Guidance – Preclinical Imaging & Testing

Ultrasound and photoacoustic imaging
Positron Emission Tomography
Magnetic Resonance Imaging
x-ray Computerized microTomography

Optical imaging

Understanding fluorescence vs. bioluminescence

In a typical in-vivo fluorescent imaging experiment an injected fluorophore or an endogenous fluorescent protein is excited by an external laser source at a given wavelength and during de-excitation it gives off light at a different wavelength that is detected by the camera of the imaging system (charge coupled device, CCD camera). In a typical in-vivo bioluminescent imaging experiment a chemical reaction between a luciferase enzyme and its substrate results in the emission of light that is detected by the camera of the imaging system. In the case of bacteria expressing lux genes bioluminescence emission does not require exogenous administration of a luciferin substrate. A primer about the differences between fluorescence and bioluminescence can be found here.

Choosing a luciferase enzyme and a substrate

Below are some links from vendors that can get you started:

Choosing a dye

There are numerous online resources to help you choose the appropriate fluorescent dye or protein for your experiment. Below are some links from vendors that can get you started:

Basic imaging workflow

Although fluorescence and bioluminescence imaging are considered easier compared to other in vivo imaging techniques there are plenty of ways to design a poor experiment. This brief guide will help you design a sound imaging workflow.

Minimizing background

If you are aiming for quantitative data you must correct for fluorescence background. Click here for more details on the various sources of fluorescent background. Bioluminescent background is typically very weak and does not affect quantitation. Guidelines on how to properly subtract fluorescent background can be found here.

Finding the optimal acquisition settings (exposure time, binning, f-stop, filters, field of view)

For a detailed description of the different acquisition settings in IVIS Spectrum click here. We highly recommend that first time users set the exposure time setting to Auto which will let the instrument select the optimal settings for the acquisition.

Please be cautious when adjusting the field of view (FoV) and the object height to avoid damage of your object or the camera. This could be the case, for example, if the object height control is set to zero for an object with an actual height of several cm and the selected FoV is small.

Using advanced techniques (spectral un-mixing, tomography)

In many cases your detected fluorescent or bioluminescent signal might be mixed, that is it might contain spectral signatures from different signals including undesired background or secondary light sources. Spectral un-mixing is a technique that allows for delineation of each signal. For utilizing the IVIS spectral un-mixing capabilities click here.

Although imaging using the IVIS Spectrum is 2-dimensional, meaning that what we actually measure is the 2D projection of the light distribution within the volume of the object at its surface, the system is configured to also perform diffuse light imaging tomography (DLIT, in the case of bioluminescence) and fluorescence imaging tomography (FLIT, in the case of fluorescence). These two techniques allow for estimation of the 3-dimensional distribution of the generated light within the object by reproducing an accurate estimate of the animal’s topography.

For information on how to setup a DLIT measurement click here.

For information on how to perform surface topography for DLIT click here.

For information on DLIT reconstruction and analysis click here.

For information on how to setup a FLIT measurement click here.

For information on how to perform surface topography for FLIT click here.

For information on FLIT reconstruction and analysis click here.

Frequently asked questions:

How do I know if my work with IVIS involves fluorescence or bioluminescence?

Simply put, if your animals have luciferase expressing cells you will be doing bioluminescence experiments. If you have tagged your vehicle with a fluorophore or your animals have fluorescent protein expressing cells you will be doing fluorescence experiments. To better understand the difference between fluorescence and bioluminescence click here.

In bioluminescence how much luciferin do I have to inject into the animal and how long should I wait before acquiring data?

The vendor spec sheet that comes with your purchased luciferin usually has a preparation protocol. A guide on planning a reproducible bioluminescence measurement can be found here.

Can any type of mouse be imaged with the IVIS Spectrum?

You will get the best results with hairless mice. White furred mice can also be imaged, but it is best to remove the hair from the area of interest. Dark furred mice will strongly attenuate light and if the pigment of the skin underneath the fur is also dark the results will be significantly compromised even after hair removal.

What fluorescent dye should I choose for my experiment?

That strongly depends on what you are trying to measure. Click here for a guide. Susceptibility to photobleaching, pH sensitivity, an adequately broad Stokes Shift (wavelength difference between peak absorption and emission of a fluorophore), a high extinction coefficient and a high quantum yield are factors to consider. All of the above properties are described here.

