Super Resolution Ultrasound Imaging

Lead Research Organisation: Imperial College London
Department Name: Dept of Physics

Abstract

Context
Many micro-vascular related diseases, such as those associated with diabetes, ischemia and cancer, exhibit changes in the micro-vascular structure and blood-flow. The measurement of micro-vascular morphology and changes in blood flow dynamics is therefore essential for early diagnosis and monitoring. Current clinical imaging modalities cannot adequately resolve the microvasculature or flow dynamics at clinically useful depths. The proposed work would generate three-dimensional (3D) super-resolved ultrasound vascular imaging and velocity mapping at clinical depths in vivo.

We have successfully demonstrated that single microbubble localisation can produce acoustic super-resolution and super-resolved flow velocity images in vivo from standard image data acquired by an unmodified clinical ultrasound system using simple post-processing localisation algorithms. We achieved visualisation and velocity measurements of vessel structures below 20 micro-metres in vivo. Our present work is limited by the underlying two dimensional acquisition strategy; this means that there is no super-resolution information in the third dimension. We propose to overcome this by incorporating recent advances in US imaging technology in order to push beyond the established resolution limits of ultrasound imaging to translate this approach into a clinically useful imaging modality.


Aims and Objectives
Our objective is to develop 3D ultrasound super-resolution imaging of the microvasculature at depths of up to 10 cm with acquisition times that are clinically useful. We aim to be the first group in the world to demonstrate this in humans. To facilitate this transition, we will develop and implement fast 3D super-resolution acquisition strategies and protocols using a combination of compounding strategies with currently available US technology and ultrafast volumetric imaging using dedicated ultrasound matrix array technology. We aim to develop optimised and automated processing algorithms which will enable more precise and efficient image acquisition. Throughout our project, we will demonstrate 3D super resolution imaging and super-resolved velocity mapping using in vitro phantoms and in vivo models, and finally provide 3D super-resolved vasculature images in human studies.

Potential Applications and benefits
The proposed work is essential for clinical translation of super-resolution ultrasound imaging. The new 3D super-resolution imaging method could underlie next generation techniques for ultrasound measurement of the structure and function of the microvasculature. A non-invasive, safe, microscopic assessment of the vasculature could prove crucial to diagnosis, prediction, and intervention in a wide range of diseases.

Publications

10 25 50
publication icon
Brown J (2019) Investigation of Microbubble Detection Methods for Super-Resolution Imaging of Microvasculature. in IEEE transactions on ultrasonics, ferroelectrics, and frequency control

publication icon
Christensen-Jeffries K (2019) Poisson Statistical Model of Ultrasound Super-Resolution Imaging Acquisition Time. in IEEE transactions on ultrasonics, ferroelectrics, and frequency control

publication icon
Christensen-Jeffries K (2017) 3-D In Vitro Acoustic Super-Resolution and Super-Resolved Velocity Mapping Using Microbubbles. in IEEE transactions on ultrasonics, ferroelectrics, and frequency control

publication icon
Christensen-Jeffries K (2020) Super-resolution Ultrasound Imaging. in Ultrasound in medicine & biology

publication icon
Christensen-Jeffries K (2017) Microbubble Axial Localization Errors in Ultrasound Super-Resolution Imaging. in IEEE transactions on ultrasonics, ferroelectrics, and frequency control

 
Description This award enabled us to develop and demonstrate a range of novel ultrasound imaging technologies for super-resolved ultrasound imaging with microbubble contrast agents. This included developments in ultrasound hardware, image processing software and microbubble contrast agents. We also performed studies using commercially-available clinically-approved ultrasound scanners to demonstrate super-resolved imaging in vivo in humans.

We successfully combined two 2D ultrasound transducers operating in parallel for super-resolution imaging for the first time. This approach enabled true 3D super-resolved ultrasound imaging of microbubbles flowing through a phantom consisting of sub-mm cellulose tubes, and we explored the relative orientations of the two transducers in terms of the size of the field of view achieved. We also demonstrated true 3D super-resolved ultrasound imaging in vitro using a clinical 2D matrix array.

We performed in-depth work to understand the trade-offs and limitations of different microbubble localisation algorithms. We explored different algorithms for microbubble localisation and showed that the onset of the signal gave the highest accuracy compared to the other methods tested. We also developed an in-depth simulation method for modelling the whole super-resolution image formation process, including propagation of the ultrasound wave to the microbubble (k-wave software), the microbubble response (Marmottant model), the propagation of the scattered field back to the transducer (k-wave), beamforming, tissue-microbubble separation (pulse inversion, differential imaging and singular value decomposition) and localisation. This resulted in a detailed understanding of the different experimental regimes where different detection methods perform best. We also developed a detailed model based on Poisson statistics that enabled us to study how the image acquisition parameters affect final the super-resolved image.

