17th 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).

Mechanisms of Remodeling in Cardiac/Pulmonary Diseases

Session chair: Gustav Strijkers (Amsterdam, Netherlands); Alkystis Phinikaridou (London, UK)
Shortcut: PS 03
Date: Wednesday, 16 March, 2022, 11:30 a.m. - 1:00 p.m.
Session type: Parallel Session


Click at talk title to open the abstract

11:30 a.m. PS 03-01

Introductory Lecture

Angelos Karlas

Munich, Germany

11:50 a.m. PS 03-02

Identification of Subtle Valvular, Myocardial and Aortic Alterations for Demarcation of Disease Transition in Murine Aortic Valve Stenosis by Multiparametric MRI

Christine Quast1, Christoph Jacoby1, Malte Kelm1, Ulrich Flögel1

1 Heinrich Heine University, Düsseldorf, Germany


Aortic valve stenosis (AS) is one of the most frequent valve diseases in the elderly with relevant prognostic impact. Because sufficient animal models were lacking, we recently refined a murine model of gradable experimental AS closely mimicking disease progression in humans[1]. Here, we aimed at developing a comprehensive MRI approach for simultaneous assessment of changes in valvular and aortic anatomy/function in this model and its impact on myocardial morphology and tissue texture.


Experimental AS was induced by wire injury [1] in male 12-week-old C57Bl/6 mice which were subjected 4 weeks after surgery to high resolution MRI at a Bruker AVANCE 9.4T Wide Bore NMR spectrometer. Images were acquired utilizing a microimaging unit with actively shielded gradient sets (1.5 T/m) and a 25-mm quadrature resonator. Cine loops were recorded using an ECG- and respiratory-gated FISP sequence (TE=1.23 ms). Aortic strain was determined from the circumferential dimensions of the aorta in end-diastole and ‑systole. Aortic flow profiles were obtained from velocity maps acquired at the atrio-ventricular level using an ECG- and respiration-triggered slice-selective 2D FLASH sequence.  Cardiac tissue characterization was carried out by T1 and T2 mapping as described previously [2].


As expected, AS mice exhibited an impaired valve opening and long axis slices revealed an altered shape and thickening of the valve as consequence of the remodelling processes (Fig. 1A-E). While control mice showed a bell-shaped flow profile in early systole, AS resulted not only in significantly enhanced peak flow velocity, but also in a fragmented flow pattern with isolated spikes of accelerated velocities across the valvular area and also negative flow (Fig. 1F-I). This resulted in enhanced transversal strain of the aortic root, which was also associated with an increase in aortic wall thickness. Of note, global cardiac function and cardiac output was unaffected in AS mice but diastolic wall thickness was significantly bigger comparable to a mild hypertrophy (Fig. 2A-D). Finally, cardiac tissue characterization by T1 and T2 mapping surprisingly revealed reduced values for both parameters in AS (Fig. 2E+F) which was associated with ultrastructural impairments in electron microscopy.


We demonstrate that in experimental AS high resolution MRI provides important information: Besides expected alterations in aortic valve function and flow profiles, incipient remodeling of the aortic wall and myocardial tissue integrity get apparent. With this, we identified a premature transition point with still compensated cardiac function but beginning textural changes relevant for early disease pathophysiology and novel therapeutic targets.


We thank Kathrin Paul-Krahé, Steffi Becher, Julia Pauli, Annika Zimmermann, and Dr. Ann Kathrin Bergmann for excellent assistance in surgical procedures, histology, and electron microscopy, respectively.


 I or one of my co-authors have no financial interest or relationship to disclose regarding the subject matter of this presentation.

