A Novel Diagnostic Tool: from Structural Health Monitoring to Tissue Quality Prediction

Lead Research Organisation: University of Liverpool
Department Name: Mathematical Sciences

Abstract

As quality of life constantly improves, the average lifespan will continue to increase. The bad news is that tissue degradation due to wear and tear in an aged body is inevitable and is different from person to person. Fortunately recent advances in sciences and technology have enabled us to work towards personalised medicine this approach calls for a collective effort of researchers from a vast spectrum of specialised subjects. This project, by an interdisciplinary team from four different UK Universities with distinct areas of expertise, aims to predict patient-specific tissue quality which is essential in devising treatments plans. While our primary concern in this study is the bone tissue, the developed framework will apply to other tissues having porous or complex microstructure.

To achieve the aim of the project, we have to overcome a few challenges. Firstly we need better mathematical models to extract the tissue microstructure accurately and automatically and also reliable mathematical methods for comparing two different samples from either a normal image (say with 1024 x 1024 pixels) or a 3D image containing shapes and embedded complex structures. Although one would expect that existing softwares can do these tasks, the reality is that these tasks are harder than we think as the images from a realistic scan contains noise. To counter noise, we use a novel mathematical technique called the variational method that actually constructs an energy quantity involving the whole image and minimises it. In doing so, a noise-free reconstruction is obtained and the precise location of the underlying tissue is identified. This project will develop efficient and robust models to extract local features only. Similarly we can construct different energies to compare two images in the so-called co-registration problem. Our new idea is to use more mathematical approaches and less statistical estimation ideas to achieve more accuracy and robustness. The CMIT research centre at Liverpool specialises in a range of mathematical models for different problem scenarios. Secondly once imaging extracts microstructural geometry, our team from Edinburgh and Heriot-Watt will develop and use a computational mechanics approach to evaluate the tissues' material properties. Since tissue quality depends on what its microstructure looks like, answers to questions such as: is it very porous, are different solid parts poorly connected and are most of the pores aligned in the same direction, should indicate how strong the tissue is. We will examine these microstructures and then conduct range of mechanical tests on the computer so as to obtain properties that tell us when and how the tissue will get damaged or fail. We then study how the microstructural geometry relates to the mechanical behaviour. We will use a process called homogenisation that will enable prediction of properties at macro (tissue) level from knowledge of micro level. Homogenisation to predict tissue failure has not been attempted before and presents a major challenge. Finally, the predicted properties must be validated from a combination of imaging and bespoke in-vitro experimentation. The validation process, to be done in Durham, will evaluate the accuracy of our models and provide model refinements. Once validated, our methodology can be used with confidence as a tool to evaluate tissue properties straight from images.

Thus the proposed project contains both technical and advanced mathematical components and real applications in scenarios where non-invasive imaging is applicable. With imaging technology getting better, cheaper and more wide-spread, the prospects for novel applications of this work are immense.

Planned Impact

The economic and societal impacts of our proposed research are as follows.

Economic Impact. Currently the combined cost of hospital and social care for patients with a hip fractures alone amounts to more than £1.73 billion per year. Tissue degradation due to wear and tear in an ageing population is inevitable. Understanding the progression of the disease process and the effectiveness of curative strategies are important. Providing treatments that are specific to a patient's tissue quality are becoming essential to healthcare. This study will lead to treatments optimised for specific patients. The pharmaceutical industry has huge financial investment in developing drugs that can maintain or modify tissue health. With respect to bone, the drugs work by either inhibiting bone resorption, stimulating bone formation or by a combination of mechanisms which have yet to be fully elucidated. Anti-resorption drugs (bisphosphonates) are being used extensively and have been shown to maintain bone density but result in brittle fractures. Understanding of structural and mechanical determinants of bone quality will help the pharmaceutical industry target relevant biological processes. The orthopaedic implant industries will benefit from this project in terms of extensive experimental data on ex-vivo tissues, computational simulations and scientific insights that focus in particular on how microstructural change influence tissue macroscopic biomechanical behaviour. Periprosthetic bone loss is common with total joint replacements; therefore, we expect this impact to continue to be facilitated with our long-term collaborative partners and new parties to perform proof-of-principle trials.