The IVIS system doesn’t seem to initialize properly. What do I do?

If you get errors during initialization, in many cases restarting the Living Image software will solve the issue. If not, please contact PMIT at aiptcore@mit.edu for assistance. If the temperature bar turns red during initialization you can check on the camera and stage temperature by clicking on the bar. If the temperature is high but is steadily dropping, wait a few minutes until it reaches the desired -90 deg value, the bar will turn automatically green. If this is not the case please contact PMIT at aiptcore@mit.edu for assistance.

I see a strange ring of high intensity in my fluorescent image. Why is that?

This artifact is a result of choosing excitation and absorption filters with a difference of less than 30 nm. To mitigate it, choose a filter pair with a wavelength difference larger than 30 nm, or let the Living image software determine the most appropriate filter pair.

How do I determine what exposure time to use for my experiments?

If you are doing the specific experiment for the first time, we highly recommend setting the exposure time to Auto. That will allow the system to choose the optimum acquisition settings for your experiment. Click here for a description of the different acquisition settings. Click here for details on how these settings affect detection sensitivity.

How do I choose among the different values for binning and f-stop?

My animal has a lot of fluorescent background, how can I mitigate that?

You have a lot of options. Depending on the source of the background you can alter the animal’s diet and/or perform ROI based background subtraction or spectral un-mixing.

Will the placement of the animal in the platform of the IVIS instrument affect my data?

Yes, it will. In general, if the approximate origin of the light source within the animal is known, you should position the animal so that the amount of tissue between the light source and the IVIS detector is minimized.

During my measurement I get a message that my image contains saturated pixels, what do I do?

This message indicates that the maximum number of counts collected per pixel is larger than 60000 which is out of the linear range or operation for the digitizer of the CCD signal. You can

1. reduce the exposure time

2. use smaller binning and/or

3. use larger f-stop.

What is the correct unit for my data? Counts, radiance or radiant efficiency?

Counts is an uncalibrated unit that depends on the acquisition settings. It is useful for checking whether your data falls within the linear range of the CCD operation. For quantifying your data use Radiance for bioluminescence and Radiant Efficiency for fluorescence. Click here for more details.

I want to draw a region of interest (ROI) and quantify the signal, what should be the shape and size that I should choose?

The minimal background in the case of bioluminescence favors large ROI’s, while the considerable background in the case of fluorescence favors an ROI shape and size closely fitted to the biology that you are observing. Click here for more information on the effect of ROI size and shape on analysis.

Can I image two different fluorescent dyes at the same measurement?

Yes. You can use spectral un-mixing to differentiate between the two.

Can I image bioluminescence and fluorescence at the same measurement?

If you have both a fluorophore tag and luciferase expressing cells within your animal, make sure to perform fluorescence before bioluminescence measurements. The reason is that the significantly longer time constant of bioluminescence emission might interfere with the fluorescent emission which is faster.

My image looks blurry, what do I do?

A blurry image can indicate too coarse binning or maladjusted focus. Click here for a description of the different acquisition settings.

There are different fields of view (FOV) to choose from, what is the best choice for my experiment?

Click here for a description of the different acquisition settings. You typically want to adjust your FOV according to the size of your object.

The spec sheet of my fluorophore has peak excitation and emission wavelengths that I can’t find among the available filters of the IVIS instruments. What do I do?

You can choose the filter pair most similar to the one in the spec sheet or let the Living image software determine the most appropriate filter pair.

Is there a way to calculate the fluorescent/bioluminescent light distribution within my animal in 3D?

For information on how to setup a DLIT measurement click here.

For information on how to perform surface topography for DLIT click here.

For information on DLIT reconstruction and analysis click here.

For information on how to setup a FLIT measurement click here.

For information on how to perform surface topography for FLIT click here.

For information on FLIT reconstruction and analysis click here.

Is my IVIS data quantitative?

Sort of. What you measure is a 2D projection on the object’s surface of the light distribution within the object’s volume. In addition, light diffuses and attenuates when it encounters tissue.