[1] Niepmann et al. Clin Res Cardiol. 2019; 108(8), 847-856.
[2] Haberkorn SM et al. Circ Cardiovasc Imaging 2017;10(8), pii: e006025.
Figure 1: Valvular structure + function and aortic flow profiles
(A+B) Valve opening at the atrio-ventricular level in sham control (left) and AS (right) mice. (C) Opening of the aortic valve  expressed as percentage of maximal valve orifice. (D) Long axis views revealing structural remodelling of the valve in AS. (E) Quantification of aortic leaflet area; n=9 each, ***P<0.001. (F+G) Anatomical reference (left) and 2D velocity maps (middle). Velocity flow profiles are displayed as 3D surface plots in early systole (right). Peak velocity was higher in AS (H) and occurred later after valve opening (I). Arrows indicate valve orifice; n=9 each, ***P<0.001.
Figure 2: Left ventricular structural and functional changes
(A+B) Global left ventricular volumes, ejection fraction and cardiac output were unaltered in AS mice. (C) Regional analysis of cardiac function; for the sake of clarity, SDs are plotted in light grey and data of untreated controls were omitted; n=9 each. (D) Anatomical reference images demonstrated a significantly increased diastolic wall thickness in AS mice, which was accompanied by reduced myocardial T1 and T2 values as revealed by MR relaxometry (E+F); n=9 each, *P<0.05, **P<0.01.
Keywords: Aortic valve stenosis, MRI, aortic flow patterns, vessel wall remodelling, tissue characterization
12:00 p.m. PS 03-03

In vitro and in vivo validation of 68Ga-FAPI-46 to assess fibroblast activation after cardiac ischemic injury

Maday Fernandez-Mayola1, Annika Hess1, Michael Willmann1, Tobias L. Ross1, Frank M. Bengel1, James T. Thackeray1

1 Hannover Medical School, Nuclear Medicine, Hannover, Germany


Activation of quiescent cardiac fibroblasts after myocardial infarction (MI) contributes to ventricle remodeling and heart failure. Thus they are an attractive therapeutic target, but the precise temporal dynamics of fibroblast activity remain equivocal. Recent development of imaging agents targeting fibroblast activation protein (FAP) may provide unique insights into fibroblast dynamics. We aimed to characterize the binding characteristics of the FAP radioligand 68Ga-FAPI-46 in relation to FAP expression. We further assessed temporal tracer binding in response to MI in a mouse model.


68Ga-FAPI-46 binding specificity for FAP was determined in human fibrosarcoma HT1080 cells transfected with human or murine FAP by incubation with excess unlabeled FAPI-46 (1nM). We further assessed 68Ga-FAPI-46 binding in quiescent human cardiac fibroblasts (HCF) and after activation by stimulation with transforming growth factor-β (TGFβ, 2nM 48h). Tracer binding was directly compared to FAP expression as measured by commercial ELISA. We then assessed the basal signal of 68Ga-FAPI-46 in C57BL/6N mice (n=5). To assess the temporal expression of FAP after ischemic injury, C57BL/6N mice underwent permanent occlusion of the left coronary artery (PO, n=6) or 60min ischemia/reperfusion (I/R, n=6). PET imaging was performed 1, 3, 7 and 21d after surgery.


68Ga-FAPI-46 displayed specific uptake by HT1080 cells overexpressing mouse or human FAP (%uptake, mFAP: 3.5±0.8, hFAP: 0.5±0.5, vs wt: 0.01±0.01, p<0.001). Coincubation with unlabeled compound lowered binding in mFAP and hFAP by 98% (p<0.001) and 94% (p=0.04), respectively. Tracer uptake correlated with FAP protein expression (r=0.79, p<0.001). In HCF, 68Ga-FAPI-46 uptake was relatively low but significantly reduced by blocking (0.06±0.02 vs 0.04±0.01, p=0.03). TGFβ stimulation resulted in 1.4 times higher tracer uptake (0.084±0.01 vs 0.06±0.02, p<0.001) and 1.3 times higher FAP expression (16±4.4 vs 11.9±5.3 ng/mL, p=0.01). 68Ga-FAPI-46 biodistribution in healthy mice showed rapid blood clearance and low cardiac signal. At 7d after PO, infarct uptake tended to be higher than healthy controls (%ID/g, 1.0±0.4 vs 0.7±0.1, p=0.1). After I/R, elevated 68Ga-FAPI-46 uptake occurred earlier, with maximal signal at 3d, declining by 7d (%ID/g, 1.3±0.4 vs 0.8±0.3, p<0.001).