Societal Impact. There is considerable concern in the society with respect to morbidity and mortality associated with bone diseases. For example, UK based research has shown mortality of 18% at 3 months following a hip fracture (www.nos.org.uk). It has also been shown that patients who survive are extremely limited in their daily living activities. Use of optimum treatments will result in a lower morbidity. Recent years have seen rapid advances in imaging technologies. Image acquisition technologies are likely to further advance forward in a major way and non-invasive and harmless imaging of live tissue structures may well become a reality. There is also considerable ongoing research that aims to derive increasing amount of information using imaging methods currently employed on patients (X-rays, CT-scans). Once it becomes clear which parameters affect bone quality it may even be possible to derive this information from images obtained from currently employed techniques. This will immediately lead to new and further follow-up work along the project approaches and the work will result in a fully interactive diagnostic tool.
This study has the advantage of having several pieces of visually displayable components such as images from scans, segmented features, co-registered features, and distribution of material properties and micro-structures of tissues. We propose to develop enlarged models of bone microstructure using 3D printing systems that are now readily available. We will use these to engage the public at annual Science Festivals. We believe that these will not only raise awareness of tissue quality (and how activities such as exercise can help maintain it), but also generate interest in the community fascinated by bio-mimic systems.

Publications

10 25 50
 
Description The project has achieved all of its planned aims and opened up new research questions being followed. However we have a few findings to report:
Durham University Team
1. Durham team has identified the importance of human bone trabecular strut thickness and connectivity of bone trabecular porous network as the key measures for bone tissue quality especially when they are associated with bone disease such as osteoporosis.
2. Durham team and Liverpool team have jointly developed method to quantitatively compute porous network connectivity accurately.
3. To overcome the natural challenge of biological variations from person to person and sample to sample, reproducible consistent porous structures were generated and manufactured in vitro.
4. Durham team has examined constitutive behaviours of human and bovine bone tissues macroscopically and at nanometer scale. There are significant local variations of mechanical property at nanometer scale. We believe that this is related to tissue composition variation at the nanometer scale.
5. In addition to tissue architecture and function, complex cell behaviour has been modelled.
Edinburgh University Team:
6. . Diagnosis and evaluation of the response of bone and bone implant systems is generally undertaken using time-independent elasticity. The UoE group undertook extensive experimental studies in which trabecular bone samples were examined via micro-CT and then subjected to creep and recovery tests at multiple load levels. Novel constitutive models were developed that have clinical applications. Important clinical implication of the present study relate to the possible role of creep mechanisms and deformations in non-traumatic bone fractures and simulation of implant loosening observed in clinical scenarios.
7. To quantify the relationship between bone microstructure and its mechanical behaviour, the Edinburgh University group undertook studies nonlinear homogenisation. In these the micro-CT images of bone samples were converted to high resolution FE models which were then subjected to numerous mechanical loading scenarios. Early during the project it was realised that commercial FE codes could not be used to carry out these studies. Consequently, parallel nonlinear codes were developed to work with the open source parallel code, ParaFEM developed at Manchester. The nonlinear algorithms included both geometrical and material nonlinearities. A successful proposal was made to the EPSRC for time on the UK supercomputer ARCHER. The developed codes have wide applications in range of large engineering problems that require modelling nonlinearities. The results show that for tension-tension and compression-compression regimes in normal strain space, the yield strains have an isotropic behaviour. However, in the tension-compression quadrants, pure shear and combined normal-shear planes, the macroscopic strain norms at yield have a relatively large variation.
Examination of the effect of the solid phase yield criterion on the macroscopic yield of trabecular bone for samples with different microstructure was considered. The study found that yield criteria cause only small differences in the macroscopic yield strains for most load cases except for those that were compression-dominated and were with high density bone.
The code developed in-house was extended to include damage modelling. Damage evolution was found to be non-isotropic and both damage and hardening were found to depend on the loading mode (tensile, compression or shear).

