A low-cost texture-based pipeline for predicting myocardial tissue remodeling and fibrosis using cardiac ultrasound

Nobuyuki Kagiyama, Sirish Shrestha, Jung Sun Cho, Muhammad Khalil, Yashbir Singh, Abhiram Challa, Grace Casaclang-Verzosa, Partho P Sengupta, Nobuyuki Kagiyama, Sirish Shrestha, Jung Sun Cho, Muhammad Khalil, Yashbir Singh, Abhiram Challa, Grace Casaclang-Verzosa, Partho P Sengupta

Abstract

Background: Maturation of ultrasound myocardial tissue characterization may have far-reaching implications as a widely available alternative to cardiac magnetic resonance (CMR) for risk stratification in left ventricular (LV) remodeling.

Methods: We extracted 328 texture-based features of myocardium from still ultrasound images. After we explored the phenotypes of myocardial textures using unsupervised similarity networks, global LV remodeling parameters were predicted using supervised machine learning models. Separately, we also developed supervised models for predicting the presence of myocardial fibrosis using another cohort who underwent cardiac magnetic resonance (CMR). For the prediction, patients were divided into a training and test set (80:20).

Findings: Texture-based tissue feature extraction was feasible in 97% of total 534 patients. Interpatient similarity analysis delineated two patient groups based on the texture features: one group had more advanced LV remodeling parameters compared to the other group. Furthermore, this group was associated with a higher incidence of cardiac deaths (p = 0.001) and major adverse cardiac events (p < 0.001). The supervised models predicted reduced LV ejection fraction (<50%) and global longitudinal strain (<16%) with area under the receiver-operator-characteristics curves (ROC AUC) of 0.83 and 0.87 in the hold-out test set, respectively. Furthermore, the presence of myocardial fibrosis was predicted from only ultrasound myocardial texture with an ROC AUC of 0.84 (sensitivity 86.4% and specificity 83.3%) in the test set.

Interpretation: Ultrasound texture-based myocardial tissue characterization identified phenotypic features of LV remodeling from still ultrasound images. Further clinical validation may address critical barriers in the adoption of ultrasound techniques for myocardial tissue characterization.

Funding: None.

Keywords: Clustering; Echocardiography; Machine learning; Radiomics; Tissue characterization.

Conflict of interest statement

Declaration of Competing Interest Nobuyuki Kagiyama is supported by a research grant from Hitachi Healthcare; Partho P. Sengupta is a consultant to Heart Sciences, Ultromics, and Kencor Health. The other authors have nothing to disclose

Copyright © 2020 The Author(s). Published by Elsevier B.V. All rights reserved.

Figures

Fig. 1
Fig. 1
Study process. This study consisted of three parts as shown in panel a to c. In the first part (panel a), we used unsupervised clustering to explore patient phenotypes and their clinical implications in total 405 patients with various stages of HF. Next, supervised machine learning was used to predict the functional (impaired LVEF and GLS) LV remodeling using the texture features in the same patient cohort (panel b). Finally, in an independent cohort of patients who underwent CMR and echocardiography, we explored the usefulness of texture-based supervised models for predicting myocardial fibrosis (panel c). WVU, West Virginia University; HF, heart failure; ML, machine learning; TDA, topological data analysis, LV, left ventricular; RFE, recursive feature elimination; CMR, cardiac magnetic resonance.
Fig. 2
Fig. 2
Patient similarity network based on myocardial texture features. Panel a: Extracted texture features were integrated using topological data analysis to create a patient similarity network. In the network, patients with similar features form a node, and adjacent nodes, including similar patients, are connected with edges. The network demonstrated the shape of a bar that was geometrically divided into two parts which had significantly different clinical and echocardiographic characteristics although the groups were created using only texture features. Panel b. Patients X, Y, Z, and W were identified in corresponding x, y, z, and w nodes, respectively. X and Y had a normal cardiac function and were found in cluster A, whereas Z and W were located in cluster B with significantly impaired cardiac function. GLS, global longitudinal strain; LVEF, left ventricular ejection fraction.
Fig. 3
Fig. 3
Clinical outcomes between clusters. Kaplan–Meier curve analyses showed that cluster B had a significantly higher incidence of cardiovascular death (panel a) and of the composite of cardiovascular death and major adverse cardiac events (MACE; panel b) compared with cluster A.
Fig. 4
Fig. 4
Direct prediction of impaired cardiac function. Panels a to c show receiver operator characteristics curves for predicting reduced LVEF (

Fig. 5

Prediction of myocardial fibrosis. The…

Fig. 5

Prediction of myocardial fibrosis. The upper panels show still ultrasound images and the…

Fig. 5
Prediction of myocardial fibrosis. The upper panels show still ultrasound images and the corresponding myocardial textures where the texture features were extracted. The lower panels show magnetic resonance images with late gadolinium enhancement.
Fig. 5
Fig. 5
Prediction of myocardial fibrosis. The upper panels show still ultrasound images and the corresponding myocardial textures where the texture features were extracted. The lower panels show magnetic resonance images with late gadolinium enhancement.

