Top

Current Cardiology Reports

Published in:

01-08-2020 | Magnetic Resonance Imaging | Cardiac PET, CT, and MRI (P Schoenhagen and P-H Chen, Section Editors)

Machine Learning-Based Segmentation of Left Ventricular Myocardial Fibrosis from Magnetic Resonance Imaging

Authors: Fatemeh Zabihollahy, S. Rajan, E. Ukwatta

Published in: Current Cardiology Reports | Issue 8/2020

Abstract

Purpose of Review

Myocardial fibrosis (MF) arises due to myocardial infarction and numerous cardiac diseases. MF may lead to several heart disorders, such as heart failure, arrhythmias, and ischemia. Cardiac magnetic resonance (CMR) imaging techniques, such as late gadolinium enhancement (LGE) CMR, enable non-invasive assessment of MF in the left ventricle (LV). Manual assessment of MF on CMR is a tedious and time-consuming task that is subject to high observer variability. Automated segmentation and quantification of MF is important for risk stratification and treatment planning in patients with heart disorders. This article aims to review the machine learning (ML)-based methodologies developed for MF quantification in the LV using CMR images.

Recent Findings

With the availability of relatively large labeled datasets supervised learning methods based on both conventional ML and state-of-the-art deep learning (DL) methods have been successfully applied for automated segmentation of MF. The incorporation of ML algorithms into imaging techniques such as 3D LGE CMR permits fast characterization of MF on CMR imaging and may enhance the diagnosis and prognosis of patients with heart disorders. Concurrently, the studies using cine CMR images have revealed that accurate segmentation of MF on non-contrast CMR imaging might be possible.

Summary

The application of ML/DL tools in CMR image interpretation is likely to result in accurate and efficient quantification of MF.

Murtha LA, Schuliga MJ, Mabotuwana NS, Hardy SA, Waters DW, Burgess JK, et al. The processes and mechanisms of cardiac and pulmonary fibrosis. Front Physiol. 2017;8:1664–042X. https://doi.org/10.3389/fphys.2017.00777.CrossRef

Weber KT, Sun Y, Bhattacharya SK, Ahokas RA, Gerling IC. Myofibroblast-mediated mechanisms of pathological remodelling of the heart. Nat Rev Cardiol. 2013;10:15–26. https://doi.org/10.1038/nrcardio.2012.158.CrossRefPubMed

Zeisberg M, Kalluri R. Cellular mechanisms of tissue fibrosis. Common and organ-specific mechanisms associated with tissue fibrosis. Am J Physiol Cell Physiol. 2013;304:C216–25. https://doi.org/10.1152/ajpcell.00328.2012.CrossRefPubMed

Jellis C, Martin J, Narula J, Marwick TH. Assessment of nonischemic myocardial fibrosis. J Am Coll Cardiol. 2010;56:89–97. https://doi.org/10.1016/j.jacc.2010.02.047.CrossRefPubMed

Disertori M, Masè M, Ravelli F. Myocardial fibrosis predicts ventricular tachyarrhythmias. Trends Cardiovasc Med. 2017;27:363–72.CrossRefPubMed

Espeland T, Lunde IG, Amundsen BH, Gullestad L, Aakhus S. Myocardial fibrosis. Tidsskr Nor Laegeforen. 2018;138. https://doi.org/10.4045/tidsskr.17.1027.

Piek A, de Boer RA, Silljé HH. The fibrosis-cell death axis in heart failure. Heart Fail Rev. 2016;21:199–211. https://doi.org/10.1007/s10741-016-9536-9.CrossRefPubMedPubMedCentral

Barison A, Grigoratos C, Todiere G, Aquaro GD. Myocardial interstitial remodelling in non-ischaemic dilated cardiomyopathy: insights from cardiovascular magnetic resonance. Heart Fail Rev. 2015;20:731–49. https://doi.org/10.1007/s10741-015-9509-4.CrossRefPubMed

Herum KM, Lunde IG, McCulloch AD, Christensen G. The soft- and hard-heartedness of cardiac fibroblasts: mechanotransduction signaling pathways in fibrosis of the heart. J Clin Med. 2017;6:E53. https://doi.org/10.3390/jcm6050053.CrossRefPubMed

10.

