Investigations with Imaging
Imaging reveals processes and events in living matter by showing the metabolic pathways a molecule travels in cellular structures or in animals, by discovering the cellular events of a disease, or by finding and identifying a missing protein.
With radio-labelled molecules, Positron Emission Tomography (PET) or Single Photon Emission Computed Tomography (SPECT) can track events and target proteins in a cell, on a cell and outside a cell.
Imaging also enables a personalized approach to the detection, evaluation, and management of disease with earlier, more accurate diagnoses and safer, more effective treatments.
Our imaging capabilities are grouped into three areas of activity:
Analysis of radiotracer uptake (in vivo PET/SPECT, biodistribution)
Radiotracer uptake is often expressed in percent of injected dose per weight of wet tissue (%ID/g). In in vivo PET/SPECT studies, Volumes or Regions of Interest (VOIs/ROIs) are drawn and regional activity data (Bq/cm3) is extracted from these VOIs/ROIs. The Time-Activity Curves are derived from the calculated data using image analysis software.
Kinetic modelling may be involved if sophisticated data is required for compartmental uptakes or clearance. In biodistribution studies, %ID/g values are calculated to express the uptakes of radiotracers in organs of interest. Ratios of target:non-target or organ:background uptakes are often calculated to serve as indications of radiotracer selectivity/specificity and clearance.
Contact scientist: Vu Nguyen
Radioligand stability studies
In vitro radioligand stability studies are often conducted in physiological solutions or body fluids (plasma/blood), mimicking physiological conditions to establish the tracer stability over time. In vivo radioligand stability studies are performed to examine the unchanged/changed fractions of the tracers in relevant organs or tissues, often the targets of the tracer. Both techniques involve activity extraction and/or analytical methodologies such as radio-HPLC, radio-TLC or appropriate chromatography methods.
Complex dynamic imaging in PET
Dynamic imaging in PET provides a wealth of information over a typical static acquisition. Dynamic imaging tells the story of what is happening within the subject over an extended period of time. This is used to develop a picture of the biological interactions which take place from the moment of injection over multiple time frames, rather than just a single timeframe in a short static scan. With some tracers there is no adequate moment of their course in the body which a static PET image represents the biological process under investigation. However a full dynamic study coupled with the analysis of the kinetic in each voxel or defined region of interest extracts the parameters of interest. When combined with other capabilities such as blood sampling, kinetic modelling, Monte-Carlo simulation, a comprehensive view of the processes being investigated is obtained.
|Contact scientists: David Zahra|
Atlas-based or ROI(region of interest)-based quantification in multiple organs
- Image-based registration of Mouse and Rat brain atlas onto the PET images
- Mesh-based registration of a whole-body mouse atlas onto the PET images (via the CT image)
- Automated extraction of the Time Activity Curves using the atlas or user-drawn region of interests (Anatomist/Brainvisa, IRW, Amide)
|Contact scientist: Marie-Claude Gregoire|
Optimisation of the quantification using Monte-Carlo simulated data
In nuclear medicine, the design, optimization and validation of image correction, reconstruction and analysis techniques often relies on the use of physical phantoms or software simulated phantoms whose geometry and contents are precisely known. Simulated data provide a greater flexibility compared to actual phantom data as they allow the simulation of realistic biological studies. The input model can be derived from actual biological data such as a segmented MR image for the anatomy and time activity curves generated from measured kinetic parameters for the function. In addition, the ability to control the factors that degrade the image is a significant advantage over the use of images based on physical phantoms.
Simulating realistic data in Positron Emission Tomography studies involves
- generation of the raw list-mode or projections data from an anatomical and functional models and
- the correction and reconstruction of the data using standard algorithms to yield PET volumes presenting the same characteristics as clinical volumes.
PET-SORTEO is a Monte Carlo-based simulator which enables fast generation of realistic PET data. This simulation model is thoroughly validated and used for the geometries of the Ecat Exact HR+ human scanner and for the R4 and P4 pre-clinical scanners. Recently this simulation tool was adapted and fully validated for the geometry of the Siemens Inveon pre-clinical scanners. This virtual acquisition platform includes the simulation of the data and their correction and reconstruction following the NEMA NU-4 standard.
PET-SORTEO is currently being intensively used by the quantification group and collaborators as a tool for designing/optimizing and validating data correction/reconstruction and analysis processes such as:
- Image reconstruction
- Image registration
- Partial Volume Effects correction
- Statistical analysis
|Figure 1: (a) Coronal slice of a simulated raclopride scan, (b) and (c) zooms of the brain extracted from a transverse and a coronal slice, respectively. (d) Graph showing the detectability of receptor density variation in a raclopride simulated experiment as a function of the number of scans included in each group.|
|Contact scientist: Marie-Claude Gregoire|
Modelling Partial Saturation Approach
For the study of neurodegenerative disease, the Partial Saturation Approach (PSA) is a promising method for the quantification of receptor binding parameters (receptor density, B'max and radioligand affinity 1/KD). The PSA has the advantage over other existing methods by deriving the B'max and the Kd individually from a very simple acquisition protocol.
An improved version of the PSA method has been developed which is able to derive a data-guided optimal time window for the analysis instead of the standard PSA method in which a static window must be initially defined by the user. The method has been intensively validated using simulated data and studies (PET-SORTEO) and the results show that the new method produces more robust and accurate parameter estimates.
