P-01 to P-15 are presented as posters only. T-01 to T-08 are also presented as talks.
All posters are displayed during the entire duration of the TOPIM TECH. DURING the DEDICATED POSTER SESSION: The presenters are present at their posters. All participants are invited to come along and discuss the work with the authors. Independent of this, two chairpersons are going to evaluate each poster. An award will be presented for the best poster presentation.
Click on an contribution to preview the abstract content.
Novel optical approaches for cancer theranostics using fluorescent quantum-dots polymer microcarriers. (#16)
Fernanda Ramos Gomes1, Galina Nifontova2, Maria Baryshnikova2, 3, Frauke Alves1, 4, Igor Nabiev2, 5, Alyona Sukhanova5
1 Max-Planck Institute for Experimental Medicine, Translational Molecular Imaging Group, Göttingen, Germany
Fluorescent imaging has already been used as a diagnostic tool to monitoring tumour progression and inflammation. Compared to other fluorescent probes, quantum dots (QDs) encapsulated into polymer microcarriers (MicroQDs) exhibit lower photobleaching and higher brightness, making them attractive probes for targeted delivery and imaging. Therefore, unfunctionalized and anti-HER2 antibody functionalized MicroQDs were imaged in live cell assays and their potential use as imaging tools and theranostic agents were explored.
MicroQDs were validated by confocal microscopy techniques, providing information about their size, morphological modifications and optical characteristics. Intracellular uptake of the unfunctionalized MicroQDs was measured using live cell imaging assays with the MH-S murine alveolar macrophage cell line during short-term (4 h) or long-term (24 h) incubations. The validation of anti-HER2 antibody functionalized MicroQDs probes were performed using live HER2 overexpressing human breast cancer cells BT474 to determine whether the produced conjugates specifically interact with HER2 positive cancer cells. Real-time monitoring of the interaction between the MicroQDs and the cancer BT474 cells were performed using IncuCyte Zoom live-cell analysis equipment and images were taken every 2 h.
We show that stable and bright unfunctionalized MicroQDs were effectively taken up by the phagocytic MH-S cells. Signs of uptake are present in both, short- and long-term incubations, where the green-fluorescent signals from the MicroQDs are seen in close contact with the red-fluorescent signals from the DRAQ5 stained cell nuclei.
Different concentrations of the anti-HER2 antibody on the MicroQDs surface were tested (3, 30 and 300 µg) in incubation assays with BT474 cancer cells. The visualization of the MicroQDs-derived green-fluorescent signals revealed that the interaction ability of the anti-HER2 functionalized MicroQDs increases with the amount of anti-HER2 on the microparticle surface due to the higher density of the covalently coupled anti-HER2 molecules. As a negative control, unfunctionalized MicroQDs were used, of which negligible amounts were detected on the cell surface of the cancer cells.
We show that the uptake of MicroQDs, unfunctionalized or antibody functionalized, confirms the possibility of their efficient use for live cell imaging and transportation of drugs and delivery within living cells. This method could provide a novel basis for the development of new cancer cell targeting theranostic probes based on encapsulated-QDs polymer microcarriers.
We thank Sarah Garbode and Bärbel Heidrich for excellent technical assistance.
Keywords: microcarriers, theranostics, quantum-dots, cancer cells
Characterisation of a New Imaging System for Gamma-Near-Infrared Fluorescence-Guided Surgery (#1)
Awad M. Almarhaby1, 2, John Lees1, Sarah Bugby1, Mohammed Alqahtani3, Layal Jambi4, William McKnighta1, Alan Perkins5
1 University of Leicester, Physics and Astronomy, Leicester, United Kingdom
A novel hand-held hybrid optical-gamma camera (HGC) has previously been described that is capable of displaying co-aligned images from both modalities in a single system. Here, a dedicated NIR imaging system for surgical guidance has been developed for combination with the HGC. This work has evaluated the performance of two NIR imaging systems using phantom studies, various fluorophores and setups. Further improvements will combine one of these systems with the HGC for dual gamma-NIR fluorescence intraoperative imaging.
Different setups were used to study system performance under a range of simulated situations with variables including dye concentrations, depth of target and target diameter. Two cameras were evaluated in this experiment to detect the fluorescence agents which were illumiated by a NIR LED ring. For cancerious targets simulations, phantoms with different sizes were constructed. In order to characterise a target at depth within the body, a tissue-like phantom with different thicknesses was used. Two identical setups were used for both cameras, namely above and to the side of the phantoms. A primary aim was to measure the signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) for both cameras and were compared to the existing NIR imaging systems.
Both cameras were able to detect ICG dye concentrations between about 0.8 to 80 μM, while the concentration ranges for 800CW dye began to be observed at 10-2 μM and plateaued at high concentrations up to 10 μM. The optimal concentration of ICG (80 μM) was still comparable to the in vivo injected dose (100 μM). The effect of the positioning of the excitation light was investigated using both dyes. When the excitation ring was in the angled position, higher CNR values were observed, such that smaller wells (0.5, 1 mm) could be identified with a lower surrounding background signal. A depth of 7 mm was found to be the maximum detection depth, after which it became difficult to identify NIR fluorescence targets even with diameters as large as 10 mm using either camera or either dye in either LED ring position. For targets with smaller diameters (between 0.5-4 mm) the maximum depth for fluorescence imaging was approximately 3 mm due to the weak signal emitted from the micro-volumes used.
Both imaging systems’ detection capabilities were determined for the purpose of imaging targets using different dye concentrations.The minimum and maximum detectable concentrations of ICG and 800CW were determined. A tissue-like phantom was also used in this evaluation to determine the maximum depth at which NIR fluorescence targets could be distinguished. Further development will combine NIR and gamma modalities in a single imaging system.
This work was support by a Science and Technologies Facilities Council (STFC) grant – CLASP ST/M007820/1 – and fellowship – ST/R00501X/1. The authors would like to thank John Holt, Space Research Centre, University of Leicester for his assistance and support. Also, the authors would like to thank Dr John Pearl, Department of Respiratory Sciences, University of Leicester, for his advice and assistance. A.M. Almarhaby, M.S. Alqahtani and L.K. Jambi have been financially supported by the Ministry of Health, King Khalid University and King Saud University, Ministry of Education, Kingdom of Saudi Arabia.
Keywords: Gamma camera, fluorescence imaging, ICG, IRDye® 800CW, cancer surgery
The role of PML in Glioblastoma physiology (#19)
Maria Tampakaki1, 2, Mariam-Eleni Oraiopoulou1, Stylianos E. Psycharakis3, Eleftheria Tzamali1, Vangelis Sakkalis1, Giannis Zacharakis3, Joseph Papamatheakis4, 5
1 Foundation for Research and Technology-Hellas, Institute of Computer Science, Heraklion, Greece
The promyelocytic leukemia protein (PML) is a tumor suppressor with multiple cellular functions regarding the cell regulation . PML also participates in neural cell migration , , via the Polycomb Repressive Complex 2 (PRC2), , whose functional component is the Enhancer of Zeste Homolog 2 (EZH2). In Glioblastoma (GB), PML inhibits the tumor growth , while it induces migration . We used Optical Microscopy (OM) and Light Sheet Fluorescence Microscopy (LSFM) to study the U87MG GB cell line PML-related growth, cell death and invasion over time.
The U87MG GB cell line (U87MG-wt) was lentivirusly transfected with the PML isoform IV (U87MG-PML), which was fused with the fluorophore DsRed. Both proteins were expressed under doxycycline (DOXY) presence. 3D spheroids were generated via the hanging drop technique and cultured with and without an extracellular matrix-like substrate. Brightfield photo-micrographs were captured every 24h to monitor the growth and invasive pattern. The sensitivity to the DZNeP, an EZH2 histone methyltransferase inhibitor, was also tested both in 2D, using the MTT viability assay and in 3D, by estimating the growth elimination. LSFM imaging was used to visualize the PML expression, using the DsRed distribution and the cell death pattern, using the Draq7 nuclear probe.
The U87MG-PML cells exhibited significant differences compared to the U87MG-wt regarding their growth and invasive properties. The U87MG-PML spheroids were smaller in size, indicating different aggregative and proliferative capacity , . The invasive pattern was common in both cell lines adopting the typical starburst morphology , yet with altered migration dynamics. The spheroid core of the U87MG-PML spheroids was smaller than the U87MG-wt, though the invasive rim was the same. The LSFM scans showed a uniform distribution of the PML in both conditions, while the cell death pattern was not considerably affected (Fig1&2).
The DZNeP effect on tumor growth expansion was not significantly altered by the presence of PML. This is in line with previous findings, indicating that EZH2 is mainly involved in tumor migration in human malignancies , while PML contributes to tumor growth inhibition by prolonging the G1 phase and decelerating the cell cycle , , .
GB expansion is attributed to both excessive proliferation and local spreading. Our results are in line with previous findings , , ,  indicating that these two functions are modulated by distinct cellular mechanisms. Unravelling the role of PML in GB invasion could set PML as a therapeutic target aiming at eliminating multiple sub-clones depending on their proliferative and/or invasive phenotype within the heterogeneous GB tumor.
Authors would like to thank Evangelos Liapis and Eirini Gonianaki for all the help they provided. This work was supported by the project “BIOIMAGING-GR” (MIS5002755) which is implemented under the Action “Reinforcement of the Research and Innovation Infrastructure”, funded by the Operational Programme "Competitiveness, Entrepreneurship and Innovation" (NSRF 2014-2020) and co-financed by Greece and the European Union (European Regional Development Fund). Also supported by the project “KRIPIS ΙΙ-VITAD (MIS 5002469)”, as well as the Hellenic Foundation for Research and Innovation (HFRI) and the General Secretariat for Research and Technology (GSRT), under the HFRI PhD Fellowship grant (GA. no. 130178/I2/31-7-2017).