68Ga-FAPI-46 selectively binds to FAP in vitro proportional to protein expression. HCF tracer uptake is sensitive to TGFβ stimulation, suggesting discrimination between quiescent and activated state. 68Ga-FAPI-46 detects transient FAP elevation in the infarct territory, which peaks within the first week after injury. The significance of this elevation for subsequent left ventricular dysfunction and remodeling warrants further investigation.


Funding from Leducq Transatlantic Network ImmunoFib and Lower Saxony support of REBIRTH Center for Tranlsational Regenerative Medicine.


  • a) I or one of my co-authors have no financial interest or relationship to disclose regarding the subject matter of this presentation.

68Ga-FAPI-46 assessment in vitro and in vivo

68Ga-FAPI-46 uptake (A) and FAP expression (B) by human cardiac fibroblasts (HCF) at baseline and after 48h stimulation by TGF-β. (C) Representative 68Ga-FAPI-46 (colorscale) and 18FDG (grayscale) fused PET images in healthy control mice and at 7d after permanent coronary artery ligation (PO) display increased tracer accumulation in the nonviable infarct region.

Keywords: 68Ga-FAPI-46, Fibroblast activation, myocardial infarction
12:10 p.m. PS 03-04

Longitudinal Characterization of Biohybrid Tissue-Engineered Vascular Grafts Scaffold Resorption and Tissue Remodelling via Molecular MR and US

Elena Rama1, Saurav R. Mohapatra2, Christoph Melcher3, Teresa Nolte1, Seyed M. Dadfar1, Ramona Brück1, Vertika Pathak1, Anne Rix1, Volkmar Schulz1, Twan Lammers1, Christian Apel2, Stefan Jockenhoevel2, Fabian Kiessling1

1 RWTH Aachen University, Institute of Experimental Molecular Imaging, Aachen, Germany
2 RWTH Aachen University, Biohybrid and Medical Textile Institute, Aachen, Germany
3 RWTH Aachen University, Institute of Textile Technology, Aachen, Germany


The longitudinal monitoring of scaffold remodeling and the onset of inflammatory reactions of implanted vascular grafts remains challenging. Moreover, the transition from in vitro conditioning to in vivo implantation of vascular grafts remains inadequately studied. To overcome this issue, we developed a non-invasive imaging approach that allowed us to monitor the degradation of textile elements of biohybrid tissue-engineered vascular grafts (TEVG) in vitro and in vivo, the deposition of extracellular matrix (ECM), and the onset of inflammation by molecular MRI and US.


Our TEVG consists of a non-degradable PVDF scaffold coated with SPION-labeled degradable PLGA fibers molded into fibrin gel containing endothelial and smooth muscle cells. The PLGA fibers degradation was longitudinally monitored by MRI, first, in different in vitro experimental setups and, subsequently, in vivo for 21 days by subcutaneously implanting the TEVGs in Lewis rats. The ECM deposition was studied in TEVGs after 14 days of maturation using elastin- and collagen type I-targeted gadolinium-containing MR molecular agents (Gd-DTPA-ESMA and EP-3533) and compared to negative and positive controls. The αvβ3 integrin expression as a marker of inflamed endothelium was assessed by molecular US imaging using RGD-targeted microbubbles (MB) and compared to control RAD-MB.


Both in vitro and in vivo, the degradation of SPION-PLGA fibers was reflected by a decrease in R2 relaxation rates due to the release of SPION (Fig.1). First, the degradation of the PLGA fibers was monitored in vitro under static and dynamic flow conditions, showing a faster degradation rate in the latter case. Subsequently, as a proof of concept, the degradation rate of our SPION-PLGA fibers and corresponding decrease in R2 relaxation rats were longitudinally monitored in vivo for 21 days in rats. Regarding the deposition of ECM components, neither molecular imaging nor histology indicated elastin production under bioreactor conditions. In contrast, strong collagen presence in TEVG was demonstrated by EP-3533 MRI. No RGD-MB binding was observed during US investigation of TEVGs indicating no expression of αvβ3 integrin. However, mimicking an inflammatory state by adding TNF-α to the TEVG, resulted in integrin expression and binding of RGD-MBs, while control RAD-MBs still did not bind.


