Skip to main content
SpringerLink
Log in

Automatic prediction of left cardiac chamber enlargement from chest radiographs using convolutional neural network

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

ABSTRACT

Objective

To develop deep learning–based cardiac chamber enlargement-detection algorithms for left atrial (DLCE-LAE) and ventricular enlargement (DLCE-LVE), on chest radiographs

Methods

For training and internal validation of DLCE-LAE and -LVE, 5,045 chest radiographs (CRs; 2,463 normal and 2,393 LAE) and 1,012 CRs (456 normal and 456 LVE) matched with the same-day echocardiography were collected, respectively. External validation was performed using 107 temporally independent CRs. Reader performance test was conducted using the external validation dataset by five cardiothoracic radiologists without and with the results of DLCE. Classification performance of DLCE was evaluated and compared with those of the readers and conventional radiographic features, including cardiothoracic ratio, carinal angle, and double contour. In addition, DLCE-LAE was tested on 5,277 CRs from a healthcare screening program cohort.

Results

DLCE-LAE showed areas under the receiver operating characteristics curve (AUROCs) of 0.858 on external validation. On reader performance test, DLCE-LAE showed better results than pooled radiologists (AUROC 0.858 vs. 0.651; p < .001) and significantly increased their performance when used as a second reader (AUROC 0.651 vs. 0.722; p < .001). DLCE-LAE also showed a significantly higher AUROC than conventional radiographic findings (AUROC 0.858 vs. 0.535–0.706; all ps < .01). In the healthcare screening cohort, DLCE-LAE successfully detected 71.0% (142/200) CRs with moderate-to-severe LAE (93.5% [29/31] of severe cases), while yielding 11.8% (492/4,184) false-positive rate. DLCE-LVE showed AUROCs of 0.966 and 0.594 on internal and external validation, respectively.

Conclusion

DLCE-LAE outperformed and improved cardiothoracic radiologists’ performance in detecting LAE and showed promise in screening individuals with moderate-to-severe LAE in a healthcare screening cohort.

Key Points

• Our deep learning algorithm outperformed cardiothoracic radiologists in detecting left atrial enlargement on chest radiographs.

• Cardiothoracic radiologists improved their performance in detecting left atrial enlargement when aided by the algorithm.

• On a healthcare-screening cohort, our algorithm detected 71.0% (142/200) radiographs with moderate-to-severe left atrial enlargement while yielding 11.8% (492/4,184) false-positive rate.

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
Fig. 5

Similar content being viewed by others

Abbreviations

ACA/AHA:

American College of Cardiology/American Heart Association

AUROC:

Area under the receiver operating characteristic curve

BNP:

Brain natriuretic peptide

CNN:

Convolutional neural network

CR:

Chest radiograph

DLCE:

Deep learning–based cardiac chamber enlargement-detection algorithms

LAd:

Anteroposterior diameter of left atrium on echocardiography

LAE:

Left atrial enlargement

LVE:

Left ventricular enlargement

ROC:

Receiver operating characteristic curve

References

  1. Mettler FA Jr, Mahesh M, Bhargavan-Chatfield M et al (2020) Patient exposure from radiologic and nuclear medicine procedures in the United States: procedure volume and effective dose for the period 2006–2016. Radiology 295:418–427

    Article  Google Scholar 

  2. Kelly AM, Keijzers G, Klim S et al (2017) An observational study of dyspnea in emergency departments: the Asia, Australia, and New Zealand Dyspnea in Emergency Departments Study (AANZDEM). Acad Emerg Med 24:328–336

    Article  Google Scholar 

  3. Gardin JM, McClelland R, Kitzman D et al (2001) M-mode echocardiographic predictors of six-to seven-year incidence of coronary heart disease, stroke, congestive heart failure, and mortality in an elderly cohort (the Cardiovascular Health Study). Am J Cardiol 87:1051–1057

    Article  CAS  Google Scholar 

  4. Patel DA, Lavie CJ, Milani RV, Shah S, Gilliland Y (2009) Clinical implications of left atrial enlargement: a review. Ochsner J 9:191–196

