WHOLE-BODY, HIGH RESOLUTION, 3D, SMALL ANIMAL PHOTOACOUSTIC AND ULTRASOUND COMPUTED TOMOGRAPHY SYSTEM
Lead Research Organisation:
University College London
Department Name: Medical Physics and Biomedical Eng
Abstract
Before a proposed new medicine reaches the stage of a clinical trial - so before it is allowed to be used on people - it must have already undergone a great deal of testing. The tests will principally examine the safety of the drug and its efficacy. Perhaps the most crucial stage in the drug development pathway prior to human trials involves testing drugs on mice. There are many reasons why mice are used, a key one being that their genetic and biological characteristics are sufficiently similar to humans that many human diseases or medical conditions can be replicated or modelled in mice. In recent years, with the advent of genetically-altered mice and the increase control it offers, mouse models have become even more useful. One of the key questions that is asked during these tests is: where has the drug ended up in the body? If the drug is designed to treat the gut, for example, does it end up in the gut or does it accumulate elsewhere in the body with potentially damaging consequences? Tests of this sort are called biodistribution studies.
The ideal tool for biodistribution studies would be a device that can provide an image of the whole mouse in 3D, in high resolution, and can pick out where in that image the drug is located, eg. by being able to detect its unique spectral signature. It would also be helpful if the imaging technique could be used on a live mouse and, furthermore, did no damage to the mouse. This latter point is especially important as it means that the same animal can be imaged at several points in time to see how the distribution of the drug changes, or see what changes are occurring in the mouse itself over time. There are several small animal imaging modalities that are in routine use today, but none of them come close to meeting this ideal.
There is rapidly growing interest in photoacoustic tomography for small animal imaging because it has the potential to become this ideal tool. Photoacoustic tomography is an emerging technique that uses short pulses of light to generate ultrasound waves within the mouse wherever the light is absorbed. The photoacoustic waves, which carry spatially-resolved information about the structure and even the molecular content of the tissue, propagate out to an array of detectors. A numerical algorithm is then used to reconstruct a 3D volumetric image of the interior of the mouse. The technique is non-invasive, harmless (as it uses non-ionising radiation), and, because it is based on optical absorption, it has the potential to identify components within the tissue based on their optical spectra (which are unique to every type of molecule).
There is one factor holding photoacoustic tomography back from becoming the default approach for small animal imaging the world over: the image quality is not yet as good as it could be. There are three reasons for this. First, the most readily-available sensors cannot detect the full frequency bandwidth of the photoacoustic signals and so fail to capture key information; second, most imaging systems do not detect from all around the animal due to the fabrication complexity and cost of the arrays that would be needed, resulting in image artefacts; third, distortions of the photoacoustic waves due to sound speed variations between and within the different tissue types leads to aberration and blurring in the image, especially at depth.
The scanner proposed here will overcome all three of these limitations of the currently available technologies, through the use of optical detection and generation of ultrasound, and by using ultrasound computed tomography as a adjunct modality to facilitate aberration correction during the image reconstruction.
The ideal tool for biodistribution studies would be a device that can provide an image of the whole mouse in 3D, in high resolution, and can pick out where in that image the drug is located, eg. by being able to detect its unique spectral signature. It would also be helpful if the imaging technique could be used on a live mouse and, furthermore, did no damage to the mouse. This latter point is especially important as it means that the same animal can be imaged at several points in time to see how the distribution of the drug changes, or see what changes are occurring in the mouse itself over time. There are several small animal imaging modalities that are in routine use today, but none of them come close to meeting this ideal.
There is rapidly growing interest in photoacoustic tomography for small animal imaging because it has the potential to become this ideal tool. Photoacoustic tomography is an emerging technique that uses short pulses of light to generate ultrasound waves within the mouse wherever the light is absorbed. The photoacoustic waves, which carry spatially-resolved information about the structure and even the molecular content of the tissue, propagate out to an array of detectors. A numerical algorithm is then used to reconstruct a 3D volumetric image of the interior of the mouse. The technique is non-invasive, harmless (as it uses non-ionising radiation), and, because it is based on optical absorption, it has the potential to identify components within the tissue based on their optical spectra (which are unique to every type of molecule).