Ultrasound and photoacoustic imaging

Power point reference for basic system operations

Click here

Understanding ultrasound imaging

Ultrasound imaging is a real-time high-resolution imaging technique that is based on contrast generated when ultrasound waves scatter and attenuate in different tissues. In ultrasound imaging the main detector element is a piezoelectric transducer, that generates ultrasound waves when a modulated voltage is applied across its terminals. The waves travel through different tissues differently as determined by each tissue’s acoustic impedance. Based on the acoustic impedance of each tissue, part of the ultrasound wave will be transmitted through the tissue and part will be reflected back to the transducer. The transducer will convert mechanical oscillations back to electrical signals from which an image will be created.

Understanding photoacoustic imaging

As opposed to pure ultrasound imaging where the transmitted and received signals are both ultrasound waves, in the case of photoacoustic imaging the transmitted signal is light and the received signal is ultrasound. Click here for a depiction of how the photoacoustic effect works. Photoacoustic imaging seeks to address the issue of light diffusion and attenuation that is common in optical (fluorescence and bioluminescence) imaging.

Ultrasound contrast agents

The typical contrast agent for ultrasound imaging is perfluorocarbon gas filled microbubbles. As opposed to linear contrast created by tissues depending on their echogenicity, microbubbles create non-linear contrast in ultrasound imaging. Their size is a few microns that allow them to stay in the blood vessels.

Photoacoustic contrast agents

A comprehensive primer on photoacoustic contrast agents and their properties can be found here. As mentioned in our page about fluorescent dyes, a good dye has high extinction coefficient and high quantum yield. On the contrary, a good photoacoustic dye has a high extinction coefficient but a low quantum yield so that the absorbed optical energy is converted into heat instead of light.

The different imaging modes that our ultrasound and photoacoustic instruments can offer are summarized here.

Frequently asked questions:

The instrument has many transducers, which one should I use?

The Vevo3100 system is equipped with a variety of transducers all of which are described here. Each transducer has a different center frequency for the emitted ultrasound. Higher frequency translates to higher spatial resolution but also to higher attenuation thus to smaller imaging depth. If you are imaging large animals a transducer with smaller center frequency should be chosen compared to imaging small animals.

How should the transducer be positioned with respect to my object?

The imaging plane in a typical 2D ultrasound is parallel to the ultrasound field emitted by the transducer. The orientation of the transducer will depend on the imaging plane that needs to be visualized. The PMIT staff will help you with transducer placement during training.

I want to measure blood flow, what imaging mode do I use?

You can use the power doppler imaging mode.

I want to measure blood directionality, what imaging mode do I use?

You can use the color doppler imaging mode.

Can I image in 3D with ultrasound?

Yes, the VevoLAZRX system is capable of 3D acquisition. The transducer can be mounted in a translation stage that automatically moves within a user defined range to provide volumetric information.

What is the smallest structure that I can image?

Our system is equipped with transducers that have a range of frequencies, spatial resolutions and imaging depths. Please click here for choosing the appropriate transducer for your application.

I want to image the heart and calculate morphological and functional properties, what imaging mode do I use?

You can use the M imaging mode and the anatomical M imaging mode. Click here for an overview of our system’s capabilities in cardiovascular research.

Can I image tumors with ultrasound?

Yes. Click here for a typical oncology study workflow with ultrasound in B imaging mode. In addition, you can use photoacoustic imaging to characterize the tumor microenvironment.

Both microbubbles and doppler image blood vessels, what’s the difference?

The difference is in the size of the of the vessel. Microbubbles can be used to visualize capillary-sized vessels while the doppler imaging mode can be used for larger vessels.

Can fluorescent dyes used for IVIS imaging be used as photoacoustic contrast agents?

Both fluorescent dyes and photoacoustic agents need to absorb the incoming light effectively, however the favorable dissipation mechanism in each case is different: In the case of fluorophores the absorbed energy should be converted to light of a different wavelength and in the case of photoacoustic agents the absorbed energy should be converted to heat. Although there are fluorophores that can be seen both by IVIS and PA (i.e. ICG), fluorescent dyes with high quantum yield are not appropriate for photoacoustic imaging.

I want to inject cells in my mouse, can I use ultrasound to guide the needle?

Yes. Our system has a mechanical support for guiding needles. Please contact us for initial consultation and training on ultrasound guided injections. Click here to see examples of image guided injections that can be performed with our system.

I have some nanostructures that I think they could be really good photoacoustic contrast agents. How can I test them?

Our system can operate in Spectro mode and characterize your contrast agent across a large wavelength range.