 R. H. Chen, Y. R. Lee, and W. C. Yuan, “The role of PML ubiquitination in human malignancies.,” J. Biomed. Sci., vol. 19, p. 81, 2012.
 V. Amodeo et al., “A PML/Slit Axis Controls Physiological Cell Migration and Cancer Invasion in the CNS,” Cell Rep., vol. 20, no. 2, pp. 411–426, 2017.
 T. Regad, C. Bellodi, P. Nicotera, and P. Salomoni, “The tumor suppressor Pml regulates cell fate in the developing neocortex,” Nat. Neurosci., vol. 12, no. 2, pp. 132–140, 2009.
 A. Iwanami et al., “PML mediates glioblastoma resistance to mammalian target of rapamycin (mTOR)-targeted therapies,” Proc. Natl. Acad. Sci., vol. 110, no. 11, pp. 4339–4344, 2013.
 C. Hadjimichael et al., “Promyelocytic Leukemia Protein Is an Essential Regulator of Stem Cell Pluripotency and Somatic Cell Reprogramming,” Stem Cell Reports, vol. 8, no. 5, pp. 1366–1378, 2017.
 M.-E. Oraiopoulou et al., “Integrating in vitro experiments with in silico approaches for Glioblastoma invasion: the role of cell-to-cell adhesion heterogeneity,” Sci. Rep., vol. 8, no. 1, p. 16200, 2018.
 D. Guan and H. Y. Kao, “The function, regulation and therapeutic implications of the tumor suppressor protein, PML,” Cell and Bioscience, vol. 5, no. 1. BioMed Central Ltd., 04-Nov-2015.
Figure 1. LSFM images of the growth condition for the U87MG-PML non-induced and induced spheroids
LSFM images of U87MG-PML non-induced (a) and induced (b) spheroids in the growth condition. LSFM brightfield scan is represented on the left side of each image. Draq 7 is excited at 635nm and illustrated with cyan, labeling the dead cell nuclei, while DsRed is excited at 488nm and illustrated with red, representing the PML expression. PML expression is homogenously distributed within the PML-induced spheroid, while the cell death pattern is not considerably affected. Scale bar is set at 100μm.
Figure 2. LSFM images of the invasive condition for the U87MG-PML non-induced and induced spheroids
LSFM images of U87MG-PML non-induced (a) and induced (b) spheroids in the invasive condition. Draq 7 is illustrated with cyan, labeling the dead cell nuclei and DsRed is illustrated with red, representing the PML expression. PML expression is homogenously distributed within the core of the PML-induced spheroid, while the cell death pattern is not considerably affected. The invasive pattern retains the radial starburst morphology in both conditions, however, the migration dynamics are altered regarding the invasive rim of the spheroid. Scale bar is set at 100μm.
Keywords: Glioblastoma, PML, Invasion, EZH2, DZNeP.
Physiological description of patient-derived Glioblastoma cells using fluorescence imaging (#20)
Mariam-Eleni Oraiopoulou1, Eleftheria Tzamali1, Stylianos Psycharakis2, Eleftheria Parasyraki3, 4, George Tzedakis1, Antonis F. Vakis5, 6, Vangelis Sakkalis1, Joseph Papamatheakis3, 4, Giannis Zacharakis2
1 FORTH, Institute of Computer Science, Heraklion, Greece
Glioblastoma (GB) is the most malignant brain cancer and is not considered a curable disease so far. A multidisciplinary framework that integrates basic and translational research is presented attempting a better understanding of its pathophysiology. In this attempt, a carefully planned combination of experimental approaches were mobilized. Specifically, patient-specific cell cultures were established and used in experimental assays, while Light Sheet Fluorescence Microscopy (LSFM) and Confocal Microscopy (CM) were used to visualize the GB pathophysiologic factors.
Tissue from naïve-from-treatment patients was excised during brain biopsy and/or resection. After the GB case confirmation, if needed, part of the tissue was transplanted to immunodeficient mice. Followingly, the primary cell cultures were established for each GB case.
The primary cell cultures were phenotypically characterized and used in experimental assays. The well-described U87MG and T98G GB cells served as control. This work primarily focuses on proliferation and invasion, two of the most dominant GB characteristics. LSFM and CM imaging were used to monitor the growth of small-sized avascular GB spheroids using optimized imaging protocols, so that GB-specific biomarkers could be identified. In addition, preclinical drug screening was used to evaluate the efficacy of specific drugs.
Focusing on proliferation, we supported that the intratumoral heterogeneity together with the overall proliferation reflected in both the proliferation rate and the mechanical cell contact inhibition, but not the cell size, can predict the evolution of different GB cell lines .
We showed that the primary GB spheroids adopt a novel, cohesive pattern mimicking perivascular brain invasion, while the U87MG and the T98G adopt a typical, starburst, invasive pattern . CM indicated alternative proliferative and adhesive characteristics of the invading cells.
We also set a preclinical drug screening tool to assess the distribution, penetration and cytotoxic potency of the Temozolomide (TMZ) and Doxorubicin (DOX) antineoplastic agents. We used LSFM in order to discriminate growth inhibition in cell division arrest from cell death. The effective doses varied over four orders of magnitude. Unlike TMZ which showed slight growth-inhibiting effects, DOX was able to accumulatively cause necrosis.
Overall, we claim that future research should be based on patient-derived GB cells and that common cell lines should only serve as landmarks to unite studies of different groups. For every primary established cell line, not only molecular, but also physiological parameters should be estimated to enable a more precise future clustering of different GB cases in order to optimize individualized therapy decisions and GB pathophysiology understanding.
Authors would like to thank Evangelos Liapis for all the help he provided, Elias Drakos for his collaboration, and Katerina Manolitsi for her advisory comments, as well as Despina Tsoukatou and Venediktos Makatounakis for the expert technical assistance. This work was supported by the project “BIOIMAGING-GR” (MIS5002755) which is implemented under the Action “Reinforcement of the Research and Innovation Infrastructure”, funded by the Operational Programme "Competitiveness, Entrepreneurship and Innovation" (NSRF 2014-2020) and co-financed by Greece and the European Union (European Regional Development Fund). Mariam-Eleni Oraiopoulou wishes to acknowledge support from the Hellenic Foundation for Research and Innovation (HFRI) and the General Secretariat for Research and Technology (GSRT), under the HFRI PhD Fellowship grant (GA. no. 130178/I2/31-7-2017) and travel support from Molecubes.
1. Oraiopoulou, M.-E., et al., In Vitro/In Silico Study on the Role of Doubling Time Heterogeneity among Primary Glioblastoma Cell Lines. BioMed Research International, 2017. 2017: p. 12.
2. Oraiopoulou, M.E., et al., Integrating in vitro experiments with in silico approaches for Glioblastoma invasion: the role of cell-to-cell adhesion heterogeneity. Scientific Reports, 2018. 8(1): p. 16200.
An invading T98G GB spheroid. Confocal Microscopy.
A.Confocal image of a T98G invasive spheroid after 24h. B. Magnified illustrated invasive region.
GFP expression (green) is indicative of the cellular protrusions and/or cytoskeleton rearrangements formed during invasion. High PKH26 (red) signal intensity indicates low proliferative activity. Nuclei are shown in blue. White arrows depict dividing cells. Scalebar is set at 100 microns.
Cell death of the live LSFM-scanned primary GB spheroids.
The intrinsic cytotoxicity of a representative untreated spheroid is depicted (upper row). A representative TMZ-treated spheroid cell death pattern is also shown (middle row), as well as a DOX-treated one (lower row). DOX is also marked with green. The max intensity z stacks are depicted for two time points. In day 11, the control and the TMZ-treated spheroids are so large that the center of the specimen could not be properly scanned reaching the penetration scanning depth limit of the modality. Scale bar is set at 100 microns.
Keywords: Glioblastoma, patient-derived cell cultures, LSFM, confocal microscopy, translational research
Applications of nonlinear microscopy in breast tissue samples diagnosis. (#13)
Vassilis Tsafas1, 3, Evangelia Gavgiotaki2, 3, George Filippidis3
1 University of Crete, Department of Physics, Heraklion, Greece
Nonlinear imaging microscopy modalities comprise powerful tools for biomedical diagnosis. These technologies provide several advantages compared to other optical techniques such as high resolution, label-free imaging capabilities, increased penetration depths and intrinsic three dimensional sectioning without any phototoxicity phenomena and energy deposition onto the biological sample . The investigation of both cells and tissues during inflammation and cancer was achieved via the application of these non destructive techniques [2,3].
The experimental apparatus of our nonlinear imaging system consisted of a femtosecond laser (1068 nm, 80 MHz, 1W, 400 fs) and a modified upright Nikon microscope. A high numerical aperture objective lens is used for the tight focusing of the beam and laser scanning procedure is performed with a pair of galvanometric mirrors. Finally, nonlinear signals are collected simultaneously from two PMTs in reflection and transmission mode.
