15th European Molecular Imaging Meeting
supported by:

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

Best of vEMIM Talks | Imaging Technologies

Shortcut: BO-04
Date: Friday, 28 August, 2020, 2:00 p.m. - 3:30 p.m.
Session type: Spotlight Symposium


Abstract/Video opens by clicking at the talk title.


A multi spin echo pulse sequence with optimized excitation pulses and a 3D cone readout for hyperpolarized 13C imaging

Vencel Somai1, 2, Alan Wright1, Maria Fala1, Friederike Hesse1, Kevin Brindle1, 3

1 University of Cambridge, Cancer Research UK Cambridge Institute, Cambridge, United Kingdom
2 University of Cambridge, Department of Radiology, Cambridge, United Kingdom
3 University of Cambridge, Department of Biochemistry, Cambridge, United Kingdom


Imaging tumour metabolism in vivo using hyperpolarized [1-13C]pyruvate is a promising technique for detecting disease, disease progression and assessing treatment response1,2. However, the transient nature of the hyperpolarization and its depletion following excitation limits the available time for imaging3. We describe here a single shot multi spin echo sequence, which improves on previously reported sequences, with a shorter readout time, isotropic point spread function (PSF) and field of view (FOV) and better signal-to-noise ratio (SNR).


The sequence uses numerically optimized excitation pulses and hyperbolic secant adiabatic refocusing pulses, all applied in the absence of slice selection gradients. The excitation pulses were designed to give constant phase in the pass-band immediately after the pulse in order to increase signal in the presence of large B0 and B1 field inhomogeneities and to minimize the achievable echo time. The gradient readout uses a 3D cone trajectory distributed among 7 spin echoes (Figure 1). Experiments were performed at 7T (Agilent, Palo Alto, CA). A 42 mm diameter birdcage volume coil was used for 1H transmit and receive and a similar volume coil for 13C transmit. A 20 mm diameter surface coil was used for 13C receive (Rapid Biomedical GMBH, Rimpar, Germany).


The maximal gradient amplitude and slew-rate were set to 4 G/cm and 20 G/cm/ms respectively to demonstrate the feasibility of clinical translation. This gave a minimal repetition time of 165 ms. The pulse sequence had an isotropic FOV of 32 mm and nominal resolution of 2 mm and gave an isotropic PSF of 2.8 mm when reconstructed at the 0.125 mm in-slice resolution of the anatomical reference 1H image. The lack of slice selection, which was possible because of the isotropic FOV and the localized sensitivity profile of the receiver coil, enabled the design of a pulse optimised for frequency selectivity. Echo formation at the end of the pulse enabled very short echo times. The optimized spectral profile showed a high degree of spectral selectivity. The sequence was demonstrated with dynamic imaging of hyperpolarized [1-13C]pyruvate and [1-13C]lactate in a murine tumour model, and the data obtained were in good agreement with previous findings in this tumour model(Figure 2).


The pulse sequence was capable of dynamic imaging of hyperpolarized 13C labelled metabolites with relatively high spatial and temporal resolution. The segmented k-space readout fully exploited the long T2 relaxation time by sampling at multiple spin echoes to maximize the SNR. With a 32 cm FOV and 2 cm resolution, typical of a clinical scanner, the readout takes only 29.133 ms, and should preserve the high degree of robustness to imperfections.