We here introduce a non-invasive imaging approach to longitudinally monitor the degradation of TEVG components during the in vitro bioreactor maturation and in vivo implantation. We also prove the detectability of PLGA fiber replacement by ECM via molecular MRI and confirm the presence of non-inflamed endothelium by molecular US. Our approach may be highly valuable for quality-control in the critical pre-and post-implantation phases of TEVG.


This work was supported by Deutsche Forschungsgemeinschaft (DFG: 403039938).


I or one of my co-authors have no financial interest or relationship to disclose regarding the subject matter of this presentation.

Figure 1. Longitudinal MR investigation of SPION-PLGA fibers.

A: MR images of PLGA fibers combined with a PVDF tubular structure with quantitative R2 analyses (B-C) showing the correlation between SPION-PLGA fiber degradation over time and the decrease of relaxation rate values. D: Degradation rate of PLGA fibers investigated in combination with cellular components under dynamic conditioning. E-F: The R2 analysis shows an accelerated degradation. G-H: R2 decrease over time is also observed in vivo when SPION-labeled TEVGs are implanted subcutaneously in rats. (*p<0.05, **p<0.01, ***p<0.001, and ****p<0.0001).

Figure 2. MRI characterization of ECM components in TEVG and US assessment of endothelial coverage.
A-D: MR images of Gd-DTPA-ESMA, EP-3533, and untargeted Gd-DTPA in TEVGs (A) and in human umbilical arteries (HUA) (D). B-F: Quantitative R2 analyses of TEVGs (B-C) and HUA (E-F) show strong EP-3533 binding in both TEVGs and HUA, whereas no Gd-DTPA-ESMA is detected in the vascular grafts. G: US longitudinal monitoring of αvβ3 integrins is shown in both untreated and TNF-α treated TEVG. H: RGD-MB higher binding is observed in the TNF-α treated TEVG compared to the untreated samples and RAD-control MB. (*p<0.05, **p<0.01, ***p<0.001, and ****p<0.0001).
Keywords: Tissue-engineered vascualr graft, Molecular MRI, Targeted US, PLGA degradation, SPION
12:20 p.m. PS 03-05

Lung imaging with hyperpolarised Xenon MRI correlates with chronic long-COVID patient breathlessness, in absence of structural damage on CT and within normal range lung function results: preliminary results.

James T. Grist1, 2, Huw Walters1, Guilhem J. Collier3, Mitchell Chen1, Rolf F. Schulte4, Gabriele Abu Eid1, Violet Matthews1, Aviana Laws1, Kenneth Jacob1, William Hickes1, Anthony McIntyre1, Marianne Durrant1, Susan Cross1, Alex Eves1, Emily Fraser5, James Wild3, Fergus Gleeson1

1 Oxford University Hospitals NHS Trust, Radiology, Oxford, United Kingdom
2 University of Oxford, Physiology, Anatomy, and Genetics, Oxford, United Kingdom
3 University of Sheffield, POLARIS, Sheffield, United Kingdom
4 GE Healthcare, Munich, Germany
5 University of Oxford, Oxford Interstitial Lung Disease Service, Oxford, United Kingdom


There is emerging evidence of long-term challenges for patients after COVID-19 infection, often called 'Long-COVID', with many reporting ongoing symptoms many weeks and months after the acute illness(1). Two of the symptoms in patients with Long-COVID are fatigue and breathlessness, commonly with no identifiable cause on imaging, lung function or blood tests(2).


This study was approved by the HRA, and all participants gave informed consent. Participants were recruited from the Post-COVID clinic. Inclusion required PCR evidence of previous COVID-19 infection, and exclusion criteria included prior hospital admission for COVID-19. Participants were imaged with hyperpolarised Xenon MRI, inspiratory CT, underwent lung function tests (all as previously described(2)) and breathlessness scored via Dyspnea-12 and modified BORG pre- and post- a 60 second sit-stand test.


CT imaging was scored by a thoracic radiologist(2), and the mean, standard deviation, and coefficient of variation of xenon MRI was calculated on a per-patient basis. Imaging and clinical data were correlated using Spearman’s correlation.