Figure 1 depicts a nonlinear multimodal image of human breast tissue by simultaneously detection of two signals. SHG signal (in green) is depicting collagen organization in tissue and is collected in backward detection. THG signal (in blue) is generated through optical inhomogeneities of the tissue and is recorded in forward direction. Previews studies indicated that THG can demonstrate the density and size of cells and could be used as a diagnostic tool for cancer prognosis in malignant tissues . Moreover, several research studies indicated that collagen's organization and integrity are changing during malignancy . Our goal is to explore collagen distribution and directionality in human breast tissues during cancer. Figure 2 presents the differences in collagen between benign and cancerous tissues. Οur target is these morphologically differences of collagen between the tissues to be expressed quantitatively for the development of an accurate objective diagnostic tool.
We anticipate that the quantification of nonlinear signals provides complementary reliable biological criteria that will be potentially proven useful as diagnostic tool to understand serious diseases such as cancer.
V. Tsafas acknowledge General Secretariat for Research and Technology (GSRT) and Hellenic Foundation for Research and Innovation (HRFI) for the financial support.
 Adur, Javier, et al. "Nonlinear microscopy techniques: principles and biomedical applications." Microscopy and Analysis. IntechOpen, 2016.
 Gavgiotaki, Evangelia, et al. "Distinction between breast cancer cell subtypes using third harmonic generation microscopy." Journal of biophotonics 10.9 (2017): 1152-1162.
 Kuzmin, N. V., et al. "Third harmonic generation imaging for fast, label-free pathology of human brain tumors." Biomedical optics express 7.5 (2016): 1889-1904.
 Gavgiotaki, Evangelia, et al. "Non-linear microscopy differentiates normal from pathological breast tissue." Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series. Vol. 10685. 2018.
 Ambekar, Raghu, et al. "Quantifying collagen structure in breast biopsies using second-harmonic generation imaging." Biomedical optics express 3.9 (2012): 2021-2035.
Keywords: Second Harmonic Generation, Third Harmonic Generation, Breast Tissue, Collagen, Diagnosis
Ultrafast adaptive beamformer based on angular coherence of plane wave compounding (#24)
Hanna Bendjador1, Thomas Deffieux1, Mickaël Tanter1
1 Physics for Medicine Paris , Inserm, ESPCI Paris, CNRS, PSL University, Paris, France
Ultrafast plane wave compounding was introduced to form high quality images at very high frame rates for exacting applications such as functional neuroimaging or quantitative ultrasound. Though, propagation through complex layers induce strong aberrations of the acoustic wave front; damaging both image quality and quantitative information assessment. In this work, we investigated a new and fast method to maximize the coherence between steered plane waves - major estimate of the focusing and image quality1,2. Essential for delays compensation3, the phase aberration profile is also computed.
To ensure a maximal amount of information, 100 plane waves were transmitted on in vitro media using a 192-element linear probe (pitch 200 µm, 6.25 MHz central frequency). To confirm our method, both numerical aberration - virtually introduced at emission - and a physical aberrating lens were used on a phantom, and on in vivo human liver data. In the case of thick aberrations, correction was further improved and performed on isoplanetic patches, giving one phase profile per patch.
Comparisons of aberrated and corrected images showed a strong diminution in clutter: contrast on in vitro anechoic compartment was improved by 11.7dB for virtual aberration, and 8.5dB for physical aberration. Also, the lateral resolution went respectively from 1.00 to 0.88 mm, and from 2.1 to 1.3 mm after correction. Aberration laws are very consistent (r2 = 99% and 95%) with expected delays. Those results demonstrate that our method can perform simultaneously efficient correction of the image, increasing both CNR and resolution, and aberration phase law extraction.
For the first time, an image correction method combines phase aberration correction and coherence optimization. Importantly, its implementation time on a high end PC computer could already be decreased to some tens of milliseconds, which opens the path to real-time adaptive beamforming of ultrafast data.
Figure 1: Adaptive beamforming results on in vitro acquisition
① Introduction of a numerical aberration law at emission. ② Propagation through an aberrating lens.
a. Phantom images of pins and anechoic cyst. b. Images after aberration c. Images after correction with our method. d. Amplitude plot showing pin resolution.
Keywords: Aberration correction, resolution, contrast, ultrafast imaging
Volumetric ultrafast Ultrasound Localisation Microscopy in vivo (#26)
Baptiste Heiles1, Vincent Hingot1, Line Rahal1, Pauline Lopez2, Claire Rabut1, Antoine Bergel1, Mathieu Pernot1, Mickael Tanter1, Olivier Couture1
1 PhysMed Paris, CNRS, INSERM, ESPCI Paris, PSL Research University, Paris, France
Ultrafast Ultrasound Localization Microscopy (ULM) allows the reconstruction of a full density map of the rat brain vasculature with 8 µm resolution[1,2]. Thanks to ultrafast capabilities, this technique also allows to measure blood flow with micrometric resolution. Despite additional successes for tumour imaging , this 2D technique suffers from out of plane microbubbles, loss of information due to projection of 3D vascular structure into a 2D image and minute long acquisitions. We present here the development of volumetric ULM using a 2D matrix array driven with a 4D echograph system.
The 2D matrix array consists of 32x32 independent transducers arranged in a plane and controlled by a customized 1024 channel system. Sprague Dawley rats were anesthesised with Ketamin-Domitor. The probe was driven at 9MHz to transmit plane waves at 12 different angles to achieve a compounded volume rate of up to 500Hz and a single sequence of 185 volumes was repeated 540 times. A 5 steps workflow was developed to perform filtering, localization, tracking, velocimetry and rendering. Localizing microbubbles was done without using conventional processes involving interpolation to avoid heavy calculations and long computation times. 3 steps were added to correct for motion by using rigid or affine registration.
Applying the workflow on 99000 volumes, a volumetric rendering of the rat’s brain vasculature (fig.1) with precision much smaller than the wavelength (i.e. 250µm). The maximum theoretical localization precision obtainable was calculated to be less than 1μm in the axial and around 5 μm in the lateral and elevational directions. For motion correction, two approaches were developed. The first one where only the sequences were corrected in between each other. Another approach consists of correcting volume to volume within a sequence and then the microbubbles position based on sequence-to-sequence motion. The latter yields better image quality without impairing contrast but is considerably heavier and longer in terms of calculations. In the case of a well stereotactised rat, such an approach can be dispensed with but will be necessary for human volumetric ULM, or other organs in rodent. Velocimetry was performed to yield a large range ([1µm/s:10cm/s]) of velocity fields in vessels.
Volumetric ULM was successfully demonstrated in vivo. A dedicated 5-8 step workflow was implemented to take into account different configurations. It was also demonstrated that 3D-velocimetry was possible. This 3D technique allows to surpass the conventional resolution in a large volume with only 200s of acquisition time. However, challenges remain such as the limited sensitivity of the 2D arrays, the data size, computation and transfer times.
 Errico et al, Nature, 2015
 Couture et al, IEEE IUFFC 2018
 Lin et al, Theranostics, 2017
Keywords: ultrasound, ultrasound localization microscopy, 4D
Ultrahigh spatial and temporal resolution fMRI with implanted CMOS-based planar microcoil at 14.1T (#27)
Marlon Pérez-Rodas1, Jonas Handwerker3, Michael Beyerlein1, Xin Yu1, Rolf Pohmann1, Jens Anders2, 3, Klaus Scheffler1, 4
1 Max Planck Institute for Biological Cybernetics, High Field Magnetic Resonance Center, Tübingen, Germany
Spatial and temporal resolution are limited by the SNR and the time required for spatial encoding. To maximize sensitivity and localization, we implanted a miniaturized coil into the brains of healthy rats, consisting of a fully-integrated CMOS1 NMR transceiver, including RF-amplifier, preamplifiers, signal conditioning electronics and microcoil on a needle-shaped PCB, thus maximizing sensitivity and reducing signal loss. This system samples the MR signal within a very confined spatial region (~10 nl) at a temporal resolution of microseconds, without the use of gradients for spatial encoding.
The fully-integrated NMR transceiver1,2 includes an on-chip broadband MR-microcoil (300 µm-diameter double-spiral coil) on a silicon substrate with a width of 450 µm, a length of 3000 µm, and a triangular tip for easier insertion into the brain tissue, with a sensitive volume of around 10 nl, directly bonded to a small supporting PCB (Fig 1). The acquired signals are amplified on a signal-conditioning PCB close to the microcoil and further processed outside the scanner room3.
The microcoil was implanted 1.5 mm to 2 mm into the somatosensory cortex of anesthetized rats and used to observe localized FIDs during rest and forepaw stimulation with repetition times between 5 ms and 1 s in a 14.1 T/26 cm horizontal magnet. For control, standard EPI measurements were performed using a surface coil.
Fig. 2 shows a comparison of BOLD responses acquired with the conventional surface coil and the microcoil at 1000 ms and 5 ms temporal resolution, respectively. The microcoil response is obtained by averaging over all stimulation epochs, extracting the area under the magnitude of the resulting FIDs, yielding a single value per FID with temporal resolutions of up to 200Hz (TR=5ms). Additionally, the functional signals were low-pass filtered with a 3Hz Gaussian filter to reduce the noise, since no visible stimulation-related features beyond that frequency were observed. Signal changes of around 2% were observed during forepaw stimulation. The contralateral responses for stimulation of the right paw showed no response in any measurement, indicating that the signals are indeed the hemodynamic response to the stimulation of the left paw. By fitting an exponential decay to the FIDs, contributions from inflow and BOLD can be distinguished as changes in amplitude and T2*-decay of the signals.