It’s hard to understand what I see in the ultrasound image, what do I do?

Ultrasound images require a trained eye for interpretation. Our staff will work to train you to identify your structure of interest. In addition to personalized training we will be working on training algorithms to do the hard work for you. Stay tuned for more details.

Positron emission tomography

Understanding PET

PET is an imaging modality that takes advantage of radionuclides emitting an elementary particle known as a Positron. The acronym PET stands for Positron Emission Tomography. Using a positron emitting isotope to help generate a 3-dimensional image.

What is a Positron?

Going back to what we’ve learned in high school physics about basic types of radiation – Alpha Particles, Beta Particles, and Gamma Rays. In laymen’s terms, positrons are beta particles that are positively charged as opposed to most beta particles that are negatively charged.

How does this work?

These “positive” beta particles do something special when they eventually encounter an electron – they annihilate!! They cease to exist…. Positrons are antimatter to electrons. When this annihilation event occurs, energy is released in the form of two gamma rays travel exactly 180 degrees from each other. We can take advantage of this by placing detectors all around where this emission of the positron could happen. The two opposing gamma rays hit detectors at the same energy of 511keV and with enough strikes we can track where this annihilation event happened in 3D space. Since we’re dealing with detecting gamma radiation created by this annihilation event – we don’t have to worry too much about sensitivity. Gamma radiation will pass right through nearly anything in its way.

How can I make this work for me and my studies?

Imagine that you had something that could target a tumor in the body of a mouse – say an antibody. That antibody could be labelled with a PET isotope and when enough has accumulated in and around the tumor (its constantly emitting positrons) we could generate a 3D image of where the annihilation event occurred! And because it is possible to capture nearly all these events – it is possible to generate real quantitative data of how much of what was administered to the mouse got to your tumor or region of interest and other regions as well. Depending on the half-life of the PET isotope; many multiple timepoints can be taken of the same mouse until biological equilibrium has been reached. With PET we can get imaging information on PK/PD, on / off targeting, clearance over time.

What’s the catch?

Like all other imaging modalities there is a tradeoff between sensitivity and resolution. In the case of PET imaging being one of the most sensitive imaging modalities; it also has a poor spatial resolution. This is due to the distance the positron travels until it encounters an electron it will annihilate with. Depending on the positron energy this positron travel can range but is typically between 1-2mm.

Another issue with PET is throughput. Enough annihilation events need to be collected in order to generate a 3D image. Depending on the amount of activity administered, a typical collection window of 10 minutes is required for most mouse whole body studies. A 3-minute CT is then run for anatomical reference. About 15 minutes total for one mouse at one single timepoint. For larger number longitudinal mouse studies this may be difficult to fit in appropriate number of mice in for needed timepoints.

What isotopes to use:

There are three isotopes that are currently available for to use on your studies. Fluorine 18 in the form of FDG. Copper 64 and Zirconium 89 – Radiometals used for molecule labeling.

FDG is “fake” glucose that is currently used in clinical PET to diagnose and monitor treatment of cancer. Many cancer cell lines will uptake glucose at a rate many times “normal” cells. FDG will accumulate in these high metabolizing cells and show higher signal in tumors of interest for most applications. F18 the isotope in FDG has a 2hr half-life and realistically is only limited to single timepoint imaging and not longitudinal studies.
Copper 64 is a radiometal that can be attached to chelators like NOTA and NODAGA. It has a 12hr half-life and would ideally be paired with small molecules that have short blood circulation kinetics. Cu64 atom itself will accumulate in the Liver.
Zirconium 89 is a radiometal that can be attached to the chelator DFO aka Desferal. It has a much longer half-life than Cu64 at 3.3 days. IgG or larger molecules with longer blood circulation kinetics would be ideally imaged using Zr89. Zr89 atom itself will accumulate in the bones and joints as it competes with Calcium.

Example case studies:

Tumor treatment studies, Cancer Progression Regression – Use FDG before you start treatment and weekly during treatment. You can compare images in the same mice to see if your treatment is effective or not.

Targeting ROI w/ Small Biologics – Attach an appropriate chelator and use Cu64 or Zr89. Adding a chelator will add 0.5kD+ mass which may affect the kinetics of whatever is being labeled.

Targeting ROI w/ Larger Biologics – Attach DFO and use Zr89. May have to add more chelators per molecule of interest to get enough signal to target.