[1] Gram A, Hansson G, Hansson L, et al. Increase in signal-to-noise ratio of. 2003:1-6. papers2://publication/uuid/B5F38F51-C552-4E4E-8B3C-6AC6B9109E56.
[2] Brindle KM. Imaging Metabolism with Hyperpolarized 13C-Labeled Cell Substrates. J Am Chem Soc. 2015;137(20):6418-6427. doi:10.1021/jacs.5b03300
[3] Day SE, Kettunen MI, Gallagher FA, et al. Detecting tumor response to treatment using hyperpolarized 13C magnetic resonance imaging and spectroscopy. Nat Med. 2007;13(11):1382-1387. doi:10.1038/nm1650
[4] Wang J, Hesketh RL, Wright AJ, Brindle KM. Hyperpolarized 13 C spectroscopic imaging using single-shot 3D sequences with unpaired adiabatic refocusing pulses. NMR Biomed. 2018;31(11):1-12. doi:10.1002/nbm.4004
[5] Wang J, Wright AJ, Hu DE, Hesketh R, Brindle KM. Single shot three-dimensional pulse sequence for hyperpolarized 13C MRI. Magn Reson Med. 2017;77(2):740-752. doi:10.1002/mrm.26168
Figure 1
The pulse sequence starts with an optimized excitation pulse, without a slice selection gradient, and contains 7 unpaired adiabatic refocusing pulses to generate 7 spin echoes and an optional 8th adiabatic pulse at the end of the sequence to flip back the spins parallel to the +z-direction in case of very long injection. At each echo a pair of two identical  cones were acquired in decreasing order with respect to cone angle. The fine structure of the first cone is not visible due to the plot linewidth. The whole sequence takes 165 ms.
Figure 2
Hyperpolarized [1-13C]pyruvate (A) and [1-13C]lactate (B) images from a tumor-bearing mouse overlaid on the corresponding 1H images (2 mm slice thickness). The 13C images were interpolated to the 128x128 in-plane matrix size of the 1H image and summed in time over the first 20 s. The relatively small signal leakage from neighbouring slices in slices 6 and 12 suggests that the simulated sampling PSF profile is well preserved.
Keywords: hyperpolarized, imaging, cone trajectory, tumour

Within-breath changes in small airway dimensions assessed in vivo by dynamic synchrotron radiation phase-contrast lung imaging in anesthetized rabbits with acute lung injury

Eva Solé Cruz1, 2, Luca Fardin1, 3, 4, Ludovic C. Broche1, Anders Larsson4, Gaetano Perchiazzi4, Alberto Bravin3, Sam Bayat1, 2

1 Inserm, UA7 STROBE Laboratory, Grenoble, France
2 Grenoble University Hospital, Dept. of Pneumology & Clinical Physiology, Grenoble, France
3 European Synchrotron Radiation Facility, Medical Beamline (ID17), Grenoble, France
4 Uppsala University, Hedenstierna Laboratory, Department of Surgical Sciences, Uppsala, Sweden


Micromechanical behaviour of individual small airways remain largely unknown in vivo in intact lung, mainly due to the lack of microscopic imaging techniques allowing for sufficient temporal and spatial resolution. We previously developed a time-resolved synchrotron radiation X-ray phase-contrast tomographic technique which allows to image intact lungs in vivo with 20 µm pixel resolution. Here we assessed the changes in small airway dimensions during the breathing cycle in anesthetized and mechanically ventilated rabbits with acute lung injury.


The experiment was performed on 4 anesthetized, tracheotomized, muscle-relaxed and mechanically ventilated rabbits. Lung injury was induced by whole lung lavage followed by injurious ventilation for 2 hours. Projection images were acquired at a constant frame rate at time resolution of 15 ms using a PCO edge 5.5 camera, coupled with optics determining a pixel size of 20 µm, during 30 minutes. Volumetric CT images were reconstructed at various phases of the respiratory cycle, during systole and diastole (at 10 and 150 ms after the ECG R wave, respectively). Small peripheral and terminal airway branches were manually segmented from the reconstructed CT images.


A sample segmented airway and daughter branches is shown in Figure 1. Figure 2 shows sequential measurements averaged along the length of a small bronchus and 2 terminal branches during the breathing cycle upon systole (S) and diastole (D). Airway pressure measured at the tracheal opening, ranged from 5 to 25 cmH2O from end-expiration to end-inspiration, respectively. Our preliminary results demonstrate that both airway pressure variation (p<0.05) and the phase of the cardiac cycle (p<0.001) significantly affected individual terminal airway dimensions, as airway radii were reduced during diastole with cardiac filling.


Using time resolved (4D) synchrotron radiation phase-contrast imaging, we demonstrate for the first time in intact in vivo lung, that both airway pressure and cardiac contractions determine small airway calibre variations in mechanically ventilated rabbits with acute lung injury. These findings bring new insight into the pathophysiology of ventilation-induced lung injury.

AcknowledgmentStudy funded by: The Swedish Reseach Council under grant 2018-02438; ESRF. 
Figure 1.

Sample segmented airway and daughter branches overlaid on greyscale tomographic image of a rabbit lung acquired with a pixel size of 20 mm. Inset shows magnification of the segmented airways. Arrows indicate daughter terminal branches.