8 patients were recruited to this study. Patient CTs were normal or near normal (mean CT score = 0.3/25 ± 0.6) and displayed normal lung function results (mean Forced Expiratory Volume = 100 ± 13% mean TLco = 78 ± 8%, range 65-90%). The mean Red Blood Cell: Tissue Plasma (RBC:TP) ratio was 0.37 ± 0.11. Patients were dyspnoeic (mean Dyspnea-12 score = 1 ± 5, mean mBORG pre- sit stand = 2 ± 2, mean mBORG post-sit stand = 7 ± 1. Example imaging can be seen in Figure 1; most patients were in the bottom 2.5 percentile for their age with an average of 27 ± 12 repetitions.

There were significant correlations (see Figure 2) between mBORG pre-sit-stand and mean RBC:TP (correlation coefficient (cc) = -0.63, p = 0.04) and RBC:TP mean and standard deviation (cc = 0.63, p = 0.05), and TLco and RBC:TP standard deviation (cc = 0.78, p = 0.02).

These results demonstrate that, in the absence of visible abnormalities on CT, there may be significant lung pathology in some patients with long-COVID that can be visualised and quantified by hyperpolarised Xenon MRI.


This study has demonstrated ongoing pathology in the lungs of long-COVID patients with ongoing dyspnoea that cannot be detected with conventional CT or lung function tests. Further studies involving larger cohorts of patients and multi-centre results are ongoing to continue to verify these initial results.


I or one of my co-authors have no financial interest or relationship to disclose regarding the subject matter of this presentation.

[1] Huang C, Huang L, Wang Y, et al. 6-month consequences of COVID-19 in patients discharged from hospital: a cohort study. Lancet 2021;397:220–232 doi: 10.1016/S0140-6736(20)32656-8.
[2] Grist JT, Chen M, Collier GJ, et al. Hyperpolarized 129Xe MRI Abnormalities in Dyspneic Participants 3 Months after COVID-19 Pneumonia: Preliminary Results Manuscript. Radiology 2021;0:0.
Figure 1

Figure 1. Example CT, proton, Gas, TP, and RBC imaging from Long-COVID patients. The top row is a patient with RBC:TP = 0.49, the middle row is a patient with RBC:TP of 0.31, and the bottom row is a patient with RBC:TP = 0.24.

Figure 2

A – a significant negative correlation between mBORG pre-sit stand and RBC:TP mean. B A significant positive correlation between Tlco (%) and RBC:TP Standard Deviation (STD). C – a significant positive correlation between RBC:TP mean and RBC:TP STD.


Keywords: COVID, Lung, Xenon, MRI
12:30 p.m. PS 03-06

Multiscale High-Resolution Phase-Contrast X-Ray Tomography of Human Lung

Jakob Reichmann1, Stijn Verleden2, Mark Kühnel3, Jan-Christopher Kamp3, Lavinia Neubert3, Thanh Quynh Bui1, Danny Jonigk3, Tim Salditt1

1 University of Göttingen, Institute for X-ray Physics, Göttingen, Lower Saxony, Germany
2 University of Antwerp, Antwerp Surgical Training, Anatomy and Research Centre (ASTARC), Wilrijk, Belgium
3 Hannover Medical School, Institute of Pathology, Hannover, Germany


The human lungs are the most important organ of the respiratory system. They are responsible for supplying oxygen to the blood, which reaches the erythrocytes by diffusion through the alveolar walls and is then distributed throughout the body. By exploiting the difference in electron density detected by a phase shift in soft tissue [1], high-resolution phase-contrast computed tomography (PCCT) can resolve biological structures in a sub-μm range [2,3], shedding new light on the three-dimensional structure of the lungs, physiological functions and pathological mechanisms.


Experiments were performed with in-lab sources as well as with synchrotron radiation at the Coherence Applications Beamline at DESY (Hamburg, Germany)[4]. In the laboratory, a micro-focus liquid metal jet X-ray source (Excillum, AB) with a CMOS camera sensor and an EasyTom scanner (RX Solutions) with a CCD camera were used. Samples were scanned freshly in alcohol, in cryo-environment as well as embedded in paraffin in cone and parallel beam geometries. For the latter, overlapping tomograms were acquired at different positions and stitched together after the reconstruction. A nonlinear model using Tikhonov regularization was used for phase retrieval before tomographic reconstruction and segmentation. A correlative approach by comparing the results with histological slices was applied.