Heriot-Watt University Team:
8. It has been shown that the fluid present in bone contributes to its mechanical response to load. Indeed, our work showed that bone samples compressed at high loading rate were stiffer that samples compressed at lower loading rate, due to the poroelastic effect.
9. Bone microstructure influences how 'liquid' phase in bone stiffens itself. The effect of liquid appears to be much higher in dense cortical bone, which forms the outer shell of long bones, compared to the more porous trabecular bone found in the head of long bones and in vertebrae.
10. It has also been shown that the type of loading, including the pore pressure and deformation induced by specific loading cases, modify how water affects bone mechanical properties.
11. It has also been shown that fluid flow in the bone matrix correlates with the new bone formation due to mechanical loading.
12. Investigation based on newly-developed collaboration (Dr. U Wolfram) is in applying the developed modelling strategies onto more than 50 human bone samples with known medical conditions and ex vivo measurement data, to further demonstrate the capability of the proposed methodologies. Current results suggests a good correlation between experimental and simulated data at the closest possible loading conditions.
13. Investigation based on newly-developed collaboration (with Sheffield) is in applying the methodologies to animal (mouse) bone tissue and investigate the role of liquid in its mechanical behaviour subject to high strain rate.
14. The proposed methodology has also been extended into the context of clinical diagnosis of soft tissue cancers, using the same principles of multiscale structural homogenisation. Results showed that the mechanical properties of soft tissue (e.g. heterogeneity and anisotropy of prostate tissue) could be an effective marker for primary cancer diagnosis and have a good correlation to the histopathological outcome.
Liverpool team
15. The Liverpool team has worked proactively with all other 3 teams to develop and provide image analysis tools.
Key findings from these studies are:
15a. We have developed 3D co-registration algorithms for computing the deformation fields form trabecular bone micro-CT images at increasing load levels.
15b. We have developed 3D segmentation algorithms for analysing human bone trabecular strut thickness and connectivity.
15c. We have developed adaptive mull-threshing segmentation algorithms to analyse properties and structures of bone tissues.
Exploitation Route Our findings have been taken forward by industry and NHS clinicians. CN Bio, our industrial partner, is keen in commercialising some of our findings. Our finding on the importance of human bone trabecular strut thickness and connectivity of bone trabecular network has been taken forward by our clinical partners at James Cook Hospital and Newcastle.

The constitutive models developed by Edinburgh university team in this project can be used in health care industry to predict the behaviour of the trabecular bone for orthopaedic implant performance and can be used to understand the progression of pathological diseases like osteoporosis.

The findings by Heriot-Watt team can be used to improve prediction of fracture risk due to osteoporosis, by adding correcting function of strain-rate dependency due to the fluid phase in bone into the traditional structure-property relationships used in bone mechanics.
Sectors Digital/Communication/Information Technologies (including Software),Healthcare,Manufacturing, including Industrial Biotechology,Pharmaceuticals and Medical Biotechnology

URL http://tissuemech.hw.ac.uk/research/tissue-diagnostic-modelling.html
 
Description The project EP/K036939/1 has a number of findings that are used in applications. (1) The work on image analysis has direct uses in this project for analysing images of tissues. They are equally applicable to other contexts. Currently we explore the possibility of adapting some of them in medical imaging companies; (2) The work on exploring the strain rate stiffening effect of bone at HWU has led to a computer programme that is compatible to a major mechanical simulation software. The investigator is currently in talk with the company for potential exploitation and integration of the code. If progress is made in IP, it is envisaged that it will be used as a part of their predictive modelling module for bone fracture prediction which they currently have a clinical trial in Belgium. (3) UoE research has developed novel experimental protocols to determine the time dependent mechanical properties of bone and used these to develop constitutive bone models. These new models permit evaluation of implant loosening with time, understanding of why non-traumatic fractures occur and reasons for vertebral deformities. There is ongoing discussion with a multi-national company to use these for the evaluation of their new implants. (4) UoE has also developed novel nonlinear homogenisation approaches which permit realistic inclusion of bone properties in biomechanical models. Software developed has been implemented on the UK supercomputer ARCHER and the intention is to make it open-source for use in other fields. This work has also found an application in the evaluation of snow properties for avalanche prediction.
First Year Of Impact 2015
Sector Digital/Communication/Information Technologies (including Software),Healthcare,Manufacturing, including Industrial Biotechology,Pharmaceuticals and Medical Biotechnology
Impact Types Societal,Economic,Policy & public services

