Our methods

Micro-computed tomography: all the tiny details

What is it?

Conventional computed tomography (CT) would be far too large and inaccurate for small animals and bone samples. This is why we use micro-computed tomography (µCT). In contrast to its ‘big brother’ in hospitals, µCT can be used to examine areas of up to 0.01 millimetres and identify them on the images. This allows us to see the bone microstructure and use it to research diseases, mineralization and fractures.

How does it work?

Before each examination, we have to secure the samples so they don’t move. Because that would lead to blurred images – similarly to photographs with low exposure. The µCT works by transmitting X-rays. In contrast to simple 2D images, we create 3D images of the samples. To do this, the computer combines numerous 2D images from different angles to produce a three-dimensional image. We can study objects with a maximum diameter of 80 millimetres and a length of up to 120 millimetres.

What can we see with it?

With µCT, we study all kinds of changes that occur in bone, due to, for example,  cancer or infections. But we also see how well implants – screws or nails – function and are accepted by the body. We are particularly interested in how well the bone grows around the respective material. It’s not only implants that we can examine with µCT,  but practically any kind of synthetic and natural material – even marine organisms. Crabs, seahorses, pipefishes, corals and microatolls wind up in our machine. Archaeologists also appreciate our work: with excavated skeletons, we can tell whether the person, for example, was injured by a sword blow. But there are many more possible applications of µCT. We can detect, for example, which minerals are contained in samples from boreholes,.

What equipment or programmes do we have available?

We operate a SCANCO VivaCT-80 in-vivo micro-CT scanner with a resolution of up to 10 micrometres (0.01mm) and a maximum sample diameter of 80mm. In the case of larger samples, we also have access by arrangement to various clinical CT scanners, but with a slightly reduced resolution of about 250 micrometres (0.25mm); here, however, the samples can have a diameter of up to 800mm. We also maintain contacts with several working groups at DESY in Hamburg that conduct research in the field of synchrotron CT.

For evaluation, we use the manufacturer’s (SCANCO) established software based on the Image Processing Language (IPL). We can also handle special evaluation scenarios using our own StructuralInsight program, an application written in C++ and based on the Imaging Toolkit (itk), which combines all of the steps of a quantitative evaluation of both clinical and preclinical study data and can be further enhanced by us as desired.

micro-CT scanner

Our SCANCO micro-CT scanner

MRT – magnetic resonance imaging

Our 7T-MRT-Scanner by Bruker

The magnet that makes pictures

What is it?

Magnetic resonance imaging (MRI) makes it possible to take extremely detailed, contrast-rich images of objects or living creatures without harmful radiation. In particular, MRI is characterised by excellent soft tissue contrast and the ability to measure various functions such as blood circulation or metabolism. MRI is less suitable for imaging bones. We use CT for this.

How does it work?

Magnetic resonance imaging measures the distribution and properties of nuclear spins, mostly in hydrogen. Since we are largely made of water (and therefore hydrogen), images can be taken from every part of the body.

What can we do with it?

The range of MRI applications are incredibly broad. Classically, anatomical structures can be imaged, measured and diagnosed; think of the size of the liver or the condition of the brain. But functional parameters can also be measured, such as blood flow to the brain, bone healing or inflammation.

What equipment or programmes do we have available?

At MOIN CC, we have two high-resolution state-of-the-art MRI machines, a Bruker Biospec 7/30 (7T) and a WB400 (9.4T). We also have two benchtop NMRs (Magritek SpinSolves) and two MRI scanners for research and teaching (Pure Devices and Magritek TerraNova). For larger objects, we use the clinical equipment at UKSH.

  • 7T, 30cm, 1H and x-nuclei MRI – usable diameter approx. 7.2cm (Bruker Biospec 7/30, Avance Neo, Paravision 360)
    • RAPID and in-house (1H, 13C, 23Na, 2H, 31P) surface and volume coils for various applications
  • 9.4T, 9cm, 1H and x-nuclei NMR/micro-MRI – usable diameter approx. 2.5–3cm. Gradient 1.5T per metre.