FDG Guide:

FDG can be ordered any day. It is delivered ready to inject from PET Net at a maximum amount of 15mCi per dose. The volume delivered is typically 1-2.5ml. Because F18 has a half-life of 2hrs, the same amount of activity to be injected in a mouse at the start vs two hours later will be double the volume. Thus we generally limit studies to be about 20mice per dose per day as our activity rapidly decays and volumes increase.

FDG looks like Glucose and because of this is going to be taken up by nearly all cells in the body. Once in cells FDG is trapped and is soon chopped up and passes through to urine. What we want to measure is what the amount over background tissue is the FDG taken up by cancer cells / tumors since they will have much higher demand of sugar than normal cells / tissues. To make sure we get good signal to noise images:

1.Fast the animals for a minimum of 4 hours and ideally overnight. This may not be possible depending on how sick animals are while on study.

2.Anesthetize animals during injection and uptake. This step is taken so that we have low muscle uptake and thus more bio availability of FDG for other cells to take up.

3.Uptake FDG for 1 hour before scanning mice. FDG needs to be taken in by cells and trapped in cells before imaging. This takes about an hour.

4.Keep animals warm during uptake. This will keep brown fat uptake low.

Several of these parameters for FDG could be omitted from being done depending on your study and health of mice.

Other caveats to be aware of:

Heart, Kidney, Bladder will have high signal not matter what is done. If tumors or tissues of interest are near these organs they may be difficult to elucidate. IE Small Lung tumors around the heart area.

Some tumors are NOT FDG AVID. They won’t uptake FDG like normal cells would and therefore before doing an FDG study please see if your cell line is FDG / Glucose AVID.

Radiometal Labeling Guide:

Radiometals such as Zr89 and Cu64 can be attached to molecules such as antibody like fragments or other novel molecules to see PK/PD in mice models. Attachment of these radiometals are done via bifunctional chelators where one part will attach to your molecule of interest and the other to a radiometal. It is important to keep in mind that the attachment of the bifunctional chelators must be such that it doesn’t affect the kinetics of the molecule. Because of the additional mass of the chelator added to the molecule can change the circulation kinetics and where the chelator attaches to the molecule can change the binding kinetics of your molecule. Although the more chelators that are attached to the molecule the stronger the imaging signal that will result but perhaps at the cost of a significant change of important kinetic information vs the molecule alone. These things must be considered when thinking about radiolabeling. The attachment of these bifunctional chelators is to be done by the investigator. The radiolabeling of the radiometal to the Molecule-Chelator construct is done by the core.

Cu64 and Zr89 can be delivered on Tuesday or Wednesday. Tuesday delivery is made at the University of Wisconsin. Wednesday delivery is made at Washington University in St Louis. Wednesday delivery is preferred due to the lower cost of activity.

Cu64:

Copper 64 has a 12hr radiological half-life and is idea for labeling smaller molecules that quickly circulate and clear from the body. Because Cu64 itself when injected will directly enter the liver; it is not ideal for liver tumor imaging without the appropriate controls. Eventually however breakdown products of molecules labeled with Cu64 will end up in the liver. Some chelators that are ideal for Cu64 are NOTA and NODAGA.

Zr89:

Zirconium 89 has a 3.3 day radiological half-life and is ideal for labeling larger molecules that take time to clear for the body. When Zr89 is directly injected it will enter into bone / joints due to the body thinking its Calcium. Thus, this isotope is not ideal for any bone tumor imaging. Eventually the breakdown products of the molecules labeled with Zr89 will end up in bone / joints and also liver. The ideal chelator or use for Zr89 is DFO / Desferal / Deferoxamine.

Questions? Please free to contact us at AIPTcore@mit.edu for PET experiment consulting and planning and feasibility.

Magnetic resonance imaging

Understanding MRI

MRI (magnetic resonance imaging) exploits the nuclear magnetic alignments of different atoms inside a magnetic field to generate images.

MRI machines consist of large magnets that generate magnetic fields around the subject. These magnetic fields cause paramagnetic atoms such as hydrogen, gadolinium, and manganese to align themselves in a magnetic dipole along the magnetic fields, created by the radiofrequency (RF) coils inside the MRI machine. The machine captures the relaxation of the atoms as they return to their normal alignment when the RF pulse is temporarily ceased.

