The fickle heart: uncertainty quantification in cardiac and cardiovascular modelling and simulation
2020; Royal Society; Volume: 378; Issue: 2173 Linguagem: Inglês
10.1098/rsta.2020.0119
ISSN1471-2962
AutoresGary R. Mirams, Steven Niederer, Richard H. Clayton,
Tópico(s)Cardiovascular Function and Risk Factors
ResumoYou have accessMoreSectionsView PDF ToolsAdd to favoritesDownload CitationsTrack Citations ShareShare onFacebookTwitterLinked InRedditEmail Cite this article Mirams Gary R., Niederer Steven A. and Clayton Richard H. 2020The fickle heart: uncertainty quantification in cardiac and cardiovascular modelling and simulationPhil. Trans. R. Soc. A.3782020011920200119http://doi.org/10.1098/rsta.2020.0119SectionYou have accessPrefaceThe fickle heart: uncertainty quantification in cardiac and cardiovascular modelling and simulation Gary R. Mirams Gary R. Mirams http://orcid.org/0000-0002-4569-4312 School of Mathematical Sciences, University of Nottingham, Mathematical Sciences Building, University Park, Nottingham, Nottinghamshire NG7 2RD, UK [email protected] Google Scholar Find this author on PubMed Search for more papers by this author , Steven A. Niederer Steven A. Niederer http://orcid.org/0000-0002-4612-6982 Division of Imaging Sciences and Biomedical Engineering, Kings College London, The Rayne Institute, 4th Floor, Lambeth Wing, St Thomas' Hospital, London SE1 7EH, UK Google Scholar Find this author on PubMed Search for more papers by this author and Richard H. Clayton Richard H. Clayton http://orcid.org/0000-0002-8438-7518 Computer Science, University of Sheffield, Regent Court, 211 Portobello, Sheffield S1 4DP, UK Google Scholar Find this author on PubMed Search for more papers by this author Gary R. Mirams Gary R. Mirams http://orcid.org/0000-0002-4569-4312 School of Mathematical Sciences, University of Nottingham, Mathematical Sciences Building, University Park, Nottingham, Nottinghamshire NG7 2RD, UK [email protected] Google Scholar Find this author on PubMed , Steven A. Niederer Steven A. Niederer http://orcid.org/0000-0002-4612-6982 Division of Imaging Sciences and Biomedical Engineering, Kings College London, The Rayne Institute, 4th Floor, Lambeth Wing, St Thomas' Hospital, London SE1 7EH, UK Google Scholar Find this author on PubMed and Richard H. Clayton Richard H. Clayton http://orcid.org/0000-0002-8438-7518 Computer Science, University of Sheffield, Regent Court, 211 Portobello, Sheffield S1 4DP, UK Google Scholar Find this author on PubMed Published:25 May 2020https://doi.org/10.1098/rsta.2020.01191. IntroductionComputational models of the heart are being used to inform drug development [1], make clinical decisions [2] and test basic science hypotheses [3]. Developing mathematical and computational models that represent interacting physical mechanisms and that are informed by experimental and clinical data is a key component of this activity [4]. Cardiac models have increased in sophistication to include subcellular, cellular, tissue and organ-scale effects of electrophysiology, mechanics, haemodynamics, perfusion, fibrosis and the autonomic nervous system [2]. As the accuracy, credibility and use of computational models of the heart have developed, more weight is being placed on their predictions to answer basic science questions, evaluate drug cardiotoxicity or inform clinical treatment decisions. However, as models have become more complicated and sophisticated, the impact of inherent biological variability and measurement error in the data used to constrain the model and its parameters can have a greater impact on predictions. This increased uncertainty needs to be examined and quantified to inform clinicians, pharmacists or physiologists of how much to rely on a prediction, to quantify confidence in model parameters and to assist model developers in targeting refinements to the models.Mathematical and computational models of the heart and cardiovascular system are typically systems of nonlinear ordinary differential equations and nonlinear