Travers JG, Kamal FA, Robbins J, Yutzey KE, Blaxall BC. Cardiac fibrosis: the fibroblast awakens. Circ Res. 2016;118:1021–40. https://doi.org/10.1161/CIRCRESAHA.115.306565.CrossRefPubMedPubMedCentral

11.

Kong P, Christia P, Frangogiannis NG. The pathogenesis of cardiac fibrosis. Cell Mol Life Sci. 2014;71:549–74. https://doi.org/10.1007/s00018-013-1349-6.CrossRefPubMed

12•.

Tao Q, Sebastiaan RP, Hildo JL, Katja Z, Rob JG. Myocardial scar surface area identified by LGE MRI is an independent predictor of mortality in post-infarction patients. J Cardiovasc Magn Reson. 2015;17(Suppl 1):P46. https://doi.org/10.1186/1532-429X-17-S1-P46The findings of this study demonstrate that MF surface area identified from LGE CMRI is a predictor of all-course mortality in post-infarction patients, independent of left ventricle ejection fraction. CrossRefPubMedCentral

13.

Lin A, Kwan C, Guppy-Coles K, Sippel J, Atherton J, Prasad S. Impact of infarct size on left ventricular diastolic function following acute myocardial infarction. Heart Lung Circ. 2015;132:A18996. https://doi.org/10.1016/j.hlc.2015.06.089.CrossRef

14.

Nguyen T, Phan J, Richards D, Moses D, Serratore M, Hogan J, et al. The impact of myocardial scar size and systolic function on diastolic dysfunction following acute st elevated myocardial infarction (STEMI) revascularisation. Heart Lung Circ. 2013;22:S200. https://doi.org/10.1016/j.hlc.2013.05.476.CrossRef

15.

Trayanova NT. Computational cardiology: the heart of the matter. ISRN Cardiol. 2012;2012:269680–15. https://doi.org/10.5402/2012/269680.CrossRefPubMedPubMedCentral

16.

Ashikaga H, Arevalo H, Vadakkumpadan F, Blake RC, Bayer JD, Nazarian S, et al. Feasibility of image-based simulation to estimate ablation target in human ventricular arrhythmia. Heart Rhythm. 2013;10:1109–16. https://doi.org/10.1016/j.hrthm.2013.04.015.CrossRefPubMedPubMedCentral

17.

Winslow RL, Trayanova N, Geman D, Miller MI. Computational medicine: Translating models to clinical care. Sci Transl Med. 2012;4:158rv11. https://doi.org/10.1126/scitranslmed.3003528.CrossRefPubMedPubMedCentral

18.

Vigmond E, Vadakkumpadan F, Gurev V, Arevalo H, Deo M, Plank G, et al. Towards predictive modelling of the electrophysiology of the heart. Exp Physiol. 2009;94:563–77. https://doi.org/10.1113/expphysiol.2008.044073.CrossRefPubMedPubMedCentral

19.

Vadakkumpadan F, Gurev V, Constantino J, Arevalo H, Trayanova N. Modeling of whole-heart electrophysiology and mechanics: toward patient-specific simulations. Modeling of whole-heart electrophysiology and mechanics: toward patient-specific simulations. In: Kerckhoffs R, editor. Patient-specific modeling of the cardiovascular system. New York: Springer; 2010.

20.

Vadakkumpadan F, Arevalo H, Jebb A, Wu KC, Trayanova N. Image-based patient-specific simulations of ventricular electrophysiology for sudden arrhythmic death risk stratification. Circulation. 2013;128:A18014. https://doi.org/10.1161/CIRCEP.117.004743.CrossRef

21.

Ukwatta E, Nikolov P, Zabihollahy F, Trayanova NA, Wright GA. Virtual electrophysiological study as a tool for evaluating efficacy of MRI techniques in predicting adverse arrhythmic events in ischemic patients. Phys Med Biol. 2018;63:225008. https://doi.org/10.1088/1361-6560/aae8b2.CrossRefPubMedPubMedCentral

22.