Figure 2.

Mean radii of a main small bronchus and 2 terminal daughter branches during the breathing cycle upon systole (S) and diastole (D); Paw: airway pressure. Arrow indicates direction of time evolution.

Keywords: Pulmonary airways, Acute Respiratory DIstress Syndrome, X-ray computed tomography, synchrotron radiation

Pre-Resonance hyperspectral stimulated Raman scattering microscopy for monitoring Amphotericin B distribution

Konstantinos G. Mavrakis1, 2, 6, Minghua Zhuge3, 2, 6, Puting Dong5, 6, Kai-Chih Huang4, 6, Jie Hui2, 6, Ji-Xin Cheng2, 4, 6

1 University of Crete, Department of Materials Science and Technology, Heraklion, Greece
2 Boston University, Department of Electrical and Computer Engineering, Boston, United States of America
3 Zhejiang University, College of Optical Science and Engineering, Hangzhou, China
4 Boston University, Department of Biomedical Engineering, Boston, United States of America
5 Boston University, Department of Chemistry, Boston, United States of America
6 Boston University, Photonics Center, Boston, United States of America


Antimicrobial resistant infections are a growing threat especially for hospitalized patients since no new antibiotic class has been discovered the last 33 years. These emerging super bugs require novel treatment methods to overcome their constant evolution and adaptation. The first step towards this direction is to gain a deeper understanding of the interaction between antibiotics and microorganisms. In that end we present a highly sensitive Stimulated Raman Scattering (SRS) microscope with high spatial and spectral resolution which is able to resolve antibiotic induced alterations on cells.


The light source was a femtosecond laser with two synchronized outputs at 80 MHz repetition rate. The Stokes beam was fixed at 1045 nm whereas the other pump beam was tuned to 966 nm. Both beams were frequency doubled to 522.5 nm and 483 nm by focusing into non-linear crystals, and covered the C=C vibrational region. The Stokes beam was modulated by an acousto-optic modulator at 2.3 MHz, and the SRS signal was extracted by de-modulating the pump beam at the same frequency via a lock-in amplifier. In order to acquire an SRS spectrum, spectral focusing is utilized to transform Raman wavenumber to time delay between the two beams. Both beams were positively chirped to ~6 picoseconds by glass rods and the temporal delay was scanned with a motorized translation stage.


Conventional SRS imaging employs NIR lasers as light source due to minimized transient absorption background and deeper penetration depth. However, the detection sensitivity is at ~10mM molecular concentrations. By shifting the frequencies to the visible range, we exploit the pre-resonance effect which enhances the Raman cross-section several orders of magnitude1 which was recently used in SRS imaging 2,3. With the absorption peak of amphotericin b at 420nm and pump beam at 480nm, a theoretical sensitivity enhancement is calculated to be ~4000. The detection limit of amphotericin b is then at ~100μM which is close to typical values of antibiotic treating dosage and thus we could monitor the intracellular distribution of amphotericin b in treated cells. By tuning the polarization of the incident beams, we were able to determine the orientation of the antibiotic molecules relative to the cell membrane, which provides valuable information about the molecular mechanism of amphotericin b.


We have demonstrated a highly sensitive pre-resonance hyperspectral SRS microscope to in situ image the distribution of antibiotics in treated cells. Due to the pre-resonance effect, we were able to image amphotericin b at micromolar concentrations and probe the orientation of amphotericin b. The reported technology can be widely used to map low-concentration drug molecules at high sensitivity and high spatial resolution.