Exceptional spatial and temporal resolution was possible with the microcoil, detecting BOLD and flow-related signal changes during electrical stimulation within 10 nl of volume and 5 ms of temporal resolution. This new technology offers the potential to detect novel effects or MR-fingerprints of neuronal activation and can be combined with other local and fast methods for neuronal recording such electrophysiology orcalcium recording,
Top: Anatomical images with micoroil and fMRI activation.
Middle: fMRI and microcoil timecourses for a 6 s forepaw stimulation of the left and right paw. Right paw stimulation is shown as control.
Bottom: Activation signal splitted by exponential fit to the single FIDs into contribution from inflow (signal amplitude) and BOLD (signal decay / T2*)
Keywords: microcoil, fMRI, ultrahigh resolution
Simultaneous measurement of BOLD-fMRI and optical imaging of intrinsic signals for determination of oxygenation and CBV at high spatial and temporal resolution (#9)
Rolf Pohmann1, Rebekka Bernard1, Klaus Scheffler1, 2
1 Max Planck Institute for Biological Cybernetics, Tübingen, Germany
To be able to better understand the formation and the dynamics of the BOLD effect, simultaneous measurements of the different components contributing to BOLD are required. We combined optical imaging of intrinsic signals, which is able to quantitatively measure oxygenation and CBV changes at high spatial and temporal resolution, with ultra high field fMRI.
A fully magnetic field compatible optical imaging setup for combined optical and MR imaging of rats was developed, with the goal of improving our understanding of the BOLD effect and contributing to a better analysis of fMRI studies.
Optical imaging of intrinsic signals takes advantage of the different absorption coefficients of oxygenated and deoxygenated hemoglobin. To obtain optical imaging and fMRI data with similar quality and resolution as in separate experiments, a tailor-made, magnetic field proof high-sensitivity camera was combined with commercial optical components. The skull of anaesthetized rats was thinned to translucency and illuminated by light with four different wavelengths for quantitative determination of oxygenation and CBV.
Brain activation was applied under electrical forepaw stimulation with varying parameters and observed with fMRI and OI (50 ms per image, four interleaved wavelengths).
The optical images were used to determine oxygenation and CBV using published pathlength data (1).
OI and fMRI images with excellent quality were acquired. Venous structures could be identified clearly in both modalities and were used to overlay the different parameter images. Fig. 2 shows timecourses for varying stimulation durations between 1 s and 5 s. BOLD, oxygenation and CBV timecourses agree well.
Concurrent MRI and OI can potentially help to improve our understanding of the BOLD effect and the hemodynamic processes that contribute to the fMRI signal. A previous approach used an endoscope to get the light out of the magnet to a conventional camera (2). By using a specially designed camera, which combines high sensitivity with full compatibility with the magnetic field, and high-performance optical components, we are able to obtain images from inside the magnet without losses in quality compared to those acquired in standard OI experiments. The combination with ultra-high field MRI ensures also a high fMRI image quality and spatial and temporal resolution.
Combining intrinsic optical imaging with fMRI can help to disentangle the different factors contributing to the BOLD effect and yield important information to enhance both OI and fMRI experiments. Furthermore, the developed setup can be combined with other optical techniques to measure further neuroscientific parameters at the same time, like Ca, speckle-flow or voltage sensitive die imaging.
1. Ma, Y., et al. (2016). "Wide-field optical mapping of neural activity and brain haemodynamics: considerations and novel approaches." Philos Trans R Soc Lond B Biol Sci 371(1705).
2. Kennerley, A. J., et al. (2005). "Concurrent fMRI and optical measures for the investigation of the hemodynamic response function." Magnetic Resonance in Medicine 54(2): 354-365.
Setup for optical imaging with fMRI
The optical imaging camera and optical components are placed in the 12 cm bore of the 14.1 T scanner. The brain is observed via a prism and a set of high-performance objectives, from which all metallic parts were carefully removed, by a magnetic field proof scientific CMOS camera. The illumination in four different wavelengths is transmitted into the scanner by optical fibers. A small surface coil right above the somatosensory cortex is used for MRI transmission and signal reception.
Keywords: fMRI, optical imaging, BOLD
Novel alpha-180 SE based LINE-scanning method (SELINE) for laminar-specific fMRI (#11)
Sangcheon Choi1, 2, Hang Zeng1, 2, Rolf Pohmann1, Klaus Scheffler1, 3, Xin Yu1
1 Max Planck Institute for Biological Cybernetics, Department of High-field Magnetic Resonance, Tuebingen, Germany
Laminar-specific functional magnetic resonance imaging (fMRI) has been successfully applied to understand neuronal circuitry across the cortical layer [1-3]. FLASH based line-scanning method was proposed with high temporal (50 ms) and spatial (50 µm) resolution to better characterize the fMRI onset time by combining 2 saturation RF pulses . However, the imperfect RF saturation performance led to poor boundary definition of the ROI from the cortex. Here, we propose an α-180˚SE based line-scanning (SELINE) method to solve this problem.
Line-scanning fMRI data was acquired in an anesthetized rat at Bruker 14.1T scanner using in-house transceiver surface coil with 6mm diameter. The conventional FLASH based and the proposed α-180 SE based line-scanning (SELINE) pulse sequence are used (Fig. 1c, Fig. 1d). To acquire the SELINE data, the 180˚RF pulse oriented perpendicular to the α RF pulse as moving the refocusing gradient to phase encoding gradient (Gp) in order to obtain high spatial resolution without reduced FOV aliasing problem. Functional activation was identified by performing a left forepaw stimulation task (3Hz, 4s, pulse width 300us, 2.5mA), followed by 1 second pre-stimulation, 4 second during electrical stimulation and 15 seconds post stimulation with a total 20 seconds for 10 min 40 sec.
To demonstrate the saturation performance, the line profiles of MRI signal intensity are shown in Fig. 1b. The proposed SELINE method has 2.1% mean suppressed brain region with 4 trials while the conventional FLASH based method has 26.9% (12.8 times of SELINE). Averaged time course and percentage change map have higher contrast-to-noise ratio (CNR) in the deeper layer of the SELINE method than FLASH-based method. Laminar-specific correlation coefficient matrices are shown in Fig 2a, 2b (2 times of SELINE). Layers L2/3-L4 (0.15 – 0.8 mm) show strong correlation with the SELINE method. tSNR plots with 3 trial are shown in Fig. 2c, 2d. The conventional FLASH-based method has 30.6 mean tSNR with slow slope in Fig. 3c (TR, 100ms, TE, 12.5ms). On the other hand, SELINE method has 50.4 mean tSNR with steep slope in Fig. 3d. The altered laminar SNR profile is due to the less homogeneous B1 field from the small transceiver surface coil.
We demonstrate the feasibility of the SELINE method for laminar-specific fMRI. The proposed scheme solves the contamination issue out of ROI. Future work will be done to improve this method. To achieve higher temporal resolution up to 100 or 200 ms by optimizing flip angle α, TR and sequence parameters, we will investigate the effect from stimulated echoes during fMRI stimulation paradigm.
 Polimeni JR, Fischl B, Greve DN, Wald LL. Laminar analysis of 7T BOLD using an imposed spatial activation pattern in human V1. Neuroimage 2010; 52:1334-1346
 Turner R. Uses, misuses, Uses, misuses, new uses and fundamental limitations of magnetic resonance imaging in cognitive science. Philos Trans R Soc Lond B 2016; 371:20150349
 Huber L, Handwerker DA, Jangraw DC, Chen G, Hall A, Stüber C, Gonzalez-Castillo J, Ivanov D, Marrett S, Guidi M, Goense J, Poser BA, Bandettini PA. High-Resolution CBV-fMRI Allows Mapping of Laminar Activity and Connectivity of Cortical Input and Output in Human M1. Neuron 2017; 96: 1253-1263
 Yu X, Qian C, Chen D, Dodd SJ, Koretsky AP. Deciphering laminar-specific neural inputs with line-scanning fMRI. Nature methods 2014; 11:55-58
FLASH based and proposed α-180 SE based line-scanning (SELINE) fMRI representation.
Fig 2. a/b) the fMRI spatiotemporal profiles of the evoked (upper panel), averaged time course from individual voxels, fMRI percentage map and voxel-by-voxel correlation coefficients as the function of time 1s off, 4s on, 15s off (lower panel), c/d) the temporal SNR (tSNR) and mean tSNR with 3 trials.
Keywords: Laminar specific fMRI, SE based line-scanning method, High spatial resolution
Combined photoacoustic and fluorescence label-free microscopy for the ex vivo investigation of pigmented ocular lesions (#6)
Konstantinos G. Mavrakis1, 3, George J. Tserevelakis1, Eleni Karamouzi1, Danai Pantazopoulou1, Efstathios Detorakis2, Giannis Zacharakis1
1 FORTH Hellas, Institute of Electronic Structure, HERAKLION, Greece
We demonstrate the application of an extended field of view microscope, combining photoacoustic (PA) and fluorescence label-free contrast modalities, for the ex vivo investigation of surgical biopsies removed from human eyes. Τhe biopsy samples presented a remarkable spatial overlap of the two signals, indicating a correlation between them. The bimodal microscopy approach presented in this work [1,2], could be employed for the differentiation between benign and malignant intraocular tumors of the uvea in surgical biopsies, simplifying the relevant procedures for this purpose.
The custom hybrid microscope consists of two optical paths, a 450nm CW laser beam employed by fluorescence and nanosecond pulses at 532nm for the PA modality. Both beams are focused by the same objective lens on the rear side of a water tank where the specimens are placed. The tank is attached on a high precision 3D translational motion system which permits raster scanning. Back-scattered fluorescence is guided to a photomultiplier tube while photoacoustic waves are detected by a focused ultrasonic transducer.