In terms of ultrasound hardware, we collaborated closely with the MSDLab at the University of Florence to develop a unique pseudo-spiral 2D array transducer with 512 elements. The new transducer design was fabricated by a professional probe manufacturer called Vermon. The 2D array was then integrated into high-frame rate ultrasound scanners (ULA-OP 256) by the Florence group by changing the probe adapters for the new design. The 3D imaging was performed with the 2D array connected to two existing ULA-OP 256 systems, where both systems were synchronized to control transmit and receive from 512 transducer elements in parallel. We then demonstrated the use of this probe to obtain 3D super-resolved ultrasound imaging in vitro. The truly parallel transmit and receive from all 512 elements was chosen to maximise the ability to accurately and precisely localise microbubble signals simultaneously across the whole 3D field of view at the same time. This system was demonstrated on microbubbles flowing through a phantom consisting of sub-mm cellulose tubes.

Using phase-change microbubble contrast agents fabricated in house in Mengxing Tang's lab, we demonstrated the first sub-diffraction limited ultrasound imaging in the absence of flow using ultrasound to activate the microbubbles, which is significant as it opens up the potential to perform super-resolved imaging in regions of the body where there is no (or very slow) inherent flow. We assessed the ultrasound pulse conditions needed to rapidly activate and image these contrast agents and developed a fast acquisition strategy where we demonstrated the first near real-time acquisition of a super-resolved ultrasound image in just 0.2 seconds, which is orders of magnitude faster than flow-based acquisition methods.

Using clinical microbubble contrast ultrasound images acquired using a commercially available and clinically-approved scanner, we demonstrated the application of a two-stage motion correction algorithm that allowed lateral motion of the transducer relative to the patient and also deformations within the tissue to be compensated. This resulted in the joint first (simultaneous publication from another lab) demonstration of 2D super-resolved human in vivo ultrasound imaging.

Working with collaborators at Denmark Hill, we have obtained a substantial body of 2D clinical data from patients with suspected testicular cancer and have processed this to obtain super-resolved ultrasound images. This data has been presented at multiple international conferences and is currently being written up for publication.
Exploitation Route We have published detailed approaches to modelling the ultrasound super-resolution image acquisition and analysis process that can be used and adapted by others to better understand the factors affecting the final image obtained. Microbubble localisation techniques can be used for other ultrasound imaging methods which require the precise knowledge of the location of scatterers in the field of view, e.g. coherent multi-transducer ultrasound imaging, measuring tissue elasticity with single bubbles.

We have published a method for two-stage motion compensation that can be applied to clinical data to enable patient motion and tissue distortions to be overcome.

Working with our collaborators at the University of Florence, we have developed a unique pseudo-spiral 2D matrix array transducer optimised for 3D contrast-enhanced ultrasound imaging in deep tissue. After our successful demonstration, another probe based on the same design was manufactured for a research project between the University of Florence and KU Leuven. This probe was used for high-frame rate echocardiography and their results were published in January 2020. The University of Florence will be able to provide this new high frame rate 3D ultrasound imaging system, consisting of two synchronized ULA-OP 256 systems and the pseudo-spiral 2D matrix array transducer, to other groups interested to pursue this approach.
Sectors Healthcare

 
Description This project provided valuable experience for project research staff who gained interdisciplinary training at the interface between engineering, physics and biomedicine. These staff have now moved on to a permanent academic position at a UK university and a prestigious research fellowship at a UK university. As part of this project, we have developed an ultrasound super-resolution image processing framework that converts raw ultrasound image data into a super-resolved ultrasound image. This software is currently being developed further as part of an Imperial Confidence in Concept award. This new project aims to produce a prototype super-resolution image processing system with greatly improved processing speed and a simple user interface for clinical users, and we will work with local clinical colleagues to demonstrate our prototype's potential. Ultimately, this system should enable much faster SR-US image processing for a wide range of clinical applications. We have an ongoing collaboration with clinicians at Denmark Hill where we are studying the potential of super-resolved ultrasound imaging to detect and/or diagnose testicular cancer and hope to publish the results of this study in the near future. Preliminary findings resulted in being awarded an EPSRC Impact Acceleration Account: Advancing Impact Award, as well as establishing new clinical collaborators who are interested in exploring the potential of ultrasound super-resolution for assessing other conditions including sepsis and kidney function. We have also ongoing collaboration with Maidstone & Tunbridge Well NHS, as well as Royal Marsden Hospital, to perform an initial evaluation the value of super-resolution imaging in managing breast cancer patients.
First Year Of Impact 2020
Sector Healthcare
 