All cores for samples < 1cm samples, selected cores for samples <2.5 cm.

  • Vital sign monitoring and triggering (respiration, heartbeat, temperature; SAI Instruments)
  • Injection and isoflurane anaesthesia
  • RF coil workstation, Rohde & Schwarz network analyser (xy) and network analyser for use in the MRI room (SDR). High-frequency oscilloscopes (Rohde & Schwarz)
  • Pure Devices 0.5T desktop MRI
    • 1H, 13C, 129Xe coils
  • Magritek Terranova earth-field MRI

At UKSH

  • Siemens Vida 3T whole-body MR scanner, with x-nuclei option
    • All standard coils for clinical care, incl. 64- and 20-channel head coil and microscopy, knee and leg angiography coils
    • Noras dental coil
    • RAPID x-nuclei coils
    • In-house developments
  • Philips Ingenia 3T WB-MRI scanner
    • All standard coils for clinical care, incl. 64- and 20-channel head coil, knee and leg angiography coil
  • Siemens Aera 1.5T whole-body MR scanner
  • Philips 1.5T whole-body MR scanner
  •  3T WB-MRI scanner

FMT – fluorescence molecular tomography

Using light to see through tissue

What is it?

Fluorescence imaging can be used to measure minute amounts of a dye in the living organism. This is particularly useful for seeing how labelled cells, bacteria or antibodies are distributed in the body. While the sensitivity is very high (a few molecules are enough), the spatial resolution is unfortunately limited. This is where cross-sectional imaging –tomography – can help, measuring from different sides to get a three-dimensional image.

How does it work?

The biological tissue is illuminated with light and then ‘photographed’. We mostly use dyes with red and near-infrared spectral ranges because this light can penetrate sufficiently deep. The biological tissue scatters and absorbs the rays depending on its composition, which naturally decreases  resolution with increasing depth of penetration. Readings can, however, be taken from different sides to achieve a certain resolution in 3D.

How does this help us?

FMT enables us to study cellular and molecular events in diseased tissue, such as tumours. But the method can also be used to test the targeting and efficacy of drugs in living objects. We have a lot of experience with using fluorescent agents for the treatment and prevention of osteoporosis in order to understand changes in bone metabolism. We are also researching inflammatory reactions of white blood cells to titanium implants. For UKSH’s neurosurgery department, we are researching a novel implant that can release drugs in the body.

  • What equipment or programmes do we have available?
  • We can acquire 2D fluorescence and bioluminescence  images in vivo using the IVIS LUMIA III (2022) and 3D fluoresescence tomographic images using the FMT 2500 (both PerkinElmer, Waltham, MA, USA).
  • The Leica THUNDER Imager 3D Tissue provides high-resolution images of tissue samples.
  • For 2D fluorescence and bioluminescence imaging when imaging organisms with higher safety requirements (BSL-2) , we use our NightOWL II from Berthold Technologies (Bad Wildbad, Germany).

PA – photoacoustics

The sound makes the picture

What is it?

The principle of photoacoustics is evident in its name: we use ‘photo’ (light) and ‘acoustics’ (signals) to ultimately create an image. Using photoacoustics, we can make various structures and functions in the body visible. We can determine the course of blood vessels, for example, and even oxygen saturation in the blood.

How do we do that?

Behind photoacoustics lies a phenomenon we’re all familiar with from everyday life: if you put metal in the sun, it gets hot. In photoacoustics, molecules in the body absorb light energy and convert it into heat. This causes the molecules to minimally expand and send a wave of pressure into their surroundings. We can measure this impulse with an ultrasound device, allowing us to draw conclusions about the tissue of origin. Since every molecule absorbs light of different wavelengths to different degrees, we can use photoacoustics to identify different substances in the body.

How does this help us?

Until now, photoacoustics was mainly used for soft tissue. But we can also visualise the medullary cavity of a mouse bone. We use the method for visualizing inflammatory reactions, too. Because we can also use contrast agents, we can identify acutely and chronically inflamed intestinal tissue in mice. And even vascular dynamics in real-time images. This allows us to draw conclusions about the blood supply. We also see the heart in 4D. This means we monitor the organ in 3D over a longer period of time – this is our 4th dimension.