The biopsies have been preserved in paraffin which was removed by heating and baths in xylene. Hydration was performed with baths of ethanol before rinsing the sections in distilled water. All procedures follow the guidelines and are approved by the Institutional Ethics Committee.
The imaging potential of the hybrid microscope was tested on choroidal biopsy specimens. High autofluorescence signals were present in all biopsies, benign and malignant. Nevertheless, parts of nevi biopsies are characterized by increased autofluorescence levels compared to the surrounding regions. This observation is validated by PA microscopy detecting the melanin distribution at the same region, indicating a positive correlation between PA amplitude and average autofluorescence intensity. Interestingly, melanomas exhibited areas of low autofluorescence which corresponded to high PA signals. For melanomas, an increase in melanin and thus increased PA signals, coincided with low autofluorescence indicating a negative correlation between PA amplitude and average autofluorescence intensity. As a result, the correlation between the signals could be used as a criterion for differentiation between malignant and benign tumors.
In this work, we have presented the application of a prototype combined optical PA label-free microscopy system for the detailed imaging of several ocular tissues. We have tested the hybrid imaging contrast on surgical biopsies with highly promising results that could be potentially exploited in the differentiation between benign and malignant intraocular tumors of the uvea, simplifying the relevant procedures for this purpose.
This work was supported by Stavros Niarchos Foundation [ARCHERS]; POLITEIA II [MIS 5002478]; H2020 Laserlab Europe [EC-GA 654148]; EU Community’s H2020 IPERION CH Project [GA n. 654028]; projects “BIOIMAGING-GR” (MIS5002755) and HELLAS CH [MIS 5002735] which are implemented under the Action “Reinforcement of the Research and Innovation Infrastructure”, funded by the Operational Programme "Competitiveness, Entrepreneurship and Innovation" (NSRF 2014-2020) and co-financed by Greece and the European Union (European Regional Development Fund).
 Tserevelakis G.J., Tsagkaraki M., Zacharakis G., "Hybrid photoacoustic and optical imaging of pigments in vegetative tissues," J. Microsc., 263 (3), 300-306 (2016).
 Tserevelakis, G.J., Avtzi, S., Tsilimbaris, M.K., Zacharakis, G., “Delineating the anatomy of the ciliary body using hybrid optical and photoacoustic imaging,” J. Biomed. Opt., 22 (6), art. no. 060501, (2017).
Keywords: choroidal melanoma, biopsy, label-free, photoacoustic microscopy, multimodal
Super-resolved 3D photoacoustic imaging (#8)
Guillaume Godefroy1, Sergey Vilov1, Bastien Arnal1, Emmanuel Bossy1
1 Université Grenoble Alpes, LIPhy, Grenoble, France
The reliability and accuracy of information provided by acoustic-resolution photoacoustic (PA) imaging depend on the resolution of the system, limited by acoustic diffraction.
A number of methods have been developed in order to overcome this limitation, including l1 -based regularization  and prior knowledge of the system's response at each point of the imaging zone. This approach can be applied to a wide class of objects considered to be sparse. Here we demonstrate the possibility to obtain 3D super resolved images with a custom 2D sparse array.
PA data was obtained from 2 nylon threads spaced by 100 µm illuminated by a 5-ns laser pulse (740 nm). The resulting PA signals were acquired by 256 elements of a custom spherical array (central frequency around 8 MHz) connected to multichannel acquisition electronic. An additional acquisition was performed to get a single-point source system's response by using a sample containing a single 10-µm diameter microbead embedded in an agarose gel. This calibration data were used to construct a propagation matrix containing the system's response at each point of the discretized imaging plane. The reconstruction images were obtained by minimizing a l1 -based regularization functional (such as used in ) with the FISTA algorithm  .
A 2D cross sectional reconstruction image of the 3D sample is represented in Fig 1. The image is filtered with a 2D Gaussian filter. The applied l1 -based reconstruction provides an image in which the 2 nylon threads spaced by 100 µm are distinguishable, whereas the conventional delay-and-sum beamforming is unable to separate them. The maximal resolution of the system, calculated on the beamformed image of a single point source, is about 175 µm in x and y direction: we here therefore break this theoretical limit for a 3D sample.
The quality of this image is influenced by several parameters, notably the regularization parameter which must be chosen cautiously. A general limitation when applying model based reconstruction method to 3D data is the resulting computational time that can be quite extensive, as it scales with the dimension of the reconstruction.
It has been demonstrated that l1 -based regularization can provide super-resolution in 3D PA imaging. The resolution obtained with our system so far reaches 100 µm, whereas the theoretical diffraction limit is about 175 µm. The next step of the project will be to apply this method on in-vivo data in order to distinguish neighboring blood vessels.
Figure 1: Classical beamformed image (a) and super-resolved image ( b)
b: Super-resolved image filtered with a 2D Gaussian filter: l1-based reconstruction recovers two distinct regions corresponding to the channels.
Keywords: Photoacoustics, sparse array, super-resolution, 3D reconstruction
P141Optical resolution photoacoustic microscopy for the study of craniosynostosis in mouse models (#5)
George J. Tserevelakis1, Konstantinos Makris2, Stella Amarioti1, George Mavrothalassitis2, Giannis Zacharakis1
1 Foundation for Research and Technology - Hellas (FORTH), Institute of Electronic Structure and Laser, Heraklion, Greece
Craniosynostosis is a serious pathological condition occurring in 1 every 2000 births, characterized by a premature fusion of infants’ cranial sutures, changing, in this manner, the growth pattern of the skull . This paper demonstrates the detailed investigation of craniosynostosis in mouse skulls, by means of an optical resolution photoacoustic microscopy approach. The promising potential of the employed imaging technique is highlighted through the extraction of quantitative data following the post processing of the recorded tomographic images from different samples.
The system  is built around a purposefully modified inverted optical microscope serving as a platform for the photoacoustic microscopy setup. A variable repetition rate diode pumped ns laser is employed for efficient signal excitation (energy per pulse: 29.4 μJ, pulse width: ~10 ns, selected repetition rate: 5 kHz) emitting infrared radiation at 1064 nm and 532nm after frequency doubling. The dyed skulls (Alizarin Red S, for bone staining and Alcian Blue for cartilage staining) are raster scanned over the focused beam via XY motorized stages and the generated acoustic waves are detected in transmission mode by a 20 MHz spherically focused ultrasonic transducer.
The generated three-dimensional reconstructions permitted the determination of skulls’ geometrical features such as the local radii of curvature, the μm precision measurement of several anatomical distances, as well as, the depth variability of cranial sutures in different specimens. Statistically significant differences between groups of healthy, treated and untreated animals were found as regards to the investigated parameters.
Optical resolution photoacoustic microscopy was proven to be a powerful diagnostic tool for the study of craniosynostosis in mouse models. Features such as the increased spatial resolution and the superior imaging contrast, render the proposed approach highly competitive when compared to state of the art techniques such as X-ray microtomography.
This work was supported by Stavros Niarchos Foundation [ARCHERS]; POLITEIA II [MIS 5002478]; H2020 Laserlab Europe [EC-GA 654148]; EU Community’s H2020 IPERION CH Project [GA n. 654028]; projects “BIOIMAGING-GR” (MIS5002755) and HELLAS CH [MIS 5002735] which are implemented under the Action “Reinforcement of the Research and Innovation Infrastructure”, funded by the Operational Programme "Competitiveness, Entrepreneurship and Innovation" (NSRF 2014-2020) and co-financed by Greece and the European Union (European Regional Development Fund). GJT acknowledges financial support from iThera Medical.
Qualitative comparison between control and pathological skulls by means of photoacoustic microscopy
Maximum amplitude projection photoacoustic microscopy images of a control (left) and pathological (right) mouse skull. The skull sutures are clearly visible in the case of the control specimen in contrast to the pathological skull with craniosynostosis.
Keywords: photoacoustic, microscopy, craniosynostosis
Two Path Optoacoustic Spectroscope for Non-invasive In-vivo sensing of metabolites in the SWIR (#25)
M. Mehdi Seyedebrahimi1, 2, Miguel A. Pleitez2, Pouyan Mohajerani2, Vasilis Ntziachristos2, 1
1 Technische Universitaet Muenchen, Computer and Electronic Eng. , Munich, Bavaria, Germany
Optoacoustic spectroscopy (OAS) offers a promising alternative to purely optical methods for non-invasive sensing. OAS potentially offers higher specificity than optical methods, since acoustic scattering is 2-3 orders of magnitude weaker than light scattering . Furthermore, earlier OA studies showed that metabolites, such as glucose, can be detected in-vivo .
We developed a novel near-infrared two-path OAS that could sense OA intensity changes due to metabolite concentration changes in-vivo (Fig. 1a). The main aim of dividing the optical path in two is to simultaneously measure two optoacoustic signals, one for sample measurements and the other for serving as an optoacoustic reference to 1) perform real-time correction of the laser emission profile of the laser source at different wavelengths and, 2) perform pulse to pulse correction to remove laser beam fluctuation and instability. In this study the OA intensity is measured by transforming OA raw signal (RS), using Hilbert transformation, and integrating the area of interest under the envelope curve, as also previously suggested .