The returning signals are converted into images by a computer based on the resonance characteristics of different tissue types.

Imaging of any part of the body can be obtained in any plane.

Advantages and Disadvantages of preclinical MRI

Advantages:

MRI can characterize and discriminate among soft tissues using their physical and biochemical properties (water, iron, fat, extravascular blood and its breakdown products)
any plane can be imaged
blood flow, contraction and relaxation of organs, both physiologic and pathologic, can be evaluated
MRI is noninvasive and does not use ionizing radiation

Disadvantages:

expensive
MRI is sensitive to motion (motion artifacts)
extreme precautions must be taken to keep metallic objects away (safety issues, image artifacts)

Most common MRI modalities

Anatomy
Dynamic contrast-enhanced MRI (DCE MRI)
Functional MRI (fMRI)
MR spectroscopy (MRS)
Diffusion weighted and diffusion tensor imaging (DWI/DTI)
X-nuclei MRI/MRS (¹⁹F, ³¹P, ¹³C, etc.)
Most common MRI Sequences for anatomy and their Approximate TR and TE times

	TR (msec)	TE (msec)
T1-Weighted (short TR and short TE)	500	12
T2-Weighted (long TR and long TE)	2500	30
Proton Density Weighted (long TE and short TE)	2000	12
Flair* (Very long TR and TE)	9000	110

* Flair: FLuid Attenuated Inversion Recovery sequence

Image contrast

Tissue	T1WI**	T2WI**	PDWI**	Flair
Fluid	Dark	Bright	Bright	Light
Muscle	Gray	Dark Gray	Gray	Gray
Spinal Cord	Gray	Light Gray
Fat (subcutaneous tissue)	Bright	Light	Light
Air	Very Dark	Very Dark	Very Dark	Very Dark
Inflammation (edema, infarction, demyelination)	Dark	Bright

** T1WI: T1-weighted image; T2WI: T2-weighted image; PDWI: proton density weighed image

Gadolinium
Iron oxide
Iron platinum
Manganese
¹⁹F – imaging hypoxia

Most commonly used MRI sequences for anatomy

Gradient echo
Spin echo
Fluid attenuated inversion recovery (Flair)

Useful links

x-ray computerized microtomography

Understanding microCT

microCT is an X-ray-based imaging technology that allows the user to non-destructively image the interior of a specimen. The Bruker Skyscan 1276 at the Koch Institute is designed to accommodate anesthetized animals such as mice and rats. It can also image inanimate, ex vivo samples with a maximum nominal resolution of 3 microns. With this type of in vivo system the sample remains in a fixed position while the x-ray source and detector rotate around the sample. During this rotation many images (from 200 to about 1200) are taken from different angles. The resulting tomographic reconstruction produces a 3D volume that can be viewed in any arbitrary virtual slice. Typical scan times range from one minute to one hour.

Although the maximum nominal resolution is 3 microns, this is never used in vivo. The field of view at this resolution is less than 12mm and the appropriate holder does not accommodate a sample the size of a mouse. Typical in vivo imaging with this system is done at 20-40 micron resolution.

Identification of a structure also relies on contrast in the image. In the case of x-ray tomography, the contrast is the result of a difference in x-ray attenuation of different materials in a specimen. This attenuation can be due to both the physical density (mass/volume) and the atomic number of the nuclei of the constituent atoms. Materials like iodine, calcium, gold, etc. attenuate strongly and can provide excellent contrast. For this reason microCT is often used to visualize the hydroxyapatite in bone, which provides significantly more attenuation than the surrounding soft tissue.

A primer about microCT can be found here.

Choosing image acquisition settings

High resolution images with low noise will take much longer compared to a lower resolution scan with more noise. The best settings for you depend on the material you are interested in imaging and what kind of information you want to extract from the images.

There are a number of preconfigured protocols provided with the microCT that are appropriate for most imaging applications. In general, the higher the resolution, the more noise will be present in the image. This manifests as grainy, ‘salt and pepper’ noise that can obscure object boundaries. A larger x-ray dose (more photons) will reduce this nose, although this may not be desirable for the animals.

Striving for an x-ray dose that is ‘As Low As Reasonably Achievable‘ it is recommended that a 20-50 micron resolution is used unless there is a strong scientific need to image at higher resolution.