 
Description Biomet Europe
Amount £30,000 (GBP)
Organisation Biomet, Inc 
Sector Private
Country United States
Start 10/2014 
End 09/2016
 
Description Biomet Europe
Amount £30,000 (GBP)
Organisation Biomet, Inc 
Sector Private
Country United States
Start 10/2015 
End 09/2017
 
Description British Council - Newton Fund - UK-China International Research Collaboration Award
Amount £8,500 (GBP)
Organisation Newton Fund 
Sector Public
Country United Kingdom
Start 01/2019 
End 03/2020
 
Description EPSRC Impact Acceleration Account Programme
Amount £65,000 (GBP)
Organisation Economic and Social Research Council 
Sector Public
Country United Kingdom
Start 04/2016 
End 03/2017
 
Description EPSRC Impact Accelerator Award
Amount £125,000 (GBP)
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Department Impact Accelerator Award
Sector Public
Country United Kingdom
Start 11/2018 
End 11/2019
 
Description EPSRC Maths for Healthcare Centre
Amount £2,004,298 (GBP)
Funding ID EP/N014499/1 
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Public
Country United Kingdom
Start 12/2015 
End 12/2019
 
Description Knowledge Transfer Account
Amount £29,700 (GBP)
Organisation University of Sheffield 
Department EPSRC KTA Knowledge Transfer Account
Sector Academic/University
Country United Kingdom
Start 01/2016 
End 04/2017
 
Title Highly scalable nonlinear parallel finite element computer codes 
Description The UoE team has developed highly scalable nonlinear parallel finite element computer codes that are in in the process of being made open source. These are likely to find wider application domains. It has also developed time-dependent constitutive models for bone and implemented them as user defined materials in computer codes, 
Type Of Material Improvements to research infrastructure 
Year Produced 2017 
Provided To Others? Yes  
Impact These are already finding applications in evaluation of implant loosening. 
 
Title Segmentation and monitoring of stents 
Description Our imaging work has led to new collaborations with the Royal Liverpool Hospital on segmentation and monitoring of stents 
Type Of Material Improvements to research infrastructure 
Year Produced 2014 
Provided To Others? Yes  
Impact The results are published in open literature. 
 
Title Characterisation of mechanical behaviour of demineralised and deproteinised trabecular bone samples 
Description Nineteen (n=19) bovine trabecular bone samples were scanned before chemical treatment by using Skyscan at high resolution (17.22 µm). The chemical treatment was to either demineralise or deproteinise the samples. Thirteen (n=13) trabecular bone cylindrical samples were successfully demineralised. The mechanical tests included fully reversed cyclic loading experiment (n=5) to investigate bone's asymmetric mechanical behaviour in tension and compression. Tests were also conducted to evaluate the timedependent behaviour of demineralised trabecular bone through tensile multiple-load-creep-unload-recovery (MLCUR) experiments (n=8). The original uCT, undertaken before chemical treatment, bone volume ratio (BV/TV) of samples for fully reversed cyclic experiment was found to be in the range of 21% - 32%; and BV/TV for MLCUR experiment samples were found in the range of 16% - 38%. Other six (n=6) trabecular bone cylindrical samples were deproteinised and subjected to monotonic loading. The bone volume ratio range from 22% to 32%. The detailed experiment methodologies can be found in related thesis chapter and/or cited journal papers. 
Type Of Material Database/Collection of data 
Year Produced 2018 
Provided To Others? Yes  
Impact We expect the dataset will permit evaluation of how bone's composite behaviour changes with change in constituents. 
URL http://dx.doi.org/10.7488/ds/2339
 
Title Characterisation of time-dependent mechanical behaviour of trabecular bone samples 
Description Twenty (n=20) untreated bovine trabecular bone cylindrical samples (without any chemical treatment) were scanned by suing uCT at high resolution (17.22 µm), before mechanical tests. The bone volume ratio (BV/TV) ranged from 15% to 54%. The provided uCT images can be used for the development of finite element models of the bone microstructure and for evaluating indices of bone microstructure (e.g. bone volume ratio, fabric tensors). Samples were subjected to compressive multiple-load-creep-unload-recovery experiments, the raw experimental data is provided for any possible further analysis. This innovative creep-recovery experiments at multiple load levels, which considered compression force equivalent to different strain levels (from 2000 µe to 25000 µe), investigated the time-dependent behaviour of trabecular bone in pre- and post-yield regimes. The detailed experiment methodology can be found from related thesis chapter and the cited journal papers. 
Type Of Material Database/Collection of data 
Year Produced 2018 
Provided To Others? Yes  
Impact This data has permitted us to evaluate implant loosening in joint replacement and fracture fixation surgeries and led to an industry supported grant. 
URL http://dx.doi.org/10.7488/ds/2340
 