Scans can be performed at the different spectral range, but at a first implementation, we were particularly interested in the spectral range between 1350 nm to 2000 nm with 10 nm step size. Averages of fifty measurements at each wavelength are typically recorded simultaneously at the two measurement paths. Fig. 1b shows OA measurements of distilled water at 1920 nm, before and after the pulse to pulse correction. It can be seen that the variation of OA intensity decreases significantly, which leads to higher signal to noise ratio (SNR). Fig. 1c shows SNR of water signal in the wavelength range 1350–2000 nm, before and after pulse to pulse correction. It can be seen that SNR improvement is significant not only at water spectra peaks at 1450 and 1920 nm but also at all other wavelengths.
In conclusion, we have developed a highly performing spectrometer for biomolecular measurements in the short wavelength. The system is now ready for broad applications in biomedical measurements and will significantly to our understanding of sensitivity and specificity of detecting various metabolites in the SWIR.
 Zhang, Hao F., et al. "Functional photoacoustic microscopy for high-resolution and noninvasive in vivo imaging." Nature biotechnology 24.7 (2006): 848.
 MacKenzie, Hugh A., et al. "Advances in photoacoustic noninvasive glucose testing." Clinical chemistry 45.9 (1999): 1587-1595.
 Ghazaryan, Ara, Saak V. Ovsepian, and Vasilis Ntziachristos. "Extended Near-Infrared Optoacoustic Spectrometry for Sensing Physiological Concentrations of Glucose." Frontiers in endocrinology 9 (2018): 112.
Schematic representation of NiR-TAOS and SNR improvement.
HWP – Half wave plate; PL – Polarizer, USD– ultrasound transducer; BS – beam splitter; AMP – Amplifier; PC – personal computer; DAQ – data acquisition card; L – lenses, M – Mirrors, b) pulse to pulse correction of water signal at 1920 nm, c) SNR improvement of water spectra at 1350 nm -2000 nm.
Keywords: Optoacoustic spectroscopy, Sensitivity improvement, Sensing metabolites
4D functional ultrasound imaging of the whole brain activity : first evidence in rodents (#15)
Claire Rabut1, Victor Finel1, Mafalda Correia1, Mathieu Pernot1, Sophie Pezet1, Thomas Deffieux1, Mickaël Tanter1
1 Physics for Medicine, Inserm U1273, ESPCI Paris, PSL Research University, CNRS FRE 2031, PARIS, France
Functional Ultrasound (fUS) Imaging is a recent powerful imaging technique to image whole-brain activation1. It has been applied to study functional response with high-resolution or Functional Connectivity (FC) in freely-moving or anesthetized conditions in different species. However, brain mechanisms are inherently 3D, it thus is crucial to develop a technology suitable to image the full brain in 3D with high sensitivity.
In this work, we have extended Multiplane Wave imaging2 to 3D Doppler (3D MWi) to allow for high sensitive 4D Ultrafast imaging by virtually increasing the emission signal amplitude without compromising the frame rate. Applied to functional neuro imaging, we can image transient changes in blood volume in the whole brain at high spatiotemporal resolution.
To highlight the spatiotemporal sensitivity of 3D fUS, we acquired the dynamics of blood volume in response to successive periodic visual stimuli and epileptiform activity, induced by a focal injection of a potassium channel blocker (4-AP) in trepanned, anesthetized rats. We also studied the FC at rest of both superficial and deep brain structures during several minutes.
High-quality and real time 3D vascular volumes (170µm3 voxel and 2s temporal resol) are obtained in rats and show the feasibility of task-activated 4D fUS. Strong correlations are observed between stimuli and vascular responses in dedicated brain areas (fig.a,b). During epileptiform seizures, we observe the 3D propagation of different waves of activity (fig c). At rest, we identified strong contrasting spatial coherence signals in low-frequency (<0.1 Hz) spontaneous 3D fUS signal fluctuations.
The ability of 4D fUS to image volumic cerebral activity at high spatiotemporal resolution, with high sensitivity is of great interest for whole-brain neuro-imaging applications.
1 : Mace&Al. Functional ultrasound imaging of the brain. 2011, 10.1038 Nat.Meth.1641
Figure: 4D Functional Ultrasound in rat
a) 3D activation map obtained when stimulating left eye of an anaesthetized trepanned rat with green LED. Map was obtained as the correlation coefficient between the Power Doppler signal and the stimulus pattern. b) Doppler Signal of task-evoked brain activation in visual areas in the rat brain. Power Doppler fUS volumes are acquired every 1.5s. Visual stimulation pattern: 15s ON light/15s OFF light. c) Propagation of a cortical depression wave from the back to the front of the brain during an epileptic seizure induced by 4AP cortical injection. Wave traveling speed = 3 ± 0.3 mm/min.
Keywords: Functional Ultrasound, 3D Brain Imaging, Innovative Imaging Modality, Pre-clinical Imaging
Transcranial Ultrasound Localization Microscopy reveals 100 µm vessels and sub-resolution blood dynamics in the adult Human brain (#14)
Charlie Demené1, Justine Robin1, Mathieu Pernot1, Fabienne Perren2, Mickaël Tanter1
1 Physics for Medicine Paris, Inserm, ESPCI Paris, CNRS, PSL Research University, Paris, France
Human brain vascular imaging is key for management of cerebrovascular and neurological pathologies. Very challenging across modalities, it requires contrast injection, ionizing (CT) or expensive (MRI) imaging devices, overlooks blood dynamics and gives ~0.5mm 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 . 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 , 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, accelerated flow in a 0.9mm-wide stenosis and parabolic speed profile on a 0.8 mm vessel section could be observed, which is impossible with any other neuroimaging modality. Complex flow pattern in a Moya-Moya syndrome could be observed, overstepping the partial information given by luminal-only clinical vascular imaging modalities.
ULM for human brain vascular imaging completely redefines the reachable boundaries of cerebrovascular imaging, with 0.1mm resolution and very local blood flow dynamics assessment. This world premiere is a breakthrough for the management of cerebro-vascular diseases.
This work was suppported by the Fond National Suisse and the European Research Council (ERC) under the European Union’s Seventh Framework Program (FP7/2007–2013)/ERC Advanced grant agreement no. 339244-FUSIMAGINE
 Errico et al, Nature, 2015
 Demené et al, IEEE TMI, 2015
Sub-resolution vessel delineation and local blood flow dynamics assessment via ULM
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: ultrasound, ultrafast imaging, ultrasound microscopy, cerebrovascular imaging, human brain
3D Ultrasound and Photoacoustic imaging of chicken embryo vasculature (#17)
Guillaume Godefroy1, Sergey Vilov1, Bastien Arnal1, Emmanuel Bossy1
1 Université Grenoble Alpes, LIPHY, Grenoble, France
Ultrasound (US) and photoacoustic (PA) imaging can be performed using a single ultrasound transducer coupled to multi-channel electronics. While 2D imaging implies using typically 128 transducers arranged linearly, the number of transducers required to perform 3D volumic imaging can be restrictive. A 2D surface has to be mapped with omnidirectional transducers resulting in thousands of elements and electronic channels implying costly and bulky devices. Strategies in using sparse arrays and reconstructions have been introduced.
Here, we investigate the use of a 256-channel spherically focused sparse array for volumic real-time US and PA acquisitions. The array was designed with an aperture of 55 mm (f=35 mm, freq. 8 MHz). We implemented both US and PA imaging on a programmable ultrasound scanner (Verasonics). US imaging consisted in using 13 different diverging waves emitted by single transducers and spatial coherent compounding was performed to enhance the image quality. Stack of RF signals were acquired and saved with a resulting frame rate of 500 Hz. For PA imaging, an OPO diode-pumped laser working at 100 Hz (Innolas) was coupled to a fiber bundle attached to the transducer. Some fertilized eggs were opened at 6 days of incubation time and blood vessels were imaged with our system.
US and PA reconstructions were performed offline and signal processing techniques using temporal dynamics were applied to the image stack to specifically image the vasculature. The resulting exploration volume is 10 x 8 x 8 mm. US and PA delay-and-sum 3D reconstructions provided different contrasts. The tissue seen with US was not all the time correlated with PA volumes. For visualizing blood flows, US image stacks were then filtered to obtain power Doppler images. Similar structures of the vasculature appeared in US and PA but complementary information can be extracted from both. Due to spatial compounding with 13 emissions, the image quality of US imaging was better than PA imaging, where the object is seen through a single light intensity pattern. The temporal dynamics of the PA signals can be exploited to reduce the clutter thus compensating the sparsity of the array by greatly enhancing the contrast.
We show that 3D real-time US and PA imaging of the vasculature is possible using a sparse array with 256 channels, with a very good image quality. This methodology has a potential for in vivo 3D real time visualization of the vasculature and other features using complementary information provided by US and PA imaging. Further work will focus on identifying specific molecular contrast using PA spectroscopy.
Keywords: photoacoustic imaging, ultrasound imaging, chicken embryo
Characterising the vascular microenvironment of breast cancer patient-derived xenografts using optoacoustic imaging (#21)
Emma L. Brown1, 2, Isabel Quiros- Gonzalez1, 2, 5, Ziqiang Huang2, Joanna Brunker1, 2, Alejandra Bruna2, 3, Carlos Caldas2, 3, 4, Sarah Bohndiek1, 2
1 University of Cambridge, Department of Physics , Cambridge , United Kingdom
Patient derived xenograft (PDX) models are emerging as vital tools in preclinical cancer research. The tumour microenvironment (TME) is key for therapy resistance. While PDXs have been extensively characterised genomically, the TME features are only beginning to be understood. Here, we use two in vivo optoacoustic imaging (OAI) modalities, raster-scanning optoacoustic mesoscopy (RSOM) and multispectral optoacoustic tomography (MSOT) to analyse the vascular microenvironment of breast PDXs of two of the major breast cancer subtypes.