AcknowledgmentKonstantinos Mavrakis was financially supported by the Fulbright Institution.
[1] E. S. Yeung, M. Helling, and G. J. Small, “Pre-resonance Raman intensities,” Spectrochim Acta, vol. 31, no. 12, pp. 1921–1931, 1975.
[2] L. Wei et al., “Super-multiplex vibrational imaging,” Nature, vol. 544, no. 7651, pp. 465–470, 2017.
[3] H. J. Lee et al., “Electronic Preresonance Stimulated Raman Scattering Imaging of Red-Shifted Proteorhodopsins: Toward Quantitation of the Membrane Potential,” J. Phys. Chem. Lett., vol. 10, no. 15, pp. 4374–4381, 2019.
[4] T. M. Anderson et al., “Amphotericin forms an extramembranous and fungicidal sterol sponge,” Nat. Chem. Biol., vol. 10, no. 5, pp. 400–406, 2014.
Models of structure and function of amphotericin b
In a, the structures of the relevant biomolecules are shown. In the ion channel model (b), amphotericin b forms aggregates inside the cell membrane that permeabilize and incudes cell death. Surface adsorption model (c) and sterol sponge model (d) which induce cell death by binding and extracting ergosterol. Notice that in ion channel model amphotericin b is parallel to the lipid layer while in the other two models it is perpendicular.
Cells treated with Amphotericin b
Efficient signal generation occurs when the polarization of the incident beams is parralel to the vibrating molecule. As it is clearly evident in every cell case,the signal is stronger in the part of the cell membrane that has the lipid bilayer parallel to the polarization. When rotating the polarization 90 degrees so does the origin of the signal. This is a clear indication that the molecules of amphotericin b are also parallel to the polarization and therefore parallel to the lipid bilayer. These results stronly support the ion channel model.
Keywords: Stimulated Raman Scattering, Pre-Resonance, Hyperspectral, antibiotics, amphotericin b

An advanced machine learning algorithm for the automated detection of tissue chromophores from Multi-spectral photoacoustic images

Valeria Grasso1, 2, Joost Holthof1, Jithin Jose1

1 FUJIFILM Visualsonics, Amsterdam, Netherlands
2 Hannover Medical School, Institute for Animal Science, Hannover, Germany


To detect the tissue chromophores, multi-spectral Photoacoustic (PA) imaging frequently uses a differential based unmixing methods with a known spectral signature as a-priori information1. For the translational research with human patients, these types of supervised spectral unmixing can be challenging, as the spectral signature of the tissues differs with respect to disease condition. So here we present a machine learning approach, the non-negative matrix factorization (NNMF), for the automated detection of tissue chromophores. This is the first time NNMF is using to explore the PA images.


The multi observations (M) for the NNMF are modeled as: M=AS where A is the weight matrix and S is the spectra matrix, which are iteratively updated under the non-negativity constraints2. In this study, the feature extraction capability of other unsupervised algorithms3,4 such as PCA, ICA, has been tested on multi spectral PA images and compared with the NNMF approach. Multi-spectral PA images were acquired by using VevoLAZR-X (FUJIFILM Visualsonics, Toronto). A tissue mimicking phantom containing tubes with different dyes (ICG and Methylene Blue) was used to evaluate the sensitivity of the algorithms. In-vivo experiments were also performed by injecting Indocyanine green (ICG) intravenously and the NNMF has been used to unmix and quantify the endogenous and exogenous tissue chromophores.


The different unsupervised machine learning approaches have been tested by performing the tissue mimicking phantom imaging. Spectral images where obtained with in the wavelength range of 680-900 nm with a step size of 5 nm. PCA, ICA and NNMF were applied to unmix the contrast agents and the Signal to noise of each algorithm was evaluated. According to the SNR values, the performance of the NNMF approach was prominent. Furthermore, the kidney-spleen region of the mouse was imaged. Single wavelength PA images were obtained at 890 nm before and after the injection of ICG. From the images it’s clear that the contrast agent is accumulating in the region as the PA signal intensity is increased. NNMF algorithm was applied on the spectral images and Fig.1 shows the spectrally unmixed components after the injection of ICG. The prominent tissue chromophores like oxy and deoxy hemoglobin are clearly visible in the spectra together with the signature of ICG and it is visualized in the Fig. 2.


Relying on the non-negative nature of the PA images, the constraint of the NNMF unsupervised machine learning approach has ensured high feature detection performance. Here, for the first time, NNMF algorithm was tested on multi-spectral PA images and the automated detection of tissue chromophores was performed. In contrast to other algorithms, NNMF offers superior sensitivity to unmix and it is promising to translate into the clinical measures.


This work has received funding from the European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie grant agreement No 811226.

