Skip to main content
SpringerLink
Log in

Deep learning algorithm to improve hypertrophic cardiomyopathy mutation prediction using cardiac cine images

  • Cardiac
  • Published:
European Radiology Aims and scope Submit manuscript

Abstract

Objectives

The high variability of hypertrophic cardiomyopathy (HCM) genetic phenotypes has prompted the establishment of risk-stratification systems that predict the risk of a positive genetic mutation based on clinical and echocardiographic profiles. This study aims to improve mutation-risk prediction by extracting cardiovascular magnetic resonance (CMR) morphological features using a deep learning algorithm.

Methods

We recruited 198 HCM patients (48% men, aged 47 ± 13 years) and divided them into training (147 cases) and test (51 cases) sets based on different genetic testing institutions and CMR scan dates (2012, 2013, respectively). All patients underwent CMR examinations, HCM genetic testing, and an assessment of established genotype scores (Mayo Clinic score I, Mayo Clinic score II, and Toronto score). A deep learning (DL) model was developed to classify the HCM genotypes, based on a nonenhanced four-chamber view of cine images.

Results

The areas under the curve (AUCs) for the test set were Mayo Clinic score I (AUC: 0.64, sensitivity: 64.29%, specificity: 47.83%), Mayo Clinic score II (AUC: 0.70, sensitivity: 64.29%, specificity: 65.22%), Toronto score (AUC: 0.74, sensitivity: 75.00%, specificity: 56.52%), and DL model (AUC: 0.80, sensitivity: 85.71%, specificity: 69.57%). The combination of the DL and the Toronto score resulted in a significantly higher predictive performance (AUC = 0.84, sensitivity: 83.33%, specificity: 78.26%), compared with Mayo I (p = 006), Mayo II (p = 022), and Toronto score (p = 0.029).

Conclusions

The combination of the DL model, based on nonenhanced cine CMR images and the Toronto score yielded significantly higher diagnostic performance in detecting HCM mutations.

Key Points

• Deep learning method could enable the extraction of image features from cine images.

• Deep learning method based on cine images performed better than established scores in identifying HCM patients with positive genotypes.

• The combination of the deep learning method based on cine images and the Toronto score could further improve the performance of the identification of HCM patients with positive genotypes.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4

Similar content being viewed by others

Abbreviations

AUC:

Area under the (receiver-operating characteristic) curve

CMR:

Cardiac magnetic resonance

DL:

Deep learning

FHHCM:

Family history of hypertrophic cardiomyopathy

FHSCD:

Family history of sudden cardiac death

HCM:

Hypertrophic cardiomyopathy

LSTM:

Long short-term memory

LVMWT:

Left ventricular maximal wall thickness

LVPWT:

Left ventricular posterior wall thickness

ROC:

Receiver-operating characteristic

ROIs:

Regions of interest

SCD:

sudden cardiac death

VUS:

Variants of uncertain significance

References

  1. Maron BJ (2002) Hypertrophic cardiomyopathy: a systematic review. JAMA 287:1308–1320

    PubMed  Google Scholar 

  2. Maron BJ, Maron MS (2013) Hypertrophic cardiomyopathy. Lancet 381:242–255

    Article  Google Scholar 

  3. Maron BJ, Maron MS, Semsarian C (2012) Genetics of hypertrophic cardiomyopathy after 20 years: clinical perspectives. J Am Coll Cardiol 60:705–715

    Article  Google Scholar 

  4. Elliott PM, Anastasakis A, Borger MA et al (2014) 2014 ESC guidelines on diagnosis and management of hypertrophic cardiomyopathy: the task force for the diagnosis and management of hypertrophic cardiomyopathy of the European Society of Cardiology (ESC). Eur Heart J 35:2733–2779

    Article  Google Scholar 

  5. Wordsworth S, Leal J, Blair E et al (2010) DNA testing for hypertrophic cardiomyopathy: a cost-effectiveness model. Eur Heart J 31:926–935

    Article  Google Scholar 

  6. Van Driest SL, Ommen SR, Tajik AJ, Gersh BJ, Ackerman MJ (2005) Sarcomeric genotyping in hypertrophic cardiomyopathy. Mayo Clin Proc 80:463–469

    Article  Google Scholar 

  7. Andersen PS, Havndrup O, Hougs L et al (2009) Diagnostic yield, interpretation, and clinical utility of mutation screening of sarcomere encoding genes in Danish hypertrophic cardiomyopathy patients and relatives. Hum Mutat 30:363–370

    Article  CAS  Google Scholar 

  8. Girolami F, Olivotto I, Passerini I et al (2006) A molecular screening strategy based on beta-myosin heavy chain, cardiac myosin binding protein C and troponin T genes in Italian patients with hypertrophic cardiomyopathy. J Cardiovasc Med (Hagerstown) 7:601–607

    Article  Google Scholar 

  9. Lopes LR, Rahman MS, Elliott PM (2013) A systematic review and meta-analysis of genotype-phenotype associations in patients with hypertrophic cardiomyopathy caused by sarcomeric protein mutations. Heart 99:1800–1811