Description Development of 2D and 3D Ultrasound Super-Resolution (US-SR) Imaging for the Clinic
Amount £1,077,817 (GBP)
Funding ID MR/S023542/1 
Organisation Medical Research Council (MRC) 
Sector Public
Country United Kingdom
Start 11/2019 
End 10/2024
 
Description Fast 3D Super-Resolution Ultrasound Imaging Through Acoustic Activation and Deactivation of Nanodroplets
Amount £1,579,302 (GBP)
Funding ID EP/T008970/1 
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Public
Country United Kingdom
Start  
 
Description Testicular Lesion Characterisation using Super-Resolution Ultrasound Imaging
Amount £44,597 (GBP)
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Public
Country United Kingdom
Start 11/2018 
End 10/2019
 
Description Project partners in Fellowship 
Organisation University of North Carolina at Chapel Hill
Country United States 
Sector Academic/University 
PI Contribution Invited talk and lab visit by Kirsten Christensen-Jeffries during the grant period. This established a future collaboration which was formally established for my Fellowship award. The university is a project partner in this fellowship, where I will spend a 6 month placement to work on super-resolution ultrasound imaging.
Collaborator Contribution The university is a project partner in this fellowship, where they have committed to host me for a 6 month placement to work on super-resolution ultrasound imaging.
Impact MRC Career Development Award Fellowship
Start Year 2019
 
Description Institute of Physics: Physical Acoustics Tutorial Day 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Postgraduate students
Results and Impact I gave a seminar on Ultrasound Imaging Super-resolution to a technical audience working in the field of acoustics.
Year(s) Of Engagement Activity 2016
URL http://ecne2016.iopconfs.org/Home
 
Description Invited talk at European Space Agency 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact Invited talk given at the European Space Agency in the Netherlands. This engaged and generated interest in an audience outside healthcare. This sparked questions and discussions. Outcomes included an exchange of knowledge in both directions, were some overlap in methods exists between ultrasound super-resolution and space-related localisation techniques.
Year(s) Of Engagement Activity 2018
 
Description Invited talk at Radiology Meeting (King's College Hospital, Denmark Hill) 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Local
Primary Audience Professional Practitioners
Results and Impact Invited talk was given to clinicians and radiologists at King's College Hospital. This generated intensive discussions, improved understanding about clinical challenges and perception. This also established new clinical collaborators looking to use super-resolution to assess vascular function/flow, as well as new sources of clinical data.
Year(s) Of Engagement Activity 2017
 
Description Invited talk at University (Danish Technical University) 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact Invited talk at Danish Technical University to discuss the current state-of-the-art, challenges, and future for ultrasound super-resolution. This generated intensive discussions, questions and helped establish a collaboration with this research group. This group is now actively working on ultrasound super-resolution following being awarded a grant to work in this area.
Year(s) Of Engagement Activity 2017
 
Description Invited talk at University (Harvard Medical School) 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact Invited talk at Harvard Medical School enabled our research to reach academic researchers who are involved in other clinical conditions not explored in this grant, for example neurological clinical conditions like stroke. This created a large interest in the possible use of ultrasound super-resolution for associated neurological vascular conditions. It also reinforced a working collaboration with an academic at this institution which has been included as a formal collaboration in the presenter's subsequent Fellowship.
Year(s) Of Engagement Activity 2018
 
Description Invited talk at University (University of Florence) 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact Physicists, engineers and electrical engineers listened to an invited talk about our research. This generated extensive discussions, advice and potential explorations for future work. Furthermore this strengthened our existing collaboration with those in the audience we already knew, and established new working relationships.
Year(s) Of Engagement Activity 2017
 
Description Invited talk at University (University of North Carolina) 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact Invited talk and lab visit at the University of North Carolina. The talk stimulated debate, interest and questions about the current state, challenges and future of super-resolution imaging. This also established new research collaborations with the University of North Carolina, which is now a formal collaborator in one of the team members Fellowship. This group are now actively involved in research related to ultrasound super-resolution.
Year(s) Of Engagement Activity 2018
 
Description Presentation at Pint of Science 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Public/other audiences
Results and Impact I was invited to give a presentation for Pint of Science. I used the opportunity to present the science of using bubbles in medical ultrasound imaging.
Year(s) Of Engagement Activity 2017
URL https://pintofscience.co.uk
 
Description STEM for Britain 2019 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Policymakers/politicians
Results and Impact Presenting and discussing my research to Members of both Houses of Parliament at Westminster. This fostered great dialogue and engagement between early-stage researchers and Members both in Westminster and in their Constituencies and encouraged personal interaction between all researchers. This allowed our work to reach a wide and non-specialist audience who have potential to influence policy.
Year(s) Of Engagement Activity 2019