References

    1. Benjamin E.J., Virani S.S., Callaway C.W., Chamberlain A.M., Chang A.R., Cheng S. Heart disease and stroke statistics-2018 update: a report from the American Heart Association. Circulation. 2018;137(12):e67–e492.
    1. Jugdutt B.I. Ventricular remodeling after infarction and the extracellular collagen matrix: when is enough enough? Circulation. 2003;108(11):1395–1403.
    1. Konstam M.A., Kramer D.G., Patel A.R., Maron M.S., Udelson J.E. Left ventricular remodeling in heart failure: current concepts in clinical significance and assessment. JACC Cardiovasc Imaging. 2011;4(1):98–108.
    1. aus dem Siepen F., Buss S.J., Messroghli D., Andre F., Lossnitzer D., Seitz S. T1 mapping in dilated cardiomyopathy with cardiac magnetic resonance: quantification of diffuse myocardial fibrosis and comparison with endomyocardial biopsy. Eur Heart J Cardiovasc Imaging. 2015;16(2):210–216.
    1. Pashakhanloo F., Herzka D.A., Mori S., Zviman M., Halperin H., Gai N. Submillimeter diffusion tensor imaging and late gadolinium enhancement cardiovascular magnetic resonance of chronic myocardial infarction. J Cardiovasc Magn Reson. 2017;19(1):9.
    1. Johnson M.R., McPherson D.D., Fleagle S.R., Hunt M.M., Hiratzka L.F., Kerber R.E. Videodensitometric analysis of human coronary stenoses: validation in vivo by intraoperative high-frequency epicardial echocardiography. Circulation. 1988;77(2):328–336.
    1. Milunski M.R., Mohr G.A., Perez J.E., Vered Z., Wear K.A., Gessler C.J. Ultrasonic tissue characterization with integrated backscatter. Acute myocardial ischemia, reperfusion, and stunned myocardium in patients. Circulation. 1989;80(3):491–503.
    1. Aerts H.J., Velazquez E.R., Leijenaar R.T., Parmar C., Grossmann P., Carvalho S. Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach. Nat Commun. 2014;5:4006.
    1. Castellano G., Bonilha L., Li L.M., Cendes F. Texture analysis of medical images. Clin Radiol. 2004;59(12):1061–1069.
    1. Gillies R.J., Kinahan P.E., Hricak H. Radiomics: images are more than pictures, they are data. Radiology. 2016;278(2):563–577.
    1. Muller H., Michoux N., Bandon D., Geissbuhler A. A review of content-based image retrieval systems in medical applications-clinical benefits and future directions. Int J Med Inform. 2004;73(1):1–23.
    1. Hoit B.D. Strain and strain rate echocardiography and coronary artery disease. Circ Cardiovasc Imaging. 2011;4(2):179–190.
    1. Muraru D., Niero A., Rodriguez-Zanella H., Cherata D., Badano L. Three-dimensional speckle-tracking echocardiography: benefits and limitations of integrating myocardial mechanics with three-dimensional imaging. Cardiovasc Diagn Ther. 2018;8(1):101–117.
    1. Nioche C., Orlhac F., Boughdad S., Reuze S., Goya-Outi J., Robert C. LIFEx: a freeware for radiomic feature calculation in multimodality imaging to accelerate advances in the characterization of tumor heterogeneity. Cancer Res. 2018;78(16):4786–4789.
    1. Sun C., Wee W.G. Neighboring gray level dependence matrix for texture classification. Comput Vis Graphics Image Process. 1983;23(3):341–352.
    1. Simpson J., Jhund P.S., Silva Cardoso J., Martinez F., Mosterd A., Ramires F. Comparing LCZ696 with enalapril according to baseline risk using the MAGGIC and EMPHASIS-HF risk scores: an analysis of mortality and morbidity in PARADIGM-HF. J Am Coll Cardiol. 2015;66(19):2059–2071.
    1. Lang R.M., Badano L.P., Mor-Avi V., Afilalo J., Armstrong A., Ernande L. Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. J Am Soc Echocardiogr. 2015;28(1):1–39. e14.
    1. Lum P.Y., Singh G., Lehman A., Ishkanov T., Vejdemo-Johansson M., Alagappan M. Extracting insights from the shape of complex data using topology. Sci Rep. 2013;3:1236.
    1. Haralick R.M., Shanmugam K., Dinstein I. Textural features for image classification. IEEE Trans Syst Man Cybern. 1973;SMC-3(6):610–621.
    1. Guo W., Banerjee A.G. 2016 IEEE international symposium on assembly and manufacturing (ISAM) 2016. Toward automated prediction of manufacturing productivity based on feature selection using topological data analysis; pp. 21–22. Aug. 2016.