Trayanova N, Boyle PM, Arevalo HJ, Zahid S. Exploring susceptibility to atrial and ventricular arrhythmias resulting from remodeling of the passive electrical properties in the heart: a simulation approach. Front Physiol. 2014;5:435. https://doi.org/10.3389/fphys.2014.00435.CrossRefPubMedPubMedCentral

23.

Ukwatta E, Arevalo H, Rajchl M, White JA, Pashakhanloo F, Prakosa A, et al. Image-based reconstruction of three-dimensional myocardial infarct geometry for patient-specific modeling of cardiac electrophysiology. Med Phys. 2015;42:4579–90. https://doi.org/10.1118/1.4926428.CrossRefPubMedPubMedCentral

24.

Schmidt A, Azevedo CF, Cheng A, Gupta SN, Bluemke DA, Foo TK, et al. Infarct tissue heterogeneity by magnetic resonance imaging identifies enhanced cardiac arrhythmia susceptibility in patients with left ventricular dysfunction. Circulation. 2007;115:2006–14. https://doi.org/10.1161/CIRCULATIONAHA.106.653568.CrossRefPubMedPubMedCentral

25.

Bello D, Fieno DS, Kim RJ, Pereles FS, Passman R, Song G, et al. Infarct morphology identifies patients with substrate for sustained ventricular tachycardia. J Am Coll Cardiol. 2005;45:1104–8. https://doi.org/10.1016/j.jacc.2004.12.057.CrossRefPubMed

26.

Kubicek J, Bryjova I, Penhaker M, Augustynek M. Extraction of myocardial fibrosis from MR using fuzzy soft thresholding algorithm. In: Proceedings - 2016 5th IIAI International Congress on Advanced Applied Informatics, IIAI-AAI 2016. 2016. https://doi.org/10.1109/IIAI-AAI.2016.143.

27.

Rajiah P, Desai MY, Kwon D, Flamm SD. MR imaging of myocardial infarction. Radiographics. 2013;33:1383–413. https://doi.org/10.1148/rg.335125722.CrossRefPubMed

28.

Mewton N, Chia Ying L, Croisille P, Bluemke D, João ACL. Assessment of myocardial fibrosis with cardiovascular magnetic resonance. J Am Coll Cardiol. 2011;57(8):891–903. https://doi.org/10.1016/j.jacc.2010.11.013.CrossRefPubMed

29.

Flett AS, Hayward MP, Ashworth MT, Hansen MS, Taylor AM, Elliott PM, et al. Equilibrium contrast cardiovascular magnetic resonance for the measurement of diffuse myocardial fibrosis: preliminary validation in humans. Circulation. 2010;122(2):138–44. https://doi.org/10.1161/CIRCULATIONAHA.109.930636 ISSN 0009-7322.CrossRefPubMed

30.

Pfeiffer MP, Biederman RW. Cardiac MRI: a general overview with emphasis on current use and indication. Med Clin North Am. 2015;99:849–61. https://doi.org/10.1016/j.mcna.2015.02.011.CrossRefPubMed

31.

Kim RJ, Fieno DS, Parrish TB, Harris K, Chen EL, Simonetti O, et al. Relationship of MRI delayed contrast enhancement to irreversible injury, infarct age, and contractile function. Circulation. 1999;100(19):1992–2002. https://doi.org/10.1161/01.CIR.100.19.1992Circulation.CrossRefPubMed

32.

Ibrahim T, Makowski MR, Jankauskas A, Maintz D, Karch M, Schachoff S, et al. Serial contrast-enhanced cardiac magnetic resonance imaging demonstrates regression of hyperenhancement within the coronary artery wall in patients after acute myocardial infarction. JACC Cardiovasc Imaging. 2009;2:580–8. https://doi.org/10.1016/j.jcmg.2008.12.029.CrossRefPubMed

33.

Farrag NA, Ramanan V, Wright GA, Ukwatta E. Effect of T1-mapping technique and diminished image resolution on quantification of infarct mass and its ability in predicting appropriate ICD therapy. Med Phys. 2018;45:1577–85. https://doi.org/10.1002/mp.12840.CrossRefPubMed

34.

Amano Y, Yanagisawa F, Tachi M, Asai K, Suzuki Y, Hashimoto H, et al. Three-dimensional cardiac MR imaging: related techniques and clinical applications. Magn Reson Med Sci. 2017;16:183–9. https://doi.org/10.2463/mrms.rev.2016-0116.CrossRefPubMedPubMedCentral

35.

