simultaneously, PET data were acquired using a 1hr static scan. Retrospective PET-MRI was achieved using IntraGate (see below).


The MRI sequence was modified to send a TTL signal of 4 ms duration at each TR loop to the PET DAQ electronics. Then, the IntraGate navigator information was used to derive the required retrospective data ordering that represents the position within the cardiac cycle (Fig.1). The list-mode PET data were rebinned accordingly to reconstruct 4-16 cardiac frames. All PET data were reconstructed using MLEM with 12 iterations, no smoothing, no PSF modeling and 0.5mm isotropic resolution.

MRI data were first evaluated providing a time-resolved cardiac movie with up to 20 frames per cardiac cycle identifying the end diastole and end systole. The same information used for MRI reconstruction was used to achieve a time synchronization of simultaneous PET/MRI. A first application of this technique has resulted in self-gated cardiac imaging, both in mice and rats. The cardiac cycle has been correctly resolved in both modalities and allows for a dynamic co-registration of the images (Fig.2).


We hereby propose for the first time a method enabling PET cardiac imaging solely based on motion information derived from MRI. Our preliminary results demonstrate the advantage of such technique for true synchronization of the PET and MRI cardiac acquisition, the easier animal handling (no ECG electrodes required), access to the full cardiac cycle and the possibility to retrospectively assess the quality of the gating information.


[1] A. Nauerth, E. Heijman, C. Diekmann. Slice refocusing signal for retrospective reconstruction of CINE cardiac MR images. Proc. Intl. Soc. Mag. Reson. Med. 14 (2006).

Figure 1

Principle of the retrospectively gated PET-MRI technique: Navigator information form the intragate MRI sequence enable to extract information about respiration and cardiac motion (red box) used to assign each image to the corresponding cardiac frame, retrospectively. PET and MRI are synchronized through a TTL signal enabling to assign each time stamp of the PET list-mode to a cardiac frame.

Figure 2

PET and MRI data fusion using 8 frames across the cardiac cycle starting from end-systole (yellow box).

Keywords: PET-MRI, cardiac, retrospective gating