There is one factor holding photoacoustic tomography back from becoming the default approach for small animal imaging the world over: the image quality is not yet as good as it could be. There are three reasons for this. First, the most readily-available sensors cannot detect the full frequency bandwidth of the photoacoustic signals and so fail to capture key information; second, most imaging systems do not detect from all around the animal due to the fabrication complexity and cost of the arrays that would be needed, resulting in image artefacts; third, distortions of the photoacoustic waves due to sound speed variations between and within the different tissue types leads to aberration and blurring in the image, especially at depth.
The scanner proposed here will overcome all three of these limitations of the currently available technologies, through the use of optical detection and generation of ultrasound, and by using ultrasound computed tomography as a adjunct modality to facilitate aberration correction during the image reconstruction.
Planned Impact
The goal of this proposal is to develop a new type of small animal scanner which will provide 3D spectroscopic information with higher image quality than current preclinical scanners. One impact of this development, supported by the demonstrations described in the 'Pathways to Impact', is expected to be a marked acceleration in the uptake of photoacoustic tomography for small animal, preclinical, imaging. We anticipate that it will become an indispensable imaging modality in every life sciences department and company worldwide involved in understanding pathological or physiological processes, or developing therapeutic drugs or techniques.
The long term implications therefore include not only an impact on the scientists and companies working to understand the basic processes of diseases, and on those scientists supporting this endeavour, eg. researchers developing new targeted contrast agents, or strains of knockout mice, but, more importantly, an impact on patients who will benefit from the new drugs developed with this imaging technology at the heart of the drug development pathway. There will also be an impact in terms of the significant reduction of the number of animals used for research, (www.nc3rs.org.uk) as photoacoustics can image animals in vivo without causing harm.
By translating the optical technology developed in this proposal into clinical photoacoustic-ultrasound scanners, there is significant potential for impact on clinical practice, eg. on radiologists and other medical professionals who will have a real-time imaging tool as easy to use as ultrasound but also providing functional information such as blood oxygenation, but more importantly on patients who receive earlier and more accurate diagnoses. Furthermore, some of the image reconstruction algorithms developed for this preclinical system will be directly applicable to other imaging scenarios, eg. 3D photoacoustic and ultrasound breast imaging for early tumour detection.
This proposal is highly multi-disciplinary, so as well as these clinical and preclinical impacts, there will be beneficial impacts on other areas of science and engineering. For example, both the broadband plane-wave ultrasound generation and the optical detection technology can be used to significantly improve the accuracy of high frequency ultrasound metrology, which is crucial for the safety of therapeutic ultrasound devices. Also, the algorithms and software developed for acoustic modelling and image reconstruction will have an impact on mathematicians and other researchers working on wave-based inverse problems, such as those in seismology or industrial non-destructive evaluation, by providing them with new and efficient tools as well as insights into a related problem.
The long term implications therefore include not only an impact on the scientists and companies working to understand the basic processes of diseases, and on those scientists supporting this endeavour, eg. researchers developing new targeted contrast agents, or strains of knockout mice, but, more importantly, an impact on patients who will benefit from the new drugs developed with this imaging technology at the heart of the drug development pathway. There will also be an impact in terms of the significant reduction of the number of animals used for research, (www.nc3rs.org.uk) as photoacoustics can image animals in vivo without causing harm.
By translating the optical technology developed in this proposal into clinical photoacoustic-ultrasound scanners, there is significant potential for impact on clinical practice, eg. on radiologists and other medical professionals who will have a real-time imaging tool as easy to use as ultrasound but also providing functional information such as blood oxygenation, but more importantly on patients who receive earlier and more accurate diagnoses. Furthermore, some of the image reconstruction algorithms developed for this preclinical system will be directly applicable to other imaging scenarios, eg. 3D photoacoustic and ultrasound breast imaging for early tumour detection.
This proposal is highly multi-disciplinary, so as well as these clinical and preclinical impacts, there will be beneficial impacts on other areas of science and engineering. For example, both the broadband plane-wave ultrasound generation and the optical detection technology can be used to significantly improve the accuracy of high frequency ultrasound metrology, which is crucial for the safety of therapeutic ultrasound devices. Also, the algorithms and software developed for acoustic modelling and image reconstruction will have an impact on mathematicians and other researchers working on wave-based inverse problems, such as those in seismology or industrial non-destructive evaluation, by providing them with new and efficient tools as well as insights into a related problem.