Contact scientist: Marie-Claude Gregoire
Partial Volume Effects correction and data de-noising in PET
Quantitative measurements in dynamic PET imaging are usually limited by the low spatial resolution of the detection system (1.5 mm and 5 mm for pre-clinical and clinical scanners respectively). The magnitude of the resulting partial volume effects (PVEs) depends on the system resolution and the size of the region of interest. Depending on the application, bias introduced by PVEs can be severe, mainly affecting small and intricate structures, as well as voxels on region boundaries. In addition to limited resolution, quantification may also be affected by the poor counting statistics particularly in short dynamic frames. Low counting statistics can introduce statistical noise in the reconstructed images which further degrade their visual aspect. This makes region identification difficult and challenges the robustness of the kinetic parameter estimation methods.
We have designed an original method for the restoration of dynamic PET images which suffer from both PVE and noise degradation. The method is based on a weighted least squares iterative deconvolution approach in the wavelet space with temporal regularization of the wavelet coefficients using basis functions. This method has been successfully validated using simulated dynamic PET data generated with the PET-SORTEO simulation program for the Inveon pre-clinical scanner.
Contact scientist: Marie-Claude Gregoire
Microscopy (Structural/functional tissue analysis using histology / immunohistochemistry)
One of the strengths of multimodal molecular imaging is the ability to relate imaged functional changes to histological data. Many biomarkers of pathological changes are identified using histological approaches such as immunohistochemistry (for detecting proteins), in situ hybridisation (for detecting mRNA of specific proteins), and histological stains (for lipids, cell matrix, etc).
|Immunohistochemistry is used routinely with our group to localise biomarkers of neuroinflammation. The process of microglia activation, as shown in E, allows imaging of the neuroinflammatory response, as the microglia move from resting to activated statea (OX42 and Iba-1 proteins are expressed at this stage).A, B: Coronal sections of rat brain in an animal model of epilepsy (A, KA) compared to vehicle treated controls (B). This is visualised as OX42 immunoreactivity localised to microglia in response to local excitotoxic brain injury. C. Microglial activation (Iba-1 immunoreactivity) is seen in hippocampus of mice.D. High magnification image of Iba-1 immunoreactive activated microglia.|
|Contact scientist: Paul Callaghan|
Longitudinal studies with PET/CT and SPECT/CT
The imaging research group is also focused on translational applications of molecular imaging techniques in various cancer models. Employing 3 Inveon dual PET/CT scanners and a tri-modality PET/SPECT/CT, there is significant capacity for longitudinal imaging studies. One of the great advantages of longitudinal imaging is the ability to image physiological changes within individual subjects over time. This is a unique experimental paradigm, reducing variability within the population investigated and modelling individual subject’s response to treatment which cannot be done using non-imaging experimental approaches.
Within the cancer research domain, tumoured animals are imaged non-invasively and longitudinally to track the progression of cancer and treatment responses over a period of time. In particular, ANSTO has capacity for functional imaging of biological processes such as metabolism, proliferation, amino acid transport, hypoxia and drug clearance in cancer models.
Within the neuroimaging domain, longitudinal imaging is used for monitoring disease progression and potential pharmacotherapeutic interventions. In particular, we have expertise in imaging glucose metabolism and neuroinflammation in animal models of multiple sclerosis, epilepsy, psychiatric disorders and excitotoxic brain injury.
Both domains use established tracers as well as those developed by ANSTO Life Sciences. The main focus of our pre-clinical program is to establish in vivo imaging biomarkers and ex vivo correlative biomarkers of response which can then be used to design clinical trials.
Radioreceptor binding and autoradiography (receptor density distribution and G protein coupling functionality)
Radioligand binding is an ideal method for determining the specificity and selectivity of uncharacterised compounds for a wide range of receptor targets. Membrane fractions, organelles, whole cells, or histological sections are used in these techniques. Homogenate receptor binding (utilising radiolabelled receptor ligands, cell harvesters and gamma/beta counters) is used for screening novel compound affinity and/or selectivity within receptor preparations.
While immunohistological approaches are useful for detecting the presence of proteins, many complexities within the in vitro methodology itself (such as permeability and access to receptor sites) limit the ability to quantify the expression of a receptor with anatomy preserved. Using radioactive probes attached to either endogenous or synthetic ligands, receptor distribution can be visualised and its density can be quantified. This workflow utilises phosphor plates/film to quantitate the amount of radioactivity in the section, using standards of known activity.
|In vivo: 18F-PBR111 PET/CT; In vitro: Autoradiography (3nM 125 I-CLINDE); In vitro: Ox42 immuno-histochemistry|
|Multimodal imaging showing neuroinflammation in an animal model of epilepsy|
(kainic acid induced status epilepticus, KASE)
[35S]GTPγS Radioreceptor binding and autoradiography.
When an agonist binds to its receptor and the receptor is known to couple with G protein, the process and its extent can be examined using [35S]GTPγS binding (membrane fractions or whole cells) or autoradiography (slides) studies.
This technique is superior to the conventional autoradiography technique because it can investigate the receptor distribution, and quantify the receptor density and functionality. The technique is suitable for longitudinal studies of receptor down-regulation, up-regulation, and desensitisation or tolerance due to chronic or acute administration of agonists. Investigation of cross interactions between several neurotransmission systems is also possible, as is the quantification of signal transductors of second messenger systems through related studies.