systems of partial differential equations constrained by conservation laws that are solved using numerical techniques, perhaps on a mesh derived from medical images. Many model parameters are difficult or impossible to obtain directly from routine experimental or clinical measurements. This has led to heart models being developed to simulate representative experimental preparations or patients. Model parameters are then inferred from indirect measurements or are obtained from large amounts of data that have been recorded in different laboratories in highly specialized experimental set-ups. In this process, model parameters are fitted to, or set to, mean measurement values and models are deemed validated, or plausible, if predictions then fall within experimental error bars.Current cardiac models rely on multiple measurements to inform model parameters. These may be recorded from different hearts or come from a single patient. All measurements are subject to variation and uncertainty that is often not reflected in this process, nor are uncertainties in fitted model parameters, assumed model parameters or underlying model assumptions included in the prediction, which tends to give a single answer—providing no context or indication of model confidence to help inform a clinician, pharmacist or basic scientist's decision.Model calibration and forward predictions from noisy, uncertain and incomplete data are performed regularly in the fields of meteorology, hydrology, cosmology and petroleum engineering and are increasingly being applied to interpreting omics data. However, these approaches have yet to be applied routinely to models of cardiac structure and function.Meeting this challenge requires an inter-disciplinary and collaborative approach, bringing together researchers with a range of expertise to create the new community that is needed to have a lasting impact on the cardiac modelling field. This special issue comprises papers that address this challenge, and are based on work done during a month-long programme on uncertainty quantification for cardiac models, held during May–June 2019 at the Isaac Newton Institute in Cambridge and entitled 'The fickle heart' [5].This special issue aims to highlight recent advances and to stimulate future progress in uncertainty quantification for cardiac and cardiovascular models.The first three papers in this special issue are perspective pieces, which aim to capture the state of the art and to highlight future directions for research. The first of these papers conducts an uncertainty audit of cardiac models [6] highlighting the uncertainties and variabilities that arise at different scales. The second focuses on the importance of 'discrepancy'—quantifiable differences between models and reality, and tests methods which attempt to account for this on the parametrization of ion channel models [7]. The third paper describes how natural variability between individuals can be captured in the creation of virtual patient cohorts [8].These perspectives are followed by 12 further research papers, which describe recent developments in the use of statistical methods for uncertainty quantification and sensitivity analysis in cardiac and cardiovascular models. The papers are organized by spatial scale, beginning with those focused on uncertainty in models of subcellular signalling mechanisms in myocytes [9] and fibroblasts [10]. The next paper addresses parametrization of ion channel models while accounting for artefacts in voltage-clamp patch clamp experiments [11]. Two papers perform sensitivity analysis of models for intracellular calcium handling [12] and the atrial action potential [13]. Then a paper looks at variability in electrophysiological properties in tissue and how this relates to arrhythmic risk [14]. The next features the application of 'data assimilation' in cardiac tissue modelling—reconstructing a system's dynamics by combining a biophysical simulation with observed data [15]. Two papers examine uncertainty in the context of