Davis CP, Hany TF, Wildermuth S, Schmidt M, Debatin JF. Postprocessing techniques for gadolinium-enhanced three-dimensional MR angiography. Radiographics. 1997;17:1061–77. https://doi.org/10.1148/radiographics.17.5.9308101.CrossRefPubMed

36.

Heiss SG, Shifrin RY, Sommer FG. Contras-enhanced three-dimensional fast spoiled gradient-echo renal MR imaging: evaluation of vascular and nonvascular disease. Radiographics. 2000;20:1341–52. https://doi.org/10.1148/radiographics.20.5.g00se251341.CrossRefPubMed

37.

Bizino M, Tao Q, Amersfoort J, Van Der Geest RJ, Lamb HJ. 3D free-breathing late gadolinium enhancement 3T MRI: comparison with standard 2D Imaging. J Cardiovasc Magn Reson. 2014;16(Suppl 1):P198. https://doi.org/10.1186/1532-429X-16-S1-P198.CrossRefPubMedCentral

38.

Viallon M, Jacquier A, Rotaru C, Delattre BMA, Mewton N, Vincent F, et al. Head-to-head comparison of eight late gadolinium-enhanced cardiac MR (LGE CMR) sequences at 1.5 tesla: from bench to bedside. J Magn Reson Imaging. 2011;34(6):1374–87. https://doi.org/10.1002/jmri.22783.CrossRefPubMed

39.

Morita K, Utsunomiya D, Oda S, Komi M, Namimoto T, Hirai T, et al. Comparison of 3D phase-sensitive inversion-recovery and 2D inversion-recovery MRI at 3.0 T for the assessment of late gadolinium enhancement in patients with hypertrophic cardiomyopathy. Acad Radiol. 2013;20(6):752–7. https://doi.org/10.1016/j.acra.2013.01.014.CrossRefPubMed

40.

Goetti R, Kozerke S, Donati OF, Sürder D, Stolzmann P, Kaufmann PA, et al. Acute, subacute, and chronic myocardial infarction: quantitative comparison of 2D and 3D late gadolinium enhancement MR imaging. Radiology. 2011;259(3):704–11. https://doi.org/10.1148/radiol.11102216.CrossRefPubMed

41.

Rajchl M, Yuan J, White JA, Ukwatta E, Stirrat J, Nambakhsh CMS, et al. Interactive hierarchical-flow segmentation of scar tissue from late-enhancement cardiac MR images. IEEE Trans Med Imaging. 2014;33(1):159–72. https://doi.org/10.1109/TMI.2013.2282932.CrossRefPubMed

42.

Xu C, Xu L, Gao Z, Zhao S, Zhang H, Zhang Y, et al. Direct delineation of myocardial infarction without contrast agents using a joint motion feature learning architecture. Med Image Anal. 2018;50:82–94. https://doi.org/10.1016/j.media.2018.09.001.CrossRefPubMed

43.

•• Xu C, Howey J, Ohorodnyk P, Roth M, Zhang H, Li S. Segmentation and quantification of infarction without contrast agents via spatiotemporal generative adversarial learning. Med Image Anal. 2020;59:101568. https://doi.org/10.1016/j.media.2019.101568The findings of this research show that LV MF characterization from non-contrast CMRI with high accuracy is possible. CrossRefPubMed

44.

Fox CS, Muntner P, Chen AY, Alexander KP, Roe MT, Cannon CP, et al. Use of evidence-based therapies in short-term outcomes of ST-segment elevation myocardial infarction and non–ST-segment elevation myocardial infarction in patients with chronic kidney disease: a report from the National Cardiovascular Data Acute Coronary Treatment and Intervention Outcomes Network Registry. Circulation. 2010;121(3):357–65. https://doi.org/10.1161/CIRCULATIONAHA.109.865352.CrossRefPubMedPubMedCentral

45.