What equipment or programmes do we have available?

  • MOIN CC is equipped with a Vevo® LAZR from Visual Sonics, based on the Vevo® 2100 platform for combined optical and ultrasound-based imaging. The system offers up to 45µm resolution in real time at 740 frames per second.
  • It is currently being used at MOIN CC in various collaborative projects, for example, for inflammation and tumour imaging, as well as for imaging oxygen supply processes in tissue. We’re also working at MOIN CC with fluorescent-labelled bacteria, which make growth in organs visible.

Hyperpolarised MRI with liquid contrast agents

Tracking down metabolism with quantum technology

What is it?

Hyperpolarised contrast agents are applied quantum technology being developed by us in Kiel. They allow us to watch the metabolism in the living organism in real time – without using any harmful radiation. This method is not an established medical product but is something we are actively researching.

How does it work?

Using quantum mechanical ‘tricks’, we can make selected molecules (the contrast agents) ‘glow’ (on the MRI) – we talk about polarising them. Strictly speaking, the atoms’ nuclear spins (tiny magnetic moments) are aligned, similar to a compass needle. When these molecules are inhaled, swallowed or otherwise administered, we can watch what happens to them in the body – and see where these processes have changed (pathologically).

What can we see with it?

MRI with hyperpolarised contrast agents allows us to measure changes in metabolism that would otherwise be invisible – in real time, non-invasively and in 3D. This means small changes in metabolism can be detected before large macroscopic changes, such as a tumour, are visible. It is also possible to identify at an early stage whether a therapy is effective or not. The added value of this technology was demonstrated in a study on prostate cancer, for example, where cancerous changes in metabolism were found in otherwise seemingly normal tissue (Nelson et al., Sci Transl Med, 2014). Another study showed that the success of breast cancer therapy could be determined only two weeks after the start of therapy (Woitek et al., Radiology 2020) – these are extremely promising results!

What equipment or programmes do we have available?

  • Polarize SpinAligner: dissolution dynamic nuclear polarizer. 1H, 13C, 15N probe heads. Option of optically-induced radicals.
  • Desktop and superconducting NMR instruments for quantification (Magritek SpinSolve 13C and SpinSolve 15N ultra; Bruker WB 400)

ESR – electron spin resonance

In contrast to the widely used nuclear spin resonance (NMR) spectrometers, electron spin resonance (ESR) spectrometers are used to detect electron rather than nuclear spins. Overall, an ESR spectrometer can provide a high density of information about the structure of the sample and its molecular dynamics. Nevertheless, the fact that ESR is not widely used is mainly due to the relative rarity of the unpaired electrons necessary for measurement.

NMR – nuclear magnetic resonance spectroscopy

The starting point for detailed images

What is it?

Nuclear magnetic resonance (NMR) spectroscopy is one of the most widely used and effective methods for studying chemical systems. We use a magnet to create images to determine the structure and interactions of molecules in solution. This shows us processes in biological tissue. We can thus also use NMR spectroscopy to identify a specific molecule. The physics of how NMR spectroscopy works is essentially the same as for magnetic resonance imaging (MRI) used in hospitals.

How do we do that?

Molecules consist of atoms with a certain orientation within the nucleus – their so-called spin. This orientation can be changed by a magnet. When molecules find themselves in an external magnetic field, their nuclear spins interact with the magnetic field and align with it. Each signal in the NMR spectrum has a specific line shape and width, which tell us the duration of the spin reaction and thus form a picture.

How does this help us?

In most cases, NMR is the step before the next step. This is because a weak point of MR-based methods is their very low sensitivity. In other words, detecting molecules present in low concentrations is difficult or impossible. The hyperpolarisation technique can overcome this weakness, allowing us to detect even low-concentration molecules within seconds. To do this, we label the molecules and make them visible. This enables us, for example, to study molecular structures and interactions of pharmaceutical substances.