    1. Kellman P., Arai A.E., McVeigh E.R., Aletras A.H. Phase-sensitive inversion recovery for detecting myocardial infarction using gadolinium-delayed hyperenhancement. Magn Reson Med. 2002;47(2):372–383.
    1. Liu W., Fang W. 2008 ISECS international colloquium on computing, communication, control, and management. 2008. Adaptive spam filtering based on fingerprint vectors; pp. 3–4. Aug. 2008.
    1. Casaclang-Verzosa G., Shrestha S., Khalil M.J., Cho J.S., Tokodi M., Balla S. Network tomography for understanding phenotypic presentations in aortic stenosis. JACC Cardiovasc Imaging. 2019;12(2):236–248.
    1. Hinks T., Zhou X., Staples K., Dimitrov B., Manta A., Petrossian T. Multidimensional endotypes of asthma: topological data analysis of cross-sectional clinical, pathological, and immunological data. Lancet. 2015;385(Suppl 1):S42.
    1. Diddams S.A., Hollberg L., Mbele V. Molecular fingerprinting with the resolved modes of a femtosecond laser frequency comb. Nature. 2007;445(7128):627–630.
    1. Johannes M., Brase J.C., Frohlich H., Gade S., Gehrmann M., Falth M. Integration of pathway knowledge into a reweighted recursive feature elimination approach for risk stratification of cancer patients. Bioinformatics. 2010;26(17):2136–2144.
    1. Shahriari B., Swersky K., Wang Z., Adams R.P., Freitas N.d. Taking the human out of the loop: a review of bayesian optimization. Proc IEEE. 2016;104(1):148–175.
    1. Hutter F., Hoos H.H., Leyton-Brown K. Learning and intelligent optimization. Springer Berlin Heidelberg; Berlin, Heidelberg: 2011. Sequential model-based optimization for general algorithm configuration. 2011//
    1. Yancy C.W., Jessup M., Bozkurt B., Butler J., Casey D.E., Jr., Drazner M.H. 2013 ACCF/AHA guideline for the management of heart failure: a report of the American College of Cardiology Foundation/American Heart Association Task Force on practice guidelines. J Am Coll Cardiol. 2013;62(16):5.
    1. Cueto-Garcia L., Reeder G.S., Kyle R.A., Wood D.L., Seward J.B., Naessens J. Echocardiographic findings in systemic amyloidosis: spectrum of cardiac involvement and relation to survival. J. Am. Coll. Cardiol. 1985;6(4):737–743.
    1. Chandrasekaran K., Aylward P.E., Fleagle S.R., Burns T.L., Seward J.B., Tajik A.J. Feasibility of identifying amyloid and hypertrophic cardiomyopathy with the use of computerized quantitative texture analysis of clinical echocardiographic data. J. Am. Coll. Cardiol. 1989;13(4):832–840.
    1. Mor-Avi V., Lang R.M., Badano L.P., Belohlavek M., Cardim N.M., Derumeaux G. Current and evolving echocardiographic techniques for the quantitative evaluation of cardiac mechanics: ASE/EAE consensus statement on methodology and indications endorsed by the Japanese Society of Echocardiography. J Am Soc Echocardiogr. 2011;24(3):277–313.
    1. Parekh V.S., Jacobs M.A. Deep learning and radiomics in precision medicine. Expert Rev Precis Med Drug Dev. 2019;4(2):59–72.
    1. Wu X., Li Y., Chen X., Huang Y., He L., Zhao K. Deep learning features improve the performance of a radiomics signature for predicting KRAS status in patients with colorectal cancer. Acad Radiol. 2020 doi: . [Epub ahead of print]
    1. Wei L., Osman S., Hatt M., El Naqa I. Machine learning for radiomics-based multimodality and multiparametric modeling. Q J Nucl Med Mol Imaging. 2019;63(4):323–338.
    1. Oikonomou E.K., Williams M.C., Kotanidis C.P., Desai M.Y., Marwan M., Antonopoulos A.S. A novel machine learning-derived radiotranscriptomic signature of perivascular fat improves cardiac risk prediction using coronary CT angiography. Eur Heart J. 2019;40(43):3529–3543.
    1. Neisius U., El-Rewaidy H., Kucukseymen S., Tsao C.W., Mancio J., Nakamori S. Texture signatures of native myocardial T1 as novel imaging markers for identification of hypertrophic cardiomyopathy patients without scar. J Magn Reson Imaging. 2020 doi: . [Epub ahead of print]
    1. Liu Y., Hamilton J., Rajagopalan S., Seiberlich N. Cardiac magnetic resonance fingerprinting: technical overview and initial results. JACC Cardiovasc Imaging. 2018;11(12):1837–1853.

Source: PubMed

Sponsors and Collaborators

Medical Conditions

Drug Interventions

3
Subscribe