Rosendahl L, Blomstrand P, Heiberg E, Ohlsson J, Björklund PG, Ahlander BM, et al. Computer-assisted calculation of myocardial infarct size shortens the evaluation time of contrast-enhanced cardiac MRI. Clin Physiol Funct Imaging. 2008;28(1):1–7. https://doi.org/10.1111/j.1475-097X.2007.00765.x.CrossRefPubMed

46.

Neizel M, Katoh M, Schade E, Rassaf T, Krombach GA, Kelm M, et al. Rapid and accurate determination of relative infarct size in humans using contrast-enhanced magnetic resonance imaging. Clin Res Cardiol. 2009;98:319–24. https://doi.org/10.1007/s00392-009-0007-3.CrossRefPubMed

47.

Kolipaka A, Chatzimavroudis GP, White RD, O’Donnell TP, Setser RM. Segmentation of non-viable myocardium in delayed enhancement magnetic resonance images. Int J Cardiovasc Imaging. 2005;21:303–11. https://doi.org/10.1007/s10554-004-5806-z.CrossRefPubMed

48.

Ukwatta E, Arevalo H, Li K, Yuan J, Qiu W, Malamas P, et al. Myocardial infarct segmentation from magnetic resonance images for personalized modeling of cardiac electrophysiology. IEEE Trans Med Imaging. 2016;35:1408–19. https://doi.org/10.1109/TMI.2015.2512711.CrossRefPubMed

49.

Ukwatta E, Yuan J, Qiu W, Wu KC, Trayanova N, Vadakkumpadan F. Myocardial infarct segmentation and reconstruction from 2D late-gadolinium enhanced magnetic resonance image. Med Image Comput Comput Assist Interv. 2014;35:1408–19. https://doi.org/10.1109/TMI.2015.2512711.CrossRef

50.

Usta F, Gueaieb W, White JA, Ukwatta E. 3D scar segmentation from LGE-MRI using a continuous max-flow method. Houston: SPIE Medical Imaging; 2018. https://doi.org/10.1117/12.2294406.CrossRef

51.

Lu Y, Connelly KA, Yang Y, Joshi SB, Wright G, Radau PE. Semi-automated analysis of infarct heterogeneity on DE-MRI using graph cuts. J Cardiovasc Magn Reson. 2012;14:T6. https://doi.org/10.1186/1532-429X-14-S1-T6.CrossRefPubMedCentral

52.

Zhu X, Vondrick C, Fowlkes CC, Ramanan D. Do we need more training data? Int J Comput Vis. 2016;119(17):76–92. https://doi.org/10.1007/s11263-015-0812-2.CrossRef

53.

LeCun Y, Bengio Y, Hinton G. Deep learning. Nature. 2015;521(7553):436–44.CrossRefPubMed

54.

Karim R, Bhagirath P, Claus P, James Housden R, Chen Z, Karimaghaloo Z, et al. Evaluation of state-of-the-art segmentation algorithms for left ventricle infarct from late Gadolinium enhancement MR images. Med Image Anal. 2016;30:95–107. https://doi.org/10.1016/j.media.2016.01.004.CrossRefPubMed

55.

Kurzendorfer T, Breininger K, Steidl S, Brost A, Forman C, Maier A. Myocardial scar segmentation in LGE-MRI using fractal analysis and random forest classification. 2018 24th International Conference on Pattern Recognition (ICPR), August 2018:3168–3173. https://doi.org/10.1109/ICPR.2018.8545636.

56.

Moccia S, Banali R, Martini C, Muscogiuri G, Pontone G, Pepi M, et al. Development and testing of a deep learning-based strategy for scar segmentation on CMR-LGE images. Magn Reson Mater Physics, Biol Med. 2019;32(2):187(9). https://doi.org/10.1007/s10334-018-0718-4.CrossRef

57.

Zabihollahy F, White JA, Ukwatta E. Myocardial scar segmentation from magnetic resonance images using convolutional neural network. Proceedings Volume 10575, Medical Imaging 2018: Computer-aided diagnosis; 105752Z. https://doi.org/10.1117/12.2293518.

58.

Zabihollahy F, White JA, Ukwatta E. Convolutional neural network-based approach for segmentation of left ventricle myocardial scar from 3D late gadolinium enhancement MR images. Med Phys. 2019;46(4):1740–51. https://doi.org/10.1002/mp.13436.CrossRefPubMed

59.