One luminal B PDX and one basal PDX were implanted in vivo in a pilot study (nLumB= 3, nBasal= 5). RSOM and MSOT of the two models were conducted weekly to allow longitudinal monitoring of vascular phenotypes as tumours developed. Tumour blood vessel volume (BVV) was extracted from RSOM images by segmenting the vessels from the background signal. Tumour total haemoglobin (THb=Hb+HbO2) and blood oxygen saturation (SO2= HbO2/THb) were extracted from MSOT images and normalized to a tissue reference region (aorta and vena cava). Once excised, tumours underwent immunohistochemistry to stain CD31 for analysis of microvessel density (MVD).
RSOM allowed high resolution imaging of tumour blood vessels in vivo (Fig.1a). BVV extracted from RSOM images was modelled linearly as tumour volume increased (Fig.1b). The basal PDX started with a lower BVV which increased at a higher rate as tumour volume increased, compared to the luminal B PDX. MSOT imaging is a whole-body imaging technique and has deeper penetration depth compared to RSOM (Fig.1c). In size-matched tumours, preliminary MSOT data showed significant differences in THb (basal 0.33±0.058 a.u. vs. lumB 0.13±0.032 a.u., p=0.04, Fig.1d) and a trend towards differences in SO2 (Fig.1e) between the two models while ex vivo histology (Fig.1f) of the same tumours showed differences in MVD (basal 0.00016±0.000023 vessels/μm2 vs. lumB 0.000066±0.000023 vessels/μm2, p=0.05, Fig.1g). Together, these results indicate that the basal PDX has a denser vascular network, which develops quicker but carries less oxygen to the tumour, compared to the luminal B PDX.
Combining longitudinal RSOM and MSOT imaging with histology gives insight into PDX vasculature across multiple scales over time. Future work will correlate BVV, THb and SO2 measured in vivo with OAI with histological markers of vasculature and vessel maturity, to link the in vivo imaging signal with the underlying tumour biology in a fully powered study.
This work was supported by Cancer Research UK (C14303/A17197).
Optoacoustic imaging of breast cancer PDXs in vivo with ex vivo histology
(A) Representative RSOM image and (B) analysis of blood vessel volume modelled linearly as tumour volume increases (bold lines: line of best fit, dashed lines: individual tumours) (C) Representative MSOT image (tumour outline in red) and analysis for (D,E) medium-size tumours (0.3-0.5cm3). (F) CD31 (black arrow) IHC for excised tumours, MVD quantified in (G). For (B,G), nBasal= 5, nLumB= 3. For (D,E) nBasal= 3, nLumB= 3.
Keywords: Optoacoustic Imaging, Vascular microenvironment, Patient-derived xenografts
Integration of thalamocortical and callosal inputs by optogenetic activation of the rat corpus callosum (CC) with MRI-guided robotic arm (MgRA) (#2)
Yi Chen1, Filip Sobczak1, Patricia Pais1, Cornelius Schwarz2, Alan P. Koretsky3, Xin Yu1, 4
1 Max Planck Institute for Biological Cybernetics, Tübingen, Baden-Württemberg, Germany
The hypothesis that CC inhibits contralateral cortex can explain dampened neural responses in cortex in human and rodents(1-4), e.g., the first stimulus suppressed neural responses to the subsequent stimulus on the other eye within a certain time. Here, we optogenetically activated CC(8) and provided direct evidence for CC-mediated interhemispheric inhibition(II), showing that the direct callosal inputs suppressed evoked calcium and BOLD signals in barrel cortex(BC) by whisker stimulation. Our work links callosal circuit-specific regulation to the global brain dynamic changes based on II(5-7).
AAV.CaMKII.ChR2.mCherry was injected into the BC of rats, expressed in callosal projection neurons (CPN) and along their axonal fiber bundles projecting to the opposite BC (Fig.1a), where the GCaMP6f was expressed (Fig.1f). Optogenetic stimulation will be delivered on corpus callosum, followed by a whisker stimulus to the whisker pad with different intervals (0-200 ms) and the paired conditions of each trail were randomized (Fig. 2c). Whole brain BOLD signals were acquired with simultaneous calcium signal in the BC while an MRI-guided robotic arm was used to precisely target the callosal fiber bundle to deliver blue light pulses (473nm) at 2Hz, 10ms width for the fMRI block design (8s on/52s off,13 epochs,Fig.2b,c). Whole brain 3D EPI: TR,1.5s, 400×400×400 μm3 spatial resolution.
Upon the optogenetic stim on CC, salient BOLD signal was detected due to the antidromic activity from the axonal fibers backward to the soma of callosal projection neurons in the ipsilateral BC (Fig.1c,d), further confirmed by LFP (Fig.1e). For the orthodromic activity, there was clear spike for each stimulus at 2Hz, while with higher frequencies, light flashes 2-16 induced responses were consistently weaker than the first response (Fig.1g), Moreover, there was a baseline drift during the whole 40Hz stimulation period (Fig.1g), therefore, confirming the CC-mediated interhemispheric inhibition. With two stimuli paradigm, the anti-dromic activity in the right cortex kept similar for 6 conditions, while the BOLD and calcium signals in the left cortex induced by paired whisker stimuli was the strongest for OW condition, kept suppressed for the O50W and O100W conditions (Fig.2d-h), almost recovered for the O200W condition.
By taking advantage of fMRI, the optogenetic stimuli on CC and cell-specific calcium signal recordings for layer 5 pyramidal excitatory neurons in the BC, we confirmed the CC-mediated interhemispheric inhibition, further provided direct evidence for the dampened neural responses to subsequent contralateral stimulus after an ipsilateral stimulus for a period of several hundred milliseconds in human and rodents at function level.
We thank Mr. Shanyi Yu for building up the first prototype of the robotic arm, Fine Mechanic and Electronic Workshop at MPI for Biological Cybernetics for MgRA system automation. The financial support of the Max-Planck-Society and the China Scholarship Council (Ph.D. fellowship to Y. Chen) are gratefully acknowledged.
1. Schnitzler A, Kessler KR, & Benecke R (1996) Transcallosally mediated inhibition of interneurons within human primary motor cortex. Exp Brain Res 112(3):381-391.
2. Bocci T, et al. (2011) Transcallosal inhibition dampens neural responses to high contrast stimuli in human visual cortex. Neuroscience 187:43-51.
3. Ogawa S, et al. (2000) An approach to probe some neural systems interaction by functional MRI at neural time scale down to milliseconds. Proc Natl Acad Sci U S A 97(20):11026-11031.
4. Nemoto M, et al. (2012) Diversity of neural-hemodynamic relationships associated with differences in cortical processing during bilateral somatosensory activation in rats. Neuroimage 59(4):3325-3338.
5. Palmer LM, et al. (2012) The cellular basis of GABA(B)-mediated interhemispheric inhibition. Science 335(6071):989-993.
6. Kawaguchi Y (1992) Receptor subtypes involved in callosally-induced postsynaptic potentials in rat frontal agranular cortex in vitro. Exp Brain Res 88(1):33-40.
7. Kumar SS & Huguenard JR (2001) Properties of excitatory synaptic connections mediated by the corpus callosum in the developing rat neocortex. J Neurophysiol 86(6):2973-2985.
8. Yu X, et al. (2013) Targeting projection fibers for optogentics and fMRI.
Anti-dromic and orthodromic activation by corpus callosum optogenetic stimulation.
a Schematic of experimental design and CaMKII.mCherry expression.
b Overview of the MgRA for optical fiber insertion inside 14.1T scanner.
c Averaged fMRI map of brain-wide activity upon optogenetic stimulation on CC from 8 rats.
D Average time courses of BOLD in right BC(n = 8) upon light stimulation. Error bars represent mean±SD.
e The representative local field potential for antidromic activation.
f Schematic of experimental design and CaMKII.mCherry expressed in the right BC while GCaMP6f in the left BC.
g Representative calcium signal changes for 8 s of the orthodromic activation responses.
Simultaneous BOLD and calcium signals upon CC opto stim and whisker stim with varying intervals
a Stimulation scheme. 6 conditions: W, O, O50W, O100W, O200W.
b Experimental setup.
c Typical calcium signals for condition W(blue dotted box) and O100W(red dotted box).
d fMRI map of brain-wide activity for 6 conditions(n=6).
e Averaged normalized calcium signal in left BC.
f Normalized calcium signal for individual rat.
g Averaged BOLD in the left BC (left) evoked by whisker stimulation and right BC (right) evoked by CC stimulation.
h Averaged normalized calcium signal changes across 6 rats for different conditions.
i Averaged normalized calcium signal changes across 4 rats for different conditions.
Keywords: Corpus Callosum, Interhemispheric Inhibition, Thalamocortical Pathway
Light Sheet Fluorescence Expansion Microscopy:Fast Mapping of Neural Circuits at Super Resolution (#12)
Martin K. Schwarz1
1 University of Bonn Medical Faculty, Experimental Epileptology and Cognition Research, Bonn, Germany
The goal of understanding the architecture of neural circuits at the synapse level with a brain-wide perspective has powered the interest in high-speed and large field-of view volumetric imaging at subcellular resolution. Here we developed a method combining tissue expansion and light sheet fluorescence microscopy to allow extended volumetric super resolution high-speed imaging of large tissue samples. We demonstrate the capabilities of this method by performing two color fast volumetric super resolution imaging of mouse CA1 and dentate gyrus molecular-, granule cell- and polymorphic layers.