Title HWU model on bone poroelastic response at high strain rate 
Description The model, which has been uploaded onto HWU PURE system therefore open-source, is capable of modelling the hardening effect of bone at high strain rate, taking into account multiscale microstructures in bone. 
Type Of Material Computer model/algorithm 
Year Produced 2015 
Provided To Others? Yes  
Impact The model has been used by 4 different research groups (known to the investigator) and the team is currently in talk with a software company for potential commercialisation opportunity based on its extension and further development. 
 
Title Potential for Deep Learning applications 
Description The selective segmentation techniques developed have the potential to help preparation of large training data for Deep Learning applications 
Type Of Material Computer model/algorithm 
Year Produced 2016 
Provided To Others? Yes  
Impact It can reduce the preparation time for training data. 
 
Description Edinburgh Parallel Computing Centre 
Organisation University of Edinburgh
Department Edinburgh Parallel Computing Centre (EPCC)
Country United Kingdom 
Sector Academic/University 
PI Contribution The Edinburgh University team provided nonlinear parallel finite element codes
Collaborator Contribution EPCC has added new solvers to the code, which will permit the code to execute a wider range of problems. This is part of a supported eCSE proposal.
Impact A nonlinear parallel code with increased solving capabilities has been developed. The work includes multiple disciplines: computational mechanics; bioengineering; computer science.
Start Year 2015
 
Description HWU new collaboration with Dr. U Wolfram 
Organisation Heriot-Watt University
Department Life Sciences
Country United Kingdom 
Sector Academic/University 
PI Contribution Investigation based on newly-developed collaboration is in applying the developed modelling strategies onto more than 50 human bone samples with known medical conditions and ex vivo measurement data, to further demonstrate the capability of the proposed methodologies. Current results suggests a good correlation between experimental and simulated data at the closest possible loading conditions.
Collaborator Contribution Dr. U Wolfram provides in situ measurement data of compression test (in various conditions) of human bone samples, and gives insights to, in particular, in bone tissue behaviour in high strain rates.
Impact A joint publication is being worked on at the moment. We will expect to submit the paper in mid-late 2017.
Start Year 2016
 
Description New collaboration with Dr Rui Zhu - Tongji University 
Organisation Tongji University
Department Medical School
Country China 
Sector Hospitals 
PI Contribution HWU group initiated new collaboration with Dr Rui Zhu, Associate Professor, and his group from the Medical School at Tongji University, China in 2018. The structure-property relationship model developed from this EPSRC project was then used to characterise a number of different soft tissues in human spinal structure, particularly on the annulus fibrosus tissue. We applied our microstructural model and successfully derived the tissue (both elastic and viscoelastic) properties, which are currently used to be correlated to clinical indices for degenerative spinal cord disease. We expect some first-hand results for this by June 2019.
Collaborator Contribution Dr Rui Zhu and his group at Tongji University provided clinical inputs, including sample histology and medical images, following approved ethics application, into our developed model. Moreover, the model parameters from the biomechanical model of spinal cord involving both spinal tissue and muscular network were introduced, in the format of biomechanical loading, into our tissue models.
Impact This collaboration is multi-disciplinary, between the groups at HWU (Biomechanics) and Tongji University (Clinical spine disease research). The immediate outcomes from this collaboration would be a predictive model that could estimate the mechanical properties of annulus fibrosus tissue from their tissue microstructure. We plan to correlate the derived mechanical properties to the clinical indices of the patient from whom the sample was retrieved, leading to a quantitative relationship between mechanics and pre-surgical conditions.
Start Year 2018
 