Publications
Bakaric M
(2021)
Measurement of the ultrasound attenuation and dispersion in 3D-printed photopolymer materials from 1 to 3.5 MHz.
in The Journal of the Acoustical Society of America
Di Sciacca G
(2022)
Evaluation of a pipeline for simulation, reconstruction, and classification in ultrasound-aided diffuse optical tomography of breast tumors.
in Journal of biomedical optics
Javaherian A
(2021)
Ray-based inversion accounting for scattering for biomedical ultrasound tomography
in Inverse Problems
Javaherian, A
(2024)
Hessian-inversion-free ray-born inversion for quantitative ultrasound tomography
in arxiv
Javaherian, A
(2022)
Hessian-inversion-free ray-born inversion for quantitative ultrasound tomography
Pan B
(2023)
On Learning the Invisible in Photoacoustic Tomography with Flat Directionally Sensitive Detector
in SIAM Journal on Imaging Sciences
Poimala J
(2024)
Compensating unknown speed of sound in learned fast 3D limited-view photoacoustic tomography.
in Photoacoustics
Rajagopal S
(2023)
The effect of source backing materials and excitation pulse durations on laser-generated ultrasound waveforms.
in The Journal of the Acoustical Society of America
Description | A 3D, high resolution dual-mode photoacoustic and ultrasound scanner has been demonstrated, with potential applications in pre-clinical imaging in the areas of drug delivery or the study of disease. |
Exploitation Route | There are three categories of ways in which the work undertaken in this study could benefit others. First, the design of the hardware for dual-mode broadband photoacoustic and ultrasound detection; second, the algorithms used to reconstruct the images; third, the instrument itself could be used in applications in biomedical imaging. |
Sectors | Digital/Communication/Information Technologies (including Software) Healthcare Pharmaceuticals and Medical Biotechnology |
Description | DICOM |
Geographic Reach | Multiple continents/international |
Policy Influence Type | Contribution to new or improved professional practice |
URL | https://www.dicomstandard.org/current |
Description | LhARA Ion Therapy Research Facility scoping project ITRF |
Amount | £17,428 (GBP) |
Funding ID | ST/X006115/1 |
Organisation | Science and Technologies Facilities Council (STFC) |
Sector | Public |
Country | United Kingdom |
Start | 09/2022 |
End | 09/2024 |
Description | k-Wave: An open-source toolbox for the time-domain simulation of acoustic wave fields |
Amount | £584,439 (GBP) |
Funding ID | EP/W029324/1 |
Organisation | Engineering and Physical Sciences Research Council (EPSRC) |
Sector | Public |
Country | United Kingdom |
Start | 08/2022 |
End | 08/2025 |
Description | Deep Learning for PAT |
Organisation | University of Arizona |
Country | United States |
Sector | Academic/University |
PI Contribution | Leonid Kunyansky (Arizona) and I have worked together previously on photoacoustic image reconstruction. I have also collaborated with Andreas Hauptmann (Oulu) previously on learned methods for image reconstruction. Most recently, I suggested we all work together to explore the use of Deep Learning for photoacoustic image reconstruction in a cylindrical geometry. |
Collaborator Contribution | Leonid has developed a fast forward model for photoacoustics. Andreas is in the process of using this to develop a partially-learned reconstruction approach for photoacoustic imaging systems with cylindrical geometry. |