    Article  Google Scholar 

  10. Lopes LR, Syrris P, Guttmann OP et al (2015) Novel genotype-phenotype associations demonstrated by high-throughput sequencing in patients with hypertrophic cardiomyopathy. Heart 101:294–301

    Article  CAS  Google Scholar 

  11. Gruner C, Ivanov J, Care M et al (2013) Toronto hypertrophic cardiomyopathy genotype score for prediction of a positive genotype in hypertrophic cardiomyopathy. Circ Cardiovasc Genet 6:19–26

    Article  Google Scholar 

  12. Li Y, Qian Z, Xu K et al (2018) MRI features predict p53 status in lower-grade gliomas via a machine-learning approach. Neuroimage Clin 17:306–311

    Article  Google Scholar 

  13. Li Y, Liu X, Xu K et al (2018) MRI features can predict EGFR expression in lower grade gliomas: a voxel-based radiomic analysis. Eur Radiol 28:356–362

    Article  Google Scholar 

  14. Zhang N, Yang G (2019) Deep learning for diagnosis of chronic myocardial infarction on nonenhanced cardiac cine MRI. Radiology 291:606–617

    Article  Google Scholar 

  15. Baessler B, Luecke C, Lurz J et al (2018) Cardiac MRI texture analysis of T1 and T2 maps in patients with infarctlike acute myocarditis. Radiology 289:357–365

    Article  Google Scholar 

  16. Cheng S, Fang M, Cui C et al (2018) LGE-CMR-derived texture features reflect poor prognosis in hypertrophic cardiomyopathy patients with systolic dysfunction: preliminary results. Eur Radiol 28:4615–4624

    Article  Google Scholar 

  17. Bello GA, Dawes TJW, Duan J et al (2019) Deep learning cardiac motion analysis for human survival prediction. Nat Mach Intell 1:95–104

    Article  Google Scholar 

  18. Bai W, Sinclair M, Tarroni G et al (2018) Automated cardiovascular magnetic resonance image analysis with fully convolutional networks. J Cardiovasc Magn Reson 20:65

    Article  Google Scholar 

  19. Bos JM, Will ML, Gersh BJ, Kruisselbrink TM, Ommen SR, Ackerman MJ (2014) Characterization of a phenotype-based genetic test prediction score for unrelated patients with hypertrophic cardiomyopathy. Mayo Clin Proc 89:727–737

  20. Van Driest SL, Ommen SR, Tajik AJ, Gersh BJ, Ackerman MJ (2005) Yield of genetic testing in hypertrophic cardiomyopathy. Mayo Clin Proc 80:739–744

    Article  Google Scholar 

  21. Richards S, Aziz N, Bale S et al (2015) Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Med 17:405–424

    Article  Google Scholar 

  22. Binder J, Ommen SR, Gersh BJ et al (2006) Echocardiography-guided genetic testing in hypertrophic cardiomyopathy: septal morphological features predict the presence of myofilament mutations. Mayo Clin Proc 81:459–467

    Article  Google Scholar 

  23. Chen LC, Zhu Y, Papandreou G, Schroff F, Adam H (2018) Encoder-decoder with atrous separable convolution for semantic image segmentation. European Conference on Computer Vision (ECCV), Munich. Available via https://doi.org/10.1007/978-3-030-01234-2_49. Accessed 06 October 2018

  24. Szegedy C, Ioffe S, Vanhoucke VAlemi A (2016) Inception-v4, Inception-ResNet and the impact of residual connections on learning. National Conference on Artificial Intelligence(AAAI) , Phoenix. Available via arXiv:1602.07261. Accessed: February 2016

  25. Deng J, Dong W, Socher R, Li LJ, Li K, Li FF (2009) ImageNet: a large-scale hierarchical image database. Computer Vision & Pattern Recognition (CVPR), Miami. Available via https://doi.org/10.1109/CVPR.2009.5206848. Accessed 18 August 2009

  26. Gers FA, Schmidhuber J, Cummins F (2000) Learning to forget: continual prediction with LSTM. Neural Comput 12:2451–2471

    Article  CAS  Google Scholar 

  27. Kingma DP, Ba J (2015) Adam: a method for stochastic optimization. International Conference on Learning Representations (ICLR), San Diego. Available via arXiv:1412.6980. Accessed 2015

  28. Chollet F (2018) Keras: the python deep learning library. Astrophysics Source Code Library. Available via https://keras.io/getting-started/faq/#how-should-i-cite-keras. Accessed June 2018

  29. Bos JM, Towbin JA, Ackerman MJ (2009) Diagnostic, prognostic, and therapeutic implications of genetic testing for hypertrophic cardiomyopathy. J Am Coll Cardiol 54:201–211

    Article  CAS  Google Scholar 

  30. Baessler B, Mannil M, Oebel S, Maintz D, Alkadhi H, Manka R (2018) Subacute and chronic left ventricular myocardial scar: accuracy of texture analysis on nonenhanced cine MR images. Radiology 286:103–112