Another parameter than can be varied is the x-ray tube voltage. This determines the range of x-ray photon energies produced by the system and can have a small affect on image quality. When imaging dense materials such as metal it can be helpful to use the highest voltage possible, in this case 100kVp. For imaging soft tissue (lung tumors, for example), a lower voltage like 50-70kVp is more appropriate.

If a rough image is acceptable (gross bone morphology, body composition measurement, identification of large lung tumors, etc.) then a 60-100 micron image with a very fast scan time may be appropriate.

X-ray beam filtering is also adjustable. When imaging metal in particular it can be advantageous to adjust the filter to remove lower energy photons and reduce beam hardening artifacts. Similarly, circular vs helical scans are worth comparing for some cases, including metal implants.

During training on the microCT system these details can be explored and an appropriate combination of settings can be determined.

Choosing a contrast agent

There are many commercially available contrast agents to aid researchers in visualizing vasculature (e.g. Fenestra, Aurovist, Exitron, Omnipaque). Specialized, experimental materials for targeting tumors or specific organs are also available. Simple iodine-based small molecules like iohexol (found in Omnipaque) can be as little as a dollar per dose (for a 20g mouse). More exotic nanoparticle-based materials such as Aurovist are closer to $50 per dose. While iohexol is cheap, it only remains in the blood pool for a limited time, and is rapidly cleared by the kidneys. There are also potential adverse events associated with some contrast agents. It would be a good idea to discuss your imaging goals with core staff prior to purchasing your materials. Also remember that any materials you use in vivo need to be included in your CAC protocol.

Things to do before you image your specimen

Plan ahead. Understand what it is you hope to get out of your images. If you need a qualitative image of your sample, it is likely that almost any combination of imaging parameters will work fine. If you have specific quantitative measurements in mind, consider the size of the features of interest and the relative contrast of the material of interest as compared to its surroundings. Do you need a high-resolution scan? Do you need a contrast agent? How you prepare your ex vivo samples, how mice are restrained, etc. will all affect image quality and your ability to quantify it.

Discuss your analysis needs with core staff prior to imaging. Determining the most appropriate method for analysis (what quantities are to be measures, what software to use, etc.) can help plan the image acquisition. For example, if the contrast between the tissue of interest and surrounding material is very low, a much longer scan or lower resolution scan may be needed. Some analysis software may work well importing DICOM files, but not tiff files. Knowing this prior to reconstruction will save time.

Working out these details in advance can save many hours later. For example, if you are imaging calvarial bones in mice, a simple bite-bar or head restrainer can help align the skulls. In this way each of your longitudinal scans quickly, easily, accurately align for accurate, nearly effortless analysis. Carefully aligning the skulls prior to scanning may take 10 seconds per mouse. Co-registering poorly aligned skulls later can require additional software, take a long time, and yield mediocre results.

If you are using a contrast agent, you should characterize it prior to use. Changes in the x-ray tube voltage and beam filtering can improve your ability to quantify small vasculature in tumors, for example. If you have metal implants, try imaging one in a cadaver first to see if helical scans or other modifications improve your images.

You are ready to begin once you have planned the imaging portion of your study and determined the best parameters for acquisition and reconstruction.

Minimizing noise/movement artifacts

Lab tape is very cheap. Using it to restrain your specimen can save a lot of time by preventing the need to re-image animals that have moved during scanning. A shift of 1mm is 20 voxels in a 50 micron scan. Even a very small shift in your sample can cause a dramatic blur and degradation of the final reconstruction. For an anesthetized animal, consider how breathing will affect the area of interest. In some cases supine is better than prone, and vice versa. For example, if you are interested in the spine, a prone subject will breath, causing the spine to move up and down while being imaged. A supine subject’s spine will not move, but breathing will cause the sternum to move up and down.

Finding the optimal acquisition settings (exposure time, binning, voltage, current)

The Skyscan 1276 control software is configured with several presets based on beam filtration. In general, larger and denser objects will require more filtration. For mice, the .5 or 1.0 mm aluminum filter is usually appropriate. For rats or larger samples, the aluminum + copper filter is more likely to yield the best results. Contact core staff for any customization that may be needed. There is some trail-and-error involved and the best possible image quality will depend on the sample being imaged and the goal of your analysis. There is always a tradeoff between image quality and imaging time. There will also be parameters involved in the reconstruction that can help improve image quality.