[1] Luke, G.P., Nam, S.Y. and Emelianov, S.Y., 2013. Optical wavelength selection for improved spectroscopic photoacoustic imaging. Photoacoustics, 1(2), pp.36-42.
[2] Lee, D.D. and Seung, H.S., 1999. Learning the parts of objects by non-negative matrix factorization. Nature401(6755), p.788.
[3] Montcuquet, A.S., Herve, L., Garcia, F.P.N.Y., Dinten, J.M. and Mars, J.I., 2010. Nonnegative matrix factorization: a blind spectra separation method for in vivo fluorescent optical imaging. Journal of biomedical optics15(5), p.056009.
[4] Glatz, J., Deliolanis, N.C., Buehler, A., Razansky, D. and Ntziachristos, V., 2011. Blind source unmixing in multi-spectral optoacoustic tomography. Optics express19(4), pp.3175-3184.
Figure 1
Figure1: (A) Expected spectral curves of oxy-deoxy hemoglobin and ICG; (B) Unmixed spectral curves obtained by NNMF
Figure 2

Figure 2: (A) Single wavelength PA image at 890nm; (B) Spectrally unmixed oxygenated hemoglobin; (C) deoxygenated hemoglobin; (D) ICG

Keywords: Photoacoustic imaging, Data quantification, Image processing

Transcranial Ultrasound Localization Microscopy reveals sub-resolution blood dynamics in aneurysms and arterial malformations in the adult human brain

Charlie Demené1, Justine Robin1, Baptiste Heiles1, Mathieu Pernot1, Fabienne Perren2, Mickael Tanter1

1 Inserm, ESPCI Paris, CNRS, PSL Research University, Physics for Medicine Paris, Paris, France
2 Université de Genève, Hôpitaux Universitaires de Genève, Clinical Neuroscience Department, LUNIC (Laboratory of Ultrafast-ultrasound Neuroimaging in Clinics), Geneva, Switzerland


Human brain vascular imaging is key for management of cerebrovascular and neurological pathologies. Challenging across modalities, it requires contrast injection, ionizing (CT) or expensive (MRI) imaging devices, overlooks blood dynamics and gives limited resolution. Ultrasound (US), conversely, is poorly used for neuroimaging due to limited sensitivity and resolution. US Localization Microscopy (ULM) has proven increased sensitivity and sub-resolution precision in the rat brain [1]. Transposed for the first time to human brain, we show that ULM is a game changer for clinical neuroimaging.


Experiments complied with the Declaration of Helsinki, patients gave informed and written consent (protocol 2017-00353 Geneva CCER). They were injected IV boluses of 0.3 mL of Sonovue before imaging through the temporal window with a 3-MHz phased array and an ultrafast scanner. Ultrafast US sequences consisted in diverging waves fired at 4800 Hz during 1s, looped every 2s, during 2 minutes. Tissue was filtered out using spatiotemporal SVD filtering [2], aberration corrections were calculated for isoplanatic patches thanks to local coherence optimization on isolated bubbles RF signatures before beamforming and motion compensation. Bubbles geometric centers were estimated using quadratic fitting, tracked and assigned to super-resolution trajectories using Hungarian algorithm.


Aberration corrections enabled to detect more bubbles and to refine the position of their geometric center. At typical f-numbers>4 in transtemporal imaging, theoretical ultrasonic lateral resolution is diffraction-limited to ~3 mm, while axial resolution is of the order of 0.8 mm. We show here that, with only 2 minutes of examination, vascular bed with diameters of the order 0.1 mm can be delineated, largely beating the diffraction limit and resolution of other clinical modalities, at depth up to 120 mm (~whole brain), with quantitative data on blood flow dynamics at a sub-resolution level.  Vortex flow in a 1.5 mm-wide aneurysm and parabolic speed profile on a 0.8 mm vessel section could be observed, which is impossible with any other neuroimaging modality. We quantified the resolution to be at least 62 µm. Complex flow pattern in a Moya-Moya syndrome could be observed, overstepping the partial information given by luminal-only clinical vascular imaging modalities.


By combining ULM with ultrafast diverging wave transmissions, skull aberration corrections and tissue motion compensation, we showed that we can redefine the reachable boundaries of cerebrovascular imaging, with 62µm resolution and very local blood flow dynamics assessment. This world premiere is a breakthrough for the management of cerebro-vascular diseases.