The multiscale approach resulted in a large FOV while resolving nuclei and fine structures in the lung parenchyma. Paraffin embedded and fresh human lung samples in a 98% EtOH solution with alveolar capillary dysplasia (ACD), with and without a misalignment of pulmonary veins were examined with this configuration, revealing unusual courses of vasculature.

In addition, paraffin embedded lungs of patients with chronic obstructive pulmonary disease (COPD) were examined with the same configuration as well as in a cryogenically fixated state. With the former, small airways as well as bronchial circulation, extensive emphysema and a potential Lambert channel were made visible. For the latter, a custom-made box was designed to keep the sample frozen during the acquisition. Thus, delicate structures could be preserved which are normally deformed during additional sample preparation. Also, due to the high difference in electron density from tissue to air a very high contrast could be achieved.


Thanks to its high sensitivity to structure with quantitative density contrast, PCCT is particularly well suited for examining lung samples with their largely air-filled compartments [5]. Compared to histology, PCCT allows non-destructive, multiscale, three-dimensional imaging of unstained human lung tissue in various environments, thus preserving fragile anatomical structures while resolving features down to the nanoscale.


We thank Markus Osterhoff, Michael Sprung, and Fabian Westermeier for support at P10. The research was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – Project-ID 432680300 – SFB 1456.


 I or one of my co-authors have no financial interest or relationship to disclose regarding the subject matter of this presentation.

[1] Bartels, M., Krenkel, Salditt, T. (2015). "X-ray holographic imaging of hydrated biological cells insolution", Physical review letters, 114(4), 048103.
[2] Marina Eckermann, Bernhard Schmitzer, Ove Hansen, Tim Salditt et al., (2021). “3d virtual histology of the human hippocampus based on phase-contrast computed-tomography”, Proceedings of the National Academy of Sciences, Nov 2021, 118 (48) e2113835118.
[3] Frohn J, Pinkert-Leetsch D, Missbach-Guntner J, Reichardt M, Osterhoff M, Alves F, Salditt T, (2020). „3D virtual histology of human pancreatic tissue by multiscale phase-contrast X-ray tomography“ , J. Synchrotron Rad. 27, 1707-1719.
[4] Salditt, T., Osterhoff, Sprung, M. (2015). „Compound focusing mirror and X-ray waveguide optics for coherent imaging and nano-diffraction”. Journal of synchrotron radiation, 22(4), 867-878.
[5] Eckermann M, Frohn J, Reichardt M, Osterhoff M, Sprung M, Westermeier F, Tzankov A, Werlein C, Kühnel M, Jonigk D, Salditt T. (2020). '3D virtual pathohistology of lung tissue from Covid-19 patients based on phase contrast X-ray tomography', eLife 2020;9:e60408.
Synchrotron lung imaging of paraffin embedded ACD-samples

a) Reconstructed slice of a lung sample with a bronchovascular bundle from patients with alveolar capillary dysplasia (ACD). Nine volumes (each ~1.6 mm3) were merged to a 3x3 tomogram (FOV: ~2.5 x 2.5 mm, cropped) and binned (2x2), resulting in a voxel size of 1.3 μm.

b) Three-dimensional visualization of the segmented bronchi (yellow) and vasculature (artery: orange; vein: red) with three orthogonal slices.

c) Rendered volume of a reconstructed lung tomogram.

d) Correlation of reconstructed images with H&E stained histological slices at region of interest.

Multiscale synchrotron imaging of samples with pulmonary diseases

Lung sample with severe idiopathic pulmonary fibrosis in 3.5 mm capton tube.

Left: Nine volumes (each ~1.6 mm3) were merged to a 3x3 tomogram (FOV: ~3.3 x 3.3 mm) and binned (2x2), resulting in a voxel size of 1.3 μm.

Right: Region of interest with 650 nm voxel size. Clearly resolved erythrocytes, vasculature and air-filled compartments in the lung. 

Keywords: virtual histology, pulmonary diseases, x-ray tomography, lung imaging, human lung