Impact | Mathematics, computer science. - Poimala, J., Cox, B.T., Hauptmann, A. "Compensating unknown speed of sound in learned fast 3D limited-view photoacoustic tomography" Photoacoustics, 37, 100597 (2024) https://doi.org/10.1016/j.pacs.2024.100597 - Hauptmann, A., Cox, B.T. "Deep Learning in Photoacoustic Tomography: Current approaches and future directions" J. Biomed. Opt., 25(11), 112903 (2020) doi: 10.1117/1.JBO.25.11.112903 - Bench, C., Hauptmann, A., Cox, B.T. "Towards accurate quantitative photoacoustic imaging: learning vascular blood oxygen saturation in 3D" J. Biomed. Opt., 25(8), 085003 - Hauptmann, A., Lucka, F., Betcke, M., Huynh, N., Adler, J., Cox, B.T., Beard, P., Ourselin, S. and Arridge, S. "Model-based learning for accelerated, limited-view 3D photoacoustic tomography" IEEE Trans. Med. Imag., 37(6), 1382-1393 (2018) 10.1109/TMI.2018.2820382 - Hauptmann, A., Cox, B.T., Lucka, F., Huynh, N., Betcke, M., Beard, P. and Arridge, S. "Approximate k-space models and Deep Learning for fast photoacoustic reconstruction" in F. Knoll et al. (Eds.). Machine Learning in Medical Image Reconstruction 2018, Lecture Notes in Computer Science, 11074, 103-111 (2018) |
Start Year | 2021 |
Description | Deep Learning for PAT |
Organisation | University of Oulu |
Country | Finland |
Sector | Academic/University |
PI Contribution | Leonid Kunyansky (Arizona) and I have worked together previously on photoacoustic image reconstruction. I have also collaborated with Andreas Hauptmann (Oulu) previously on learned methods for image reconstruction. Most recently, I suggested we all work together to explore the use of Deep Learning for photoacoustic image reconstruction in a cylindrical geometry. |
Collaborator Contribution | Leonid has developed a fast forward model for photoacoustics. Andreas is in the process of using this to develop a partially-learned reconstruction approach for photoacoustic imaging systems with cylindrical geometry. |
Impact | Mathematics, computer science. - Poimala, J., Cox, B.T., Hauptmann, A. "Compensating unknown speed of sound in learned fast 3D limited-view photoacoustic tomography" Photoacoustics, 37, 100597 (2024) https://doi.org/10.1016/j.pacs.2024.100597 - Hauptmann, A., Cox, B.T. "Deep Learning in Photoacoustic Tomography: Current approaches and future directions" J. Biomed. Opt., 25(11), 112903 (2020) doi: 10.1117/1.JBO.25.11.112903 - Bench, C., Hauptmann, A., Cox, B.T. "Towards accurate quantitative photoacoustic imaging: learning vascular blood oxygen saturation in 3D" J. Biomed. Opt., 25(8), 085003 - Hauptmann, A., Lucka, F., Betcke, M., Huynh, N., Adler, J., Cox, B.T., Beard, P., Ourselin, S. and Arridge, S. "Model-based learning for accelerated, limited-view 3D photoacoustic tomography" IEEE Trans. Med. Imag., 37(6), 1382-1393 (2018) 10.1109/TMI.2018.2820382 - Hauptmann, A., Cox, B.T., Lucka, F., Huynh, N., Betcke, M., Beard, P. and Arridge, S. "Approximate k-space models and Deep Learning for fast photoacoustic reconstruction" in F. Knoll et al. (Eds.). Machine Learning in Medical Image Reconstruction 2018, Lecture Notes in Computer Science, 11074, 103-111 (2018) |
Start Year | 2021 |
Description | Proton beam monitoring |
Organisation | University of Birmingham |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | Jamie Guggenheim and Tony Price, from the University of Birmingham, and I, are collaborating on a grant proposal (also with a commercial partner) on developing a system for measuring the acoustic emissions from proton therapy beams as a means of monitoring the beams. I will bring numerical acoustic modelling expertise, and the image reconstruction methods are very similar to photoacoustic approaches. |
Collaborator Contribution | Tony brings proton therapy expertise, Jamie brings expertise in high sensitivity acoustic (ultrasonic) detection. |
Impact | None yet, but we hope to submit a grant proposal soon. |
Start Year | 2022 |
Title | Ray based quantitative ultrasound tomography |
Description | A github repository of Matlab functions useful for 3D ray-based ultrasound tomography. In particular, functions that solve the 3D ray-linking problem (I know of no other equivalent functions available open source). |
Type Of Technology | Software |
Year Produced | 2023 |
Impact | It has been used to date for 3D ultrasound tomography of the breast in volunteers and patients in a study conducted at the University of Twente, Netherlands. |
URL | https://github.com/Ash1362/ray-based-quantitative-ultrasound-tomography |