AcknowledgmentThe authors want to thank C Papadacci, V Hingot, A Dizeux for their help; and the European Research Council under the European Union’s Seventh
Framework Program (FP7/2007-2013)/ERC Advanced grant agreement n ̊ 339244-FUSIMAGINE , the technical support of the INSERM Technology Research Accelerator in Biomedical Ultrasound, and the Fond National Suisse.
[1]  Errico et al, Nature, 2015

Transcranial ULM in an aneurysm patient

A. ULM image obtained transcranially on a patient presenting a middle cerebral artery aneurysm. B. Zoom on the 1.7 mm-wide sub-resolution aneurysm exhibiting a vortex flow, as visible on the speed vector field reconstructed from bubble trajectories. C. A subwavelength analysis of the bubble speeds on the cross section (blue line) of a ~0,8mm diameter vessel show significant differences (* p-value < 0,01, ** p-value <0,005), revealing a typical parabolic profile.

Keywords: cerebrovascular, superresolution, ultrasound, ultrafast imaging

Aptamer-functionalized microbubbles: Ultrasound molecular imaging using an anti-P-selectin aptamer for imaging acute bowel inflammation.

Una Goncin1, Ronald Geyer2, Steven Machtaler1

1 University of Saskatchewan, Department of Medical Imaging, College of Medicine, Saskatoon, Canada
2 University of Saskatchewan, Department of Pathology and Laboratory Medicine, College of Medicine, Saskatoon, Canada


Aptamers are oligonucleotides that bind with high affinity and specificity to a range of targets. There are two limitations to their use in molecular imaging: they are quickly degraded by nucleases and rapidly filtrated via the kidneys. Ultrasound molecular imaging (US-MI) uses targeted microbubbles (MBs) that bind vascular disease markers 4-10 mins after injection, allowing for rapid detection. Our goal is to combine strengths of aptamers and MBs to create a targeted MB to P-selectin (a vascular inflammatory marker) and use it to image inflammation in a murine model of acute bowel colitis.


MBs were produced by sonicating a perfluorobutane-sparged solution containing solubilized phospholipids including DSPE-PEG2000-DBCO. MBs were incubated with either a fluorescent P-selectin-aptamer (P-Ap) (5’-azide-P-Ap-3’-Cy3) or non-fluorescent (5’-azide-P-Ap) for 15 min at 37°C, placed on ice for 5 min, and washed via centrifugation. Acute colitis was induced through rectal administration of TNBS in 5 of 8 mice. Mice were imaged using a preclinical ultrasound system following i.v. bolus of 1x108 MBs. MBs circulated for 4 min before the US-MI signal was collected. Each mouse received a bolus of both non-targeted and P-Ap-MBs separated by 20 min. Following imaging, bowels were excised and harvested for histology. Images were analyzed using VEVOCQ.


Successful labeling of MBs with P-Ap-Cy3 was verified with confocal microscopy (Figure 1A). There was a low US-MI signal detected using both non-targeted (0.4 ± 0.3 a.u.) and P-Ap-MBs (0.6 ± 0.8 a.u.) in control (no inflammation) mice. There was a significant increase in the US-MI signal in mice with acute colitis using the P-Ap-MBs (13.5 ± 8.1 a.u.)  in comparison to non-targeted MBs (3.2 ± 2.2 a.u.)(Figure 1B & C). US-MI signals correlated well to H&E histology (Figure 1D).


We constructed a targeted MB to P-selectin using an aptamer, which generated a detectable US-MI signal in mice with acute colitis. This approach for constructing quick, cost-efficient targeted MBs may represent a new generation of US-MI contrast agents that can be clinically translated.

Ultrasound molecular imaging (US-MI) using P-selectin aptamer for imaging acute bowel inflammation.
A) Schematic of perfluorobutane-filled lipid shelled MBs and successful labelling with fluorescent P-selectin aptamer (P-Ap; red). B) US imaging with non-targeted (left) and P-Ap targeted (right) MBs in control (bottom) and acute colitis mice (top) (B-mode: left; US-MI signal: right; green ROI highlights cross section of bowel). C) Bar graph (mean ± standard deviation) of US-MI signal using non-targeted and P-Ap targeted MBs in control (n=3) and acute colitis mice (n=5). D) H&E staining of bowels for control (left) and acute colitis mice (right; note large infiltration of immune cells).
Keywords: aptamer, P-selectin, inflammation, ultrasound, ultrasound molecular imaging