left ventricular contraction [16,17]. The final three papers develop methods for patient-specific approaches: for afterload modelling [18]; uncertainty in inferring conduction velocity from clinical measurements [19]; and parameter subset reduction for a simplified model of mechanics and haemodynamics [20].As evidenced by the breadth of applications covered in this special issue, computational cardiac models are being used beyond tools to replicate and rationalize observations and are now being used to formally test hypotheses, identify optimal model structure and inform decisions. These new applications often require not only a prediction but also a realistic measure of confidence in that prediction. To address these new questions will require the adoption of statistical techniques to create and analyse cardiac models to account for and quantify uncertainty as part of best practice in computational cardiac modelling.Data accessibilityThis article has no additional data.Competing interestsWe declare we have no competing interests.FundingThe Isaac Newton Institute is supported by the UK Engineering and Physical Sciences Research Council through grant no. EP/R014604/1.AcknowledgementsThe editors would like to thank the Isaac Newton Institute for Mathematical Sciences in Cambridge, UK, for support and hospitality during the Fickle Heart programme held in May 2019, when work on these papers was initiated.FootnotesOne contribution of 16 to a theme issue 'Uncertainty quantification in cardiac and cardiovascular modelling and simulation'.© 2020 The Author(s)Published by the Royal Society. All rights reserved.References1. Davies MRet al.2016Recent developments in using mechanistic cardiac modelling for drug safety evaluation. Drug Discov. Today 21, 924–938. (doi:10.1016/j.drudis.2016.02.003) Crossref, PubMed, ISI, Google Scholar2. Niederer SA, Lumens J, Trayanova NA. 2019Computational models in cardiology. Nat. Cardiol. Rev. 16, 100–111. (doi:10.1038/s41569-018-0104-y) Crossref, PubMed, ISI, Google Scholar3. Brook BS, Kohl P, King JR. 2011Towards the virtual physiological human: mathematical and computational case studies'. Phil. Trans. R. Soc. A 369, 4145–4148. (doi:10.1098/rsta.2011.0278) Link, ISI, Google Scholar4. Hoekstra AG, van Bavel E, Siebes M, Gijsen F, Geris L. 2018Virtual physiological human 2016: translating the virtual physiological human to the clinic. Interface Focus 8, 20170067. (doi:/10.1098/rsfs.2017.0067) Link, ISI, Google Scholar5. Isaac Newton Institute. 2020The fickle heart. See https://www.newton.ac.uk/event/fht. Google Scholar6. Clayton RHet al.2020An audit of uncertainty in multi-scale cardiac electrophysiology models. Phil. Trans. R. Soc. A 378, 20190335. (doi:10.1098/rsta.2019.0335) Link, ISI, Google Scholar7. Lei Cet al.2020Considering discrepancy when calibrating a mechanistic electrophysiology model. Phil. Trans. R. Soc. A 378, 20190349. (doi:10.1098/rsta.2019.0349) Link, ISI, Google Scholar8. Niederer SAet al.2020Creation and application of virtual patient cohorts of cardiac models. Phil. Trans. R. Soc. A 378, 20190558. (doi:10.1098/rsta.2019.0558) Link, ISI, Google Scholar9. Cao S, Aboelkassem Y, Wang A, Saucerman JJ, Omens JH, McCulloch AD. 2020Quantification of model and data uncertainty in a network model of cardiac myocyte mechanosignaling. Phil. Trans. R. Soc. A 378, 20190336. (doi:10.1098/rsta.2019.0336) Link, ISI, Google Scholar10. Wang A, Cao S, Aboelkassem Y, Valdez-Jasso D. 2020Quantification of uncertainty in a new network model of pulmonary arterial adventitial fibroblast pro-fibrotic signaling. Phil. Trans. R. Soc. A 378, 20190338. (doi:10.1098/rsta.2019.0338) Link, ISI, Google Scholar11. Lei C, Clerx M, Whittaker D, Gavaghan DJ, de Boer T, Mirams G. 2020Accounting for variability in ion current recordings using a mathematical model of artefacts in voltage-clamp experiments. Phil. Trans. R. Soc. A 378, 20190348. (doi:10.1098/rsta.2019.0348) Link, ISI, Google Scholar12. Lo ACY, Bai J, Gladding PA, Fedorov VV, Zhao J. 2020Abnormal