What equipment or programmes do we have available?

  • High-resolution NMR spectrometer: Bruker WB400 9.4T/400MHz NMR and micro-MRI, 1.5T/m gradients for imaging.
    • Temperature control
    • Life signs monitoring
    • 5mm smartProbe, 10mm broadband coil, 25mm coils: 1H/13C, 1H/2H
  • Desktop NMR spectrometer: Magritek 1T Spinsolve 13C and Spinsolve 15N ultra

Hyperpolarised MRI with gases

Using a quantum mechanical trick to track down lung function

What is it?

Taking pictures of air isn’t easy, even with MRI. Yet imaging the lungs is often extremely important for detecting a disease or monitoring the healing process. Hyperpolarisation comes in handy here, too: using quantum mechanical ‘tricks’, we can make different gases, such as xenon, glow on the MRI. If this ‘hyperpolarised’ gas is then inhaled by someone, we can get excellent images of their lungs, seeing not only where the gas goes (ventilation) but also how much of it is absorbed into the tissue or taken up by red blood cells. We can also take measurements of the microstructure of the alveoli. The method, available at only a limited number of centers worldwide, is unique for all of these reasons.

How do we do that?

To make the noble gas visible, we have to ‘polarise’ it. This means we have to align the gas’s nuclear spins. We do this with a powerful laser that first polarises rubidium, then xenon – in other words, changes them so that they ‘glow’ in the MRI. We then bring this so-called hyperpolarised gas to the patient in a plastic bag. The patient inhales the gas before the MRI examination.

Structure of the xenon polariser: the main components are – (1) QPC Ultra-500 180W laser diode; (2) control unit: vacuum valve, cell pressure valve, flow valve, oven temperature valve, power supply for the diode and diode cooling; (3) coils for the electromagnetic field; (4) optical path with lens; (5) gas distributor; (6) optical chamber for hyperpolarisation; (7) power supply and control unit, vacuum pump; (8) water cooling; (9) optical spectrometer for laser detection.

How does this help us?

One area we use polarised xenon for is medical research. Our long-term goal is for this examination method to also be used in clinical trials. Prof. Wild, an expert in xenon imaging of the lungs from the University of Sheffield, helped develop this method. But xenon is also excellent for studying gas flows – how gases behave in certain environments. This helps us understand gas phase reactions in large reactors. Xenon is unique, too, because of its ability as a large polarisable electron cloud to ‘feel’ the environment. We can use it to study the smallest spaces and channels in solids because we can make them visible.

What equipment and programmes do we have available?

  • Spin exchange optical pumping (SEOP) polariser from Polaris, Sheffield, UK, for the production of hyperpolarised xenon of technical and medical quality.

Histology

Histology: the standard in research

What is it?

Whether it’s the course of a disease or the progress of therapy, histology can help us see the changes in living organisms. It’s a methodology that has been used in research and clinical practice for centuries. In histology, we take wafer-thin samples – of tissue, for example – and examine them under the microscope. We do this again and again over time and see what changes there are in the samples.

How do we do that?

After the samples have been taken, we preserve them. To do this, we dehydrate them and embed them in paraffin or methacrylate. Then we cut the samples into micrometre-thin slices and mount them on a glass slide. In the case of tissue samples, we may sometimes have to shock-freeze them in liquid nitrogen beforehand so that we can cut them better. The next step is to make cells, cell nuclei or individual tissue types visible with the help of dyes. To do this, we use either chemical dyes or antibodies that can bind structures in the samples. We can labell these antibodies with a dye to make antibody binding to individual structures visible.

How does this help us?

Whether it’s tumour burden or iron content, we see everything under the microscope. Including whether and how well therapies work in laboratory animals. We can even examine bones. But because the procedure is time-consuming, we work together with other institutes. We can also, however, examine human tissue to see if our findings are transferable to animals. Last but not least, we can use histology to check how well or how poorly new testing methods work.

What equipment or programmes do we have available?