Lau F, Hendriks T, Lieman-Sifry J, Berk N, Sall S, Golden D, et al. ScarGAN: chained generative adversarial networks to simulate pathological tissue on cardiovascular MR scans. Deep Learn. Med. Image Anal. Multimodal Learn. Clin. Decis. Support, 2018:343–350. https://doi.org/10.1007/978-3-030-00889-5_39,.

60.

Fahmy AS, Rausch J, Ulf N, Chan RH, Maron MS, Appelbaum E, et al. Automated cardiac MR scar quantification in hypertrophic cardiomyopathy using deep convolutional neural networks. JACC Cardiovasc Imaging. 2018;11(12):1917–8. https://doi.org/10.1016/j.jcmg.2018.04.030.CrossRefPubMedPubMedCentral

61••.

Zabihollahy F, Rajchl M, White JA, Ukwatta E. Fully automated segmentation of left ventricular scar from 3D late gadolinium enhancement magnetic resonance imaging using a cascaded multi-planar U-Net (CMPU-Net). Med Phys. 2020. https://doi.org/10.1002/mp.14022This document demonstrates the fully automated quantification of LV MF from 3D LGE CMRI with high accuracy is feasible.

62.

Chen M, Fang L, Zhuang Q, Liu H. Deep learning assessment of myocardial infarction from MR image sequences. IEEE Access. 2019;7:5438–46. https://doi.org/10.1109/ACCESS.2018.2889744.CrossRef

63.

Xu C, Xu L, Gao Z, Zhao S, Zhang H, Zhang Y, et al. Direct detection of pixel-level myocardial infarction areas via a deep-learning algorithm. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). 2017. arXiv:1706.03182v1.

64.

Zhang N, Yang G, Gao Z, Xu C, Zhang Y, Shi R, et al. Deep learning for diagnosis of chronic myocardial infarction on nonenhanced cardiac cine MRI. Radiology. 2019;291(3):606–17. https://doi.org/10.1148/radiol.2019182304.CrossRefPubMed

65.

Xu C, Howey J, Ohorodnyk P, Roth M, Zhang H, Li S. Segmentation and quantification of infarction without contrast agents via spatiotemporal generative adversarial learning. Med Image Anal. 2020;291(3):606–17. https://doi.org/10.1148/radiol.2019182304.CrossRef

66.

Sun, Wong AKC, Kamel MS. Classification of imbalanced data: a review. Int J Pattern Recognit Artif Intell. 2009;23(04):687–719. https://doi.org/10.1142/S0218001409007326.CrossRef

Title: Machine Learning-Based Segmentation of Left Ventricular Myocardial Fibrosis from Magnetic Resonance Imaging
Authors: Fatemeh Zabihollahy
S. Rajan
E. Ukwatta
Publication date: 01-08-2020
Publisher: Springer US
Keywords: Magnetic Resonance Imaging
Magnetic Resonance Imaging
Published in: Current Cardiology Reports / Issue 8/2020
Print ISSN: 1523-3782
Electronic ISSN: 1534-3170
DOI: https://doi.org/10.1007/s11886-020-01321-1

Keynote webinar | Spotlight on medication adherence

Springer Medicine

Machine Learning-Based Segmentation of Left Ventricular Myocardial Fibrosis from Magnetic Resonance Imaging

Abstract

Purpose of Review

Recent Findings

Summary

Keynote webinar | Spotlight on medication adherence

Springer Medicine

Abstract

Purpose of Review

Recent Findings

Summary

Please log in to get access to this content

Other articles of this Issue 8/2020

The PASCAL Device—Early Experience with a Leaflet Approximation Device: What Are the Benefits/Limitations Compared with the MitraClip?

Artificial Intelligence and Myocardial Contrast Enhancement Pattern

Heart and Brain: Complex Relationships for Left Ventricular Dysfunction

Appropriate Use Criteria for Echocardiography in the Era of Value-Based Care: Mission Accomplished or Future Mandates?

The Role of MRI in Prognostic Stratification of Cardiomyopathies

Metabolic Regulation of Human Pluripotent Stem Cell-Derived Cardiomyocyte Maturation