Description Pontificia Universidade Catolica do Rio de Janeiro, Brazil 
Organisation Pontifical Catholic University of Rio de Janeiro
Country Brazil 
Sector Academic/University 
PI Contribution Edinburgh team provided the expertise in computational modelling required to understand the mechanical behaviour of bone
Collaborator Contribution Dr E Sales, from Brazil, spent two years with the Edinburgh team in which she supported the project through her expertise of Micro_CT image segmentation and analysis
Impact Dr Sales is a named author on several publications that emerged from this collaboration (please see publications). This collaboration was multidisciplinary and involved imaging and mechanics.
Start Year 2013
 
Description SheffieldGroup 
Organisation University of Sheffield
Country United Kingdom 
Sector Academic/University 
PI Contribution Micro CT scans were taken at many compressive strain levels for cylindrical trabecular bone samples at university of Edinburgh. Images were reconstructed and passed on to Dr. Enrico Dall'Ara and Marco Viceconti group, Sheffield.
Collaborator Contribution At university of Sheffield, Enrico's group has performed image registration analysis on the images we sent, and a paper has been submitted together (currently revised version is under review).
Impact One journal article has been submitted, and is under review.
Start Year 2014
 
Description University of Manchester 
Organisation University of Manchester
Department Centre for Biostatistics
Country United Kingdom 
Sector Academic/University 
PI Contribution Edinburgh team developed new nonlinear codes that can be used the linear finite element code ParaFEM available at Manchester. This new code will be made open source.
Collaborator Contribution Dr Lee Margetts from the University of Manchester provided and supported the the linear finite element code ParaFEM for execution on HPCs.
Impact New open source code developed for implementation on HPCs. A number of joint publications have emerged that are reported in the publications section.
Start Year 2013
 
Description University of Sheffield 
Organisation University of Sheffield
Country United Kingdom 
Sector Academic/University 
PI Contribution Edinburgh analysed trabecular bone under micro-CT and how it changed with loading. We contributed this data to the Sheffield group.
Collaborator Contribution The Sheffield group used image registration algorithms to evaluate local responses in bone.
Impact Edinburgh-Sheffield joint publications have emerged which have been reported in the publications section.
Start Year 2014
 
Title Tissue Structure Scaffolds 
Description A method of forming highly interconnected collagen scaffolds mimicking natural tissue ECM structure. 
IP Reference WO2017182676 
Protection Patent application published
Year Protection Granted 2017
Licensed No
Impact CN Bio has used this tissue scaffolds and have seen notable effects.
 
Title Nonlinear driver programs for HPCs 
Description Driver programs that incorporate plasticity, damage and nonlinear geometry were developed to work with the open source parallel code ParaFEM. These have been shown to be highly scalable with the number of core used in computation. They will be made open source in due course. 
Type Of Technology Software 
Year Produced 2016 
Impact We expect these codes will find a wide range of applications in different engineering fields that require solution of large nonlinear problems. 
URL http://parafem.org.uk/news/general/179-geometric-and-material-nonlinearity
 
Title Nonlinear visco-elastic and visco-plastic software 
Description User defined material routines have been developed to simulate the nonlinear viscoelastic and viscoplastic behaviour of bone. 
Type Of Technology Software 
Year Produced 2017 
Impact We expect these to be able to simulate orthopaedic implant loosening under cyclic loads and help understand why non-traumatic fractures occur. 
 
Title Poroelasticity in bone microstructure 
Description The software is based on python and MATLAB, working compatibly with ABAQUS, and is capable of modelling poroelastic effects due to fluid movement in bone microstructure subject to a vast range of strain rate, using homogenisation method with a number of different boundary conditions e.g. PBC and KUBC. 
Type Of Technology Software 
Year Produced 2017 
Impact Our modelling work has received attention from research groups, and initiated a number of new collaborations. It is now being tested on various samples (human and small animals) and under many loading conditions (for different loading scenarios, physiological and lab-based). 
 
Description "Imaging to modelling: Workflows for clinical practice" 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Professional Practitioners
Results and Impact Prof Pankaj organized a national meeting in Edinburgh entitled "Imaging to modelling: Workflows for clinical practice" which had participation from physical scientists, engineers and clinicians.
Year(s) Of Engagement Activity 2016