    Article  Google Scholar 

  31. Tao Q, Yan W, Wang Y et al (2019) Deep learning-based method for fully automatic quantification of left ventricle function from cine MR images: a multivendor, multicenter study. Radiology 290:81–88

    Article  Google Scholar 

  32. Ellims AH, Iles LM, Ling LH et al (2014) A comprehensive evaluation of myocardial fibrosis in hypertrophic cardiomyopathy with cardiac magnetic resonance imaging: linking genotype with fibrotic phenotype. Eur Heart J Cardiovasc Imaging 15:1108–1116

    Article  Google Scholar 

  33. Fujita T, Fujino N, Anan R et al (2013) Sarcomere gene mutations are associated with increased cardiovascular events in left ventricular hypertrophy: results from multicenter registration in Japan. JACC Heart Fail 1:459–466

    Article  Google Scholar 

  34. Olivotto I, Girolami F, Ackerman MJ et al (2008) Myofilament protein gene mutation screening and outcome of patients with hypertrophic cardiomyopathy. Mayo Clin Proc 83:630–638

    Article  CAS  Google Scholar 

Download references

Funding

This paper is supported by the National Natural Science Foundation of China under Grant Nos. 81922040, 81930053, 81227901, 81527805, and 81772012; the major international (regional) joint research project of National Science Foundation of China under Grant No. 81620108015; Beijing Natural Science Foundation (under Grant No. 7182109); National Key Research and Development Plan of China (under Grant Nos. 2017YFA0205200, 2016YFA0100900, and 2016YFA0100902); and Youth Innovation Promotion Association CAS (Grant No. 2019136).

Author information

Author notes

  1. Hongyu Zhou and Lu Li contributed equally to this work.

Authors and Affiliations

  1. Paul C. Lauterbur Research Center for Biomedical Imaging, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, SZ University Town, Shenzhen, 518055, China

    Hongyu Zhou, Kankan Zhao & Hairong Zheng

  2. CAS Key Lab of Molecular Imaging, Institute of Automation, Chinese Academy of Sciences, Beijing, 100190, China

    Hongyu Zhou, Zhenyu Liu & Jie Tian

  3. University of Chinese Academy of Sciences, Beijing, 100080, China

    Hongyu Zhou, Zhenyu Liu & Jie Tian

  4. Department of Magnetic Resonance Imaging, Fuwai Hospital, State Key Laboratory of Cardiovascular Disease, National Center for Cardiovascular Diseases of China, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, 100037, China

    Lu Li, Xiuyu Chen, Minjie Lu, Gang Yin & Shihua Zhao

  5. Department of Cardiology, Fuwai Hospital, National Center for Cardiovascular Diseases of China, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, 100037, China

    Lei Song

  6. Beijing Advanced Innovation Center for Big Data-Based Precision Medicine, School of Medicine, Beihang University, 100191, Beijing, China

    Jie Tian

  7. Engineering Research Center of Molecular and Neuro Imaging of Ministry of Education, School of Life Science and Technology, Xidian University, 710126, Xi’an, China

    Jie Tian

Authors
  1. Hongyu Zhou
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Lu Li
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Zhenyu Liu
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Kankan Zhao
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Xiuyu Chen
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Minjie Lu
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Gang Yin
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Lei Song
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. Shihua Zhao
    View author publications

    You can also search for this author in PubMed Google Scholar

  10. Hairong Zheng
    View author publications

    You can also search for this author in PubMed Google Scholar

  11. Jie Tian
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

Lu Li made contributions to the conception and design of the study; Lu Li and Hongyu Zhou drafted the manuscript; Kankan Zhao and Lu Li were responsible for statistical analysis of the data, Zhenyu Liu, Min-jie Lu, and Xiuyu Chen made critical revisions to the manuscript; Lu Li and Lei Song collected conventional CMR and genetic data; Gang Yin was in charge of image segmentation; Shihua zhao, Jie Tian, and Hairong Zheng made a contribution to study conduction. All authors read and approved the final manuscript.

Corresponding authors

Correspondence to Shihua Zhao or Jie Tian.

Ethics declarations

Guarantor

The scientific guarantor of this publication is Shihua Zhao.

Conflict of interest

The authors of this manuscript declare no relationships with any companies whose products or services may be related to the subject matter of the article.

Statistics and biometry

No complex statistical methods were necessary for this paper.

Informed consent

Written informed consent was obtained from all subjects (patients) in this study.

Ethical approval

Institutional Review Board approval was obtained.

Methodology

• Retrospective

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

ESM 1

(DOCX 21.3 kb)

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Zhou, H., Li, L., Liu, Z. et al. Deep learning algorithm to improve hypertrophic cardiomyopathy mutation prediction using cardiac cine images. Eur Radiol 31, 3931–3940 (2021). https://doi.org/10.1007/s00330-020-07454-9

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00330-020-07454-9

Keywords

Search

Navigation