  • A Leica Thunder Tissue 3D Imager, an inverted fluorescence microscope with a variable excitation wavelength of 365nm to 770nm. Imaging filters are available for DAPI, GFP, Texas Red, Cy5 and Cy7.
  • An automated Leica Histocore AUTOCUD rotary microtome
  • SLEE MPS P1 paraffin dispensing unit

Image analysis

Methods of image analysis, both with and without artificial intelligence, are developed in the Section for Biomedical Imaging. These methods are then available within MOIN CC and our cooperation projects. Research in this area happens in the intelligent imaging lab, the i2lab

To the website of the Intelligent Lab Team

Image Registration

Image registration is an iterative technique where one image is fitted onto another image of the same structure or location using a combination of rotations and translations in 2D or 3D space. Image registration can be applied to images of the same modality at multiple time-points to quantify local structural change (longitudinal or single-modal registration), or can be applied to two images obtained from different modalities at the same time point to link two signals to the same location (multi-modal registration). At MOIN-CC, image registration procedures are developed for both of these applications.

Longitudinal Registration

A registration procedure has been developed to register in vivo micro-CT images in order to produced time-lapse images for the quantification of bone formation and resorption, and to assess the severity of metastatic lesions. The programs have been custom-built in ScancoMedical’s image processing language (IPL) and Matlab.

Multi-modal Registration

Using the software Amira, micro-CT images are registered with fluorescent images to link functional information to specific skeletal locations. For example the kinetic uptake of fluorescent bisphosphonates can be linked to regions of high or low bone tissue mineral density. This has been studied in a MOIN-CC project to develop a method for localized bone turnover quantification using binding kinetics.

What equipment and programmes are available?

  • StructuralInsight
  • MOIN spin library

AI – Artificial Intelligence

Artificial Intelligence detects injuries in the lungs of a COVID-19 patient and marks them in red on the image.

Artificial intelligence (AI): computers learn to think

What is it?

AI stands for artificial intelligence, and this is what it’s all about. Computer programmes take over the work of humans and evaluate images independently. These may be CT scans or x-ray examinations, for example. The aim is to recognise the areas in the images that indicate diseases or injuries. Broken bones are one such example. But a computer programme like this can also recognise diseased tissue, as in the case of the damaged lungs of COVID-19 patients.

How do we do that?

The computer programme must first, just like humans, be fed a lot of information until it can eventually work and think independently. Since this is very time-consuming and tedious when using classical image analysis programmes, we use ‘deep neural networks’ (DNNs). These can automatically recognise features and characteristics in images and thus develop their own experience base. This ability of DNNs to learn features from image data is essentially based on so-called convolutional filters. These are features that a certain section of the image has. The more closely the programme examines the image, the more features emerge that define it. Sometimes we have to correct this evaluation in the beginning to help the programme learn. Eventually, a whole series of deeper and deeper layers with more and more features emerges from a single image. Because DNNs go deeper and deeper into an image, they are also known as ‘deep convolutional neural networks’ (DCNNs).

How does this help us?

One of our main areas of focus is the diagnosis of osteoporotic vertebral fractures. We are working on automated determination of the risk of future fractures of the vertebral bodies and upper neck. In addition to applications for osteoporosis and COVID-19, we have started to adapt our methods to other disciplines, for example, in the study of dental diseases. We are evaluating our models on larger data sets from the University of California, San Francisco and the AGES Reykjavik Study in Iceland. But we can also use our methods to examine zoological samples – including old bone samples to investigate diseases during this period. Cooperations for the automatic analysis of satellite images or underwater scans are also possible.

What equipment or programmes do we have available?

Very different software systems are used for model development, training and inference, depending on the application. In addition to several smaller AI workstations, the Intelligent Imaging Lab currently operates two powerful AI servers:

  • ‘K9’ GPU cluster with 8x NVIDIA V100 SMX2 (Volat GPUs, 16GB each and 14 teraflops)
  • ‘aIfur’ GPU cluster with 4x NVIDIA Quadro A40 PCIe (Ampere GPUs, 48GB each and 37 teraflops)

You have heard about a new method and would like to use it for your project? Contact us!