Gelation and Expansion: The expansion microscopy protocol was adopted from F. Chen, P. W. Tillberg, and E. S. Boyden, “Optical imaging. Expansion microscopy.,” Science 347.
Light Microscopy: For light-sheet microscopy we used a custom-built setup based on a Nikon Eclipse Ti-U inverted microscope (Nikon, Düsseldorf, Germany).
Characterization of optical resolution: The lateral resolution theoretically achievable with an objective lens is given by the Rayleigh criterion, dR, which quantifies the distance between the maximum and first minimum of the point spread function:d=0.61L/NA.
Image Processing: 3D stacks of raw 16-bit images were processed using custom-written MATLAB scripts, which allowed parallel data processing.
In this study we focused on a fast super resolution analysis of large GFP-labeled granule cells ensembles in mouse dorsal DG. Wefirst compared our approach to conventional airy scan confocal imaging to demonstrate the the gain in contrast and axial resolution achieved by LSEM. We next showed the potential of our method to image extended dendritic networks in nanoscale resolution resolving individual spines. In order to show the possibility of neural connectivity mapping we performed experiments demonstrating multicolor labelling of pre- and post-synaptic proteins. To this end we generated DG samples containing sparsely expressing EGFP positive GCs and visualized the mossy fibers within hilus of the DG and identified GABAergic cells as postsynaptic targets of mossy fibers boutons (see attached Figure). Colectively we present an imaging approach that allows the analysis of extended neural networks in super resolution and facilitating short imaging times.
Finally we would like to stress, that a decisive factor in imaging extended neural networks by expansion light sheet microscopy is imaging duration. Thus we are currently optimizing our setup and expect to increase the imaging rate by a factor of 10 to 20. Such a device would allow nanoscale imaging of our specimen (3.93 mm3) in about 10 hours.
This work was supported by the German Research Foundation (grant numbers KU 2474/13-1, SCHW 1578/2-1 and INST 217/886-1).
Light-sheet fluorescence expansion microscopy: fast mapping of neural circuits at super resolution.
Bürgers J, Pavlova I, Rodriguez-Gatica JE, Henneberger C, Oeller M, Ruland JA, Siebrasse JP, Kubitscheck U, Schwarz MK.
Neurophotonics. 2019 Jan;6(1):015005. doi: 10.1117/1.NPh.6.1.015005. Epub 2019 Feb 8.
Cortical column and whole-brain imaging with molecular contrast and nanoscale resolution.
Gao R, Asano SM, Upadhyayula S, Pisarev I, Milkie DE, Liu TL, Singh V, Graves A, Huynh GH, Zhao Y, Bogovic J, Colonell J, Ott CM, Zugates C, Tappan S, Rodriguez A, Mosaliganti KR, Sheu SH, Pasolli HA, Pang S, Xu CS, Megason SG, Hess H, Lippincott-Schwartz J, Hantman A, Rubin GM, Kirchhausen T, Saalfeld S, Aso Y, Boyden ES, Betzig E.
Science. 2019 Jan 18;363(6424).
Two color imaging of mossy fibers and GABAergic interneurons in DG
(A) Mossy fibers in the hilus area expressing EGFP. The endogenous EGFP fluorescence was enhanced by antibody staining against EGFP, the secondary antibody was conjugated to Alexa 488 (green). Parvalbumin staining identified GABAergic interneurons shown in red (video 5, MP4, 65 MB). 1500 optical slices were acquired with a step size of 0.3 µm, the shown data was deconvolved. Volume size 456 x 945 x 390 μm3. (B) Side view of the data shown in (A). (C) Segmented parvalbumin cells and mossy fibers reconstructed in 3D. Magnification of the ROI marked in B, showing connection between the cells.
Keywords: Connectomics, super resolution, tissue expansion, dentate gyrus, light sheet fluorescence microscopy
A novel illumination system based on long-range electromagnetic surface waves for fluorescence microscopy (#7)
Dmitry Bagrov1, 2, Kirill Prusakov1, Dmitry Basmanov1, Dmitry Klinov1
1 Federal State Budgetary Institution Federal Research and Clinical Center of Physical-Chemical Medicine Federal Medical Biological Agency, Laboratory of medical nanotechnology, Moscow, Russian Federation
One-dimensional photonic crystal (1D PC) is a flat multilayer structure which consists of layers with alternating refractive indices. Under special conditions, the long-range electromagnetic surface waves can travel along the surface of a 1D PC . Here we have shown that these waves can be used to excite fluorescence in the samples adsorbed onto the surface of the 1D PC. We developed a novel illumination system for a fluorescence microscope, which exploited a 1D PC substrate , and used this system to image bacterial and eukaryotic cells.
Our illumination systems are shown in figure 1; we used a modified Kretschmann scheme. The studied cells with the fluorescent labels were adsorbed onto (or grown on) the surface of the 1D PC and mounted to a prism. A cylindrical lens focused the laser (either 473 nm or 660 nm) to the sample plane. The scheme was mounted either in the upright or the inverted configuration (figure 1). The inverted configuration was mounted on the Nikon Eclipse microscope; the upright configuration was built from optical components. The studied cells were E.coli (expressing GFP), B.subtilis (stained with Holosens) and HeLa (expressing GFP). For comparison, the images of the 1D PC surface were captured the help of a wide-field fluorescence microscope.
The experiments showed that fluorescence was excited only inside the thin near-surface layer (~150 nm) of a specimen. Figure 2 compares the images of HeLa cells obtained using traditional epi-fluorescence and the novel illumination system. When the surface waves were used for excitation, the microscopic filopodia became visible at the cell-substrate interface. Similarly, when the bacterial cells were imaged, the signal to noise ratio was on the average seven times higher than using the epi-fluorescence. We obtained images which were similar to the ones obtained using Total Internal Reflection Fluorescence (TIRF) microscope. The most commercial TIRF systems are objective-based and require special high-NA objectives; however, our system can operate with almost any objectives. Our system looks similar to the prism-based TIRF. However, we can change the penetration of the surface waves into the sample in the approximate range of 70-500 nm by changing the parameters of the 1D PC substrate.
The fluorescence of biological samples can be excited by the electromagnetic surface waves at the interface between the liquid medium and the 1D PC. It was demonstrated using both procaryotic and eucaryotic cells. Our illumination system can be used in either upright or inverted configuration and does not require fine tuning or specialized objectives as the TIRF systems.
This work was supported by the Russian science foundation (project № 17-75-30064).
 V.N. Konopsky, E. V. Alieva, Long-range propagation of plasmon polaritons in a thin metal film on a one-dimensional photonic crystal surface, Phys. Rev. Lett. 97 (2006) 1–4. doi:10.1103/PhysRevLett.97.253904.
 K.A. Prusakov, D. V. Basmanov, D. V. Klinov, Patent 2626269 (RU), 2017.
Keywords: Fluorescence microscopy, Photonic crystals, TIRF, Cell imaging
Phase retrieved tomography for alignment-free and hidden 3D optical imaging (#18)
Daniele Ancora1, 2, 5, Diego Di Battista2, 5, Georgia Giasafaki2, Styliiianos Psycharakis2, Evangelos Liapis2, Jorge Ripoll3, 4, Giannis Zacharakis2
1 Politecnico di Milano, Department of Physics, Milano, Italy
In the field of biomedical optics, the co-registration of volumetric images acquired by SPIM or scan OPT must deal with problematics connected with measurement inaccuracies and misplacement of the sample during the measurement. This augments the complexity of the problem and leads to the formation of errors in the reconstruction. Here we describe the ideas behind the project HI-PHRET project that deals with novel computational methods for accurate quantitative tomography, trying to overcome the needs for setup and data alignment with the usage of phase retrieval-based algorithms.
Our approach is implemented in three-dimension using iterative Gerchberg–Saxton methods, mapping the autocorrelation of the object as estimation of its Fourier modulus. The phase connected to such volume is then retrieved via the introduction of appropriate object constraints, relying on the fact that the reconstructed volume should be real and positive. By appropriately implementing these algorithm classes, we show how to make use of the autocorrelation sinograms even in case the signal trespasses optically turbid media, being able to perform an aligned reconstruction even in such difficult measurement scenario.
We worked on model specimen such as human tumor spheroids expressing DRAQ7 fluorescence in the necrotic cells. Using an optical projection tomography approach relying on the autocorrelation sinogram inversion, we shown the effectiveness of the reconstruction technique in the case of hidden three-dimensional imaging [1-3] and, at the same time, to reconstruct fluorescent distribution in a combined light sheet microscopy experiment [4,5]. The protocol proposed is completely automated and has the potential to replace existent tomographic algorithms to reconstruct even beyond turbid curtains.
The advantages of our method are numerous: its usability in case of imaging through opaque samples and insensitiveness to mechanical misalignments make it a promising tool for the definition of a robust protocol for any three-dimensional tomographic reconstructions. Our will is to keep on developing this approach, extending it to additional dimensionalities and importing novel approaches for the solution of the phase problem.
The work was supported by the EU Marie Curie ITN “OILTEBIA” (PITN-GA-2012-317526) and the EU Marie Curie Individual Fellowship “HI-PHRET”, MSCA-IF-2017 grant number 799230 under the framework Horizon 2020.
Keywords: tomography, phase retrieval, scattering imaging, fluorescent microscopy