Using advanced techniques (respiratory gating, stitching, multi-modality coregistration)

Respiratory gating can help improve image quality of the lungs for anesthetized animals. In the case of mice, this can be helpful, but it also increases acquisition time and x-ray dose. If the increased image quality is worth spending 5x a long imaging, then respiratory gating may be right for your application. Applications such as multimodality coregistration require an animal holder that is compatible with both imaging systems. Consult core staff about appropriate options.

Frequently asked questions:

I have an amazing new nanoparticle! Can I image it in the microCT?

Probably not. Commercially available contrast agents like Aurovist are provided at a very high concentration. In the case of Aurovist, the 40mg of gold nanoparticles are suspended in 0.2mL PBS. Is your gold concentration 200 mg/ml? If so, the solution is black and you can expect good contrast with soft tissue when imaged by microCT. Is your nanoparticle solution clear like water? If so, it is very unlikely you will be able to see it in vivo, even if it accumulates in tumors or a small organ (e.g. lymph node). Prior to planning any in vivo experiment, try imaging a small vial of your highest concentration of particles next to a vial of water. Is the attenuation different?

I have a metal implant in my animals! Can I image it in the microCT?

Maybe. With the Skyscan 1276, metal causes some image artifacts that can prevent images from being used quantitatively, and in some cases even qualitatively. If you need to see the growth of bone on a titanium screw, the interface between bone and metal can be hard to visualize accurately. The interior of porous metal structures are challenging to image. Using the helical (spiral) scan mode can help minimize artifacts due to the metal and provide qualitative images of your specimen. Quantification of bone densities, for example, may not be possible in close proximity to metal.

Why are my mouse images so blurry?

A common problem is that the mouse moved during scanning. A typical scan may take anywhere from two to fifteen minutes. If the animal moves during that time, there will be significant image artifacts. This can manifest as a blurry, hazy image if there was continuous motion (perhaps due to breathing). Alternatively, it may result in a sort of double image where you see, for example, two sets of ribs in a mouse. This is because the mouse shifted halfway through the scan and you see a somewhat faint image of where it was during each half of the scan. This can be because the mouse itself moved due to insufficient anesthesia, or the mouse holder moved due to an unsecured component of the holder. Be careful to properly restrain both the mouse and any custom holder being used to isolate the area of interest.

An additional concern with the Skyscan 1276 is the misalignment correction done during reconstruction. In many cases the initial guess provided by NRecon is accurate (if you are imaging a mouse in the standard mouse holder). When it guesses wrong (common when imaging rats), and you don’t manually fix it, the result can be a blurry image. Double check the misalignment value and see if this fixes your problem.

What is the highest resolution I can get with this scanner?

Resolution is a bit complicated. There is the so-called “nominal resolution” which is simply the size of the volume elements (voxels) being reconstructed. This will typically be 40 micron for an in vivo mouse scan. This 40 micron value does not mean that you can clearly resolve the size/shape/density of a 40 micron object. The ability to distinguish adjacent objects will also depend on their density relative to the background material (e.g. bone in muscle or a tumor in aerated lung). We have resolution phantoms that can be imaged using your scan parameters (voltage, current, filtration, binning, exposure time, etc) to determine what the actual resolution of your scans are. That said, the nominal resolution of this system goes down to 3 micron. This is only achievable for small samples (12mm or smaller in diameter). Higher resolution also requires more x-rays to achieve a given contrast-to-noise ratio. This means a much higher dose of ionizing radiation to your mouse. For that reason, in vivo scans are typically limited to around 20 microns.

Can I reconstruct my data on computers other than the one the runs the scanner?

Yes. Bruker supplies NRecon for free and it can be run on any Windows computer. NRecon is the client program that allows you to control the reconstruction parameters. It will use a server program running in the background to do the actual reconstruction. You can choose to use either the NReconServer or the GPUReconServer. The GPU accelerated version is much faster, but requires an NVIDIA brand graphics card. In informal tests, an NVIDIA GTX480 (about $80) takes about 20-30% longer to reconstruct a typical lung CT scan as compared to a Quadro RTX 5000 ( $2300). If you want to reconstruct your data at home, save some money and stick with an old graphics card. For large, high resolution scans it may help to have more memory on the graphics card, which could skew the performance results towards the more expense card (16GB vs 1.5GB).