calcium handling in atrial myocytes with modulated SERCA uptake: a sensitivity analysis. Phil. Trans. R. Soc. A 378, 20190557. (doi:10.1098/rsta.2019.0557) Link, ISI, Google Scholar13. Houston C, Marchand B, Engelbert L, Cantwell CD. 2020Reducing complexity and unidentifiability when modelling human atrial cells. Phil. Trans. R. Soc. A 378, 20190339. (doi:10.1098/rsta.2019.0339) Link, ISI, Google Scholar14. Lawson BAJ, Oliveira RS, Berg LA, Silva PAA, Burrage K, dos Santos RW. 2020Variability in electrophysiological properties controls arrhythmic risk in a patient-specific model of infarction. Phil. Trans. R. Soc. A 378, 20190341. (doi:10.1098/rsta.2019.0341) Link, ISI, Google Scholar15. Hoffman MJ, Cherry EM. 2020Sensitivity of a data-assimilation system for reconstructing 3D cardiac tissue dynamics. Phil. Trans. R. Soc. A 378, 20190388. (doi:10.1098/rsta.2019.0388) Link, ISI, Google Scholar16. Longobardi S, Lewalle A, Coveney S, Sjaastad I, Espe EKS, Louch WE, Musante CJ, Sher A, Niederer SA. 2020Predicting left ventricular contractile function via Gaussian process emulation in aortic banded rats. Phil. Trans. R. Soc. A 378, 20190334. (doi:10.1098/rsta.2019.0334) Link, ISI, Google Scholar17. Campos JO, Sundnes J, dos Santos RW, Rocha BM. 2020Uncertainty quantification and sensitivity analysis of left ventricular function during the full cardiac cycle. Phil. Trans. R. Soc. A 378, 20190381. (doi:10.1098/rsta.2019.0381) Link, ISI, Google Scholar18. Marx L, Gsell MAF, Rund A, Caforio F, Prassl AJ, Toth-Gayor G, Kuehne T, Augustin CM, Plank G. 2020Personalisation of electro-mechanical models of the pressure-overloaded left ventricle: fitting of Windkessel-type afterload models. Phil. Trans. R. Soc. A 378, 20190342. (doi:10.1098/rsta.2019.0342) Link, ISI, Google Scholar19. Coveney Set al.2020Gaussian process manifold interpolation for probabilistic atrial activation maps and uncertain conduction velocity. Phil. Trans. R. Soc. A 378, 20190345. (doi:10.1098/rsta.2019.0345) Link, ISI, Google Scholar20. van Osta Net al.2020Parameter subset reduction for patient-specific modelling of arrhythmogenic cardiomyopathy related mutation carriers in the CircAdapt model. Phil. Trans. R. Soc. A 378, 20190347. (doi:10.1098/rsta.2019.0347) Link, ISI, Google Scholar Next Article VIEW FULL TEXT DOWNLOAD PDF FiguresRelatedReferencesDetailsCited byWhittaker D, Wang J, Shuttleworth J, Venkateshappa R, Kemp J, Claydon T and Mirams G (2022) Ion channel model reduction using manifold boundaries, Journal of The Royal Society Interface, 19:193, Online publication date: 1-Aug-2022.Whittaker D, Capel R, Hendrix M, Chan X, Herring N, White N, Mirams G and Burton R (2021) Cardiac TdP risk stratification modelling of anti-infective compounds including chloroquine and hydroxychloroquine, Royal Society Open Science, 8:4, Online publication date: 1-Apr-2021. Kügler P (2020) Modelling and Simulation for Preclinical Cardiac Safety Assessment of Drugs with Human iPSC-Derived Cardiomyocytes, Jahresbericht der Deutschen Mathematiker-Vereinigung, 10.1365/s13291-020-00218-w, 122:4, (209-257), Online publication date: 1-Dec-2020. Pathmanathan P, Galappaththige S, Cordeiro J, Kaboudian A, Fenton F and Gray R (2020) Data-Driven Uncertainty Quantification for Cardiac Electrophysiological Models: Impact of Physiological Variability on Action Potential and Spiral Wave Dynamics, Frontiers in Physiology, 10.3389/fphys.2020.585400, 11 This Issue12 June 2020Volume 378Issue 2173Theme issue 'Uncertainty quantification in cardiac and cardiovascular modelling and simulation' compiled and edited by Gary R. Mirams, Steven A. Niederer and Richard H. Clayton Article InformationDOI:https://doi.org/10.1098/rsta.2020.0119PubMed:32448073Published by:Royal SocietyPrint ISSN:1364-503XOnline ISSN:1471-2962History: Manuscript accepted30/04/2020Published online25/05/2020Published in print12/06/2020 License:© 2020 The Author(s)Published by the Royal Society. All rights reserved. Citations and impact Subjectscomputational biology
Referência(s)