    PubMed  PubMed Central  Google Scholar 

  5. Sahin H, Chowdhry DN, Olsen A, Nemer O, Wahl L (2019) Is there any diagnostic value of anteroposterior chest radiography in predicting cardiac chamber enlargement? Int J Cardiovasc Imaging 35:195–206

    Article  Google Scholar 

  6. Ernst ER, Shub C, Bailey KR, Brown LR, Redfield MM (2001) Radiographic measurements of cardiac size as predictors of outcome in patients with dilated cardiomyopathy. J Card Fail 7:13–20

    Article  CAS  Google Scholar 

  7. Dimopoulos K, Giannakoulas G, Bendayan I et al (2013) Cardiothoracic ratio from postero-anterior chest radiographs: a simple, reproducible and independent marker of disease severity and outcome in adults with congenital heart disease. Int J Cardiol 166:453–457

    Article  Google Scholar 

  8. Taskin V, Bates MC, Chillag SA (1991) Tracheal carinal angle and left atrial size. Arch Intern Med 151:307–308

    Article  CAS  Google Scholar 

  9. Murray J, Brown A, Anagnostou E, Senior R (1995) Widening of the tracheal bifurcation on chest radiographs: value as a sign of left atrial enlargement. AJR Am J Roentgenol 164:1089–1092

    Article  CAS  Google Scholar 

  10. Ngo LH (2019) Using a deep learning network to diagnose congestive heart failure. Radiology 290:523–524

    Article  Google Scholar 

  11. Seah JC, Tang JS, Kitchen A, Gaillard F, Dixon AF (2019) Chest radiographs in congestive heart failure: visualizing neural network learning. Radiology 290:514–522

    Article  Google Scholar 

  12. Duler L, LeBlanc N, Cooley S, Nemanic S, Scollan K (2018) Interreader agreement of radiographic left atrial enlargement in dogs and comparison to echocardiographic left atrial assessment. J Vet Cardiol 20:319–329

    Article  CAS  Google Scholar 

  13. Toba S, Mitani Y, Yodoya N et al (2020) Prediction of pulmonary to systemic flow ratio in patients with congenital heart disease using deep learning–based analysis of chest radiographs. JAMA Cardiol 5:449–457

    Article  Google Scholar 

  14. Lang (2016) Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging (vol 28, pg 1, 2015). J Am Soc Echocardiogr 29:276–276

    Article  Google Scholar 

  15. Lang RM, Bierig M, Devereux RB et al (2006) American Society of Echocardiography’s Nomenclature and Standards Committee; Task Force on Chamber Quantification; American College of Cardiology Echocardiography Committee; American Heart Association; European Association of Echocardiography, European society of Cardiology. Recommendations for chamber quantification. Eur J Echocardiogr 7:79–108

    Article  Google Scholar 

  16. Bouzas-Mosquera A, Broullón FJ, Álvarez-García N et al (2011) Left atrial size and risk for all-cause mortality and ischemic stroke. CMAJ 183:E657–E664

  17. Tsang TS, Barnes ME, Gersh BJ, Bailey KR, Seward JB (2002) Left atrial volume as a morphophysiologic expression of left ventricular diastolic dysfunction and relation to cardiovascular risk burden. Am J Cardiol 90:1284–1289

    Article  Google Scholar 

  18. Cubuk ED, Zoph B, Mane D, Vasudevan V, Le QV (2019) Autoaugment: learning augmentation strategies from data. Proceedings of the IEEE conference on computer vision and pattern recognition, pp 113–123

  19. Lim S, Kim I, Kim T, Kim C, Kim S (2019) Fast autoaugment. Advances in Neural Information Processing Systems, pp 6665–6675

  20. Simonyan K, Zisserman A (2014) Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:14091556

  21. Huang G, Liu Z, Van Der Maaten L, Weinberger KQ (2017) Densely connected convolutional networks. Proceedings of the IEEE conference on computer vision and pattern recognition, pp 4700–4708

  22. Szegedy C, Vanhoucke V, Ioffe S, Shlens J, Wojna Z (2016) Rethinking the inception architecture for computer vision. Proceedings of the IEEE conference on computer vision and pattern recognition, pp 2818–2826

  23. He K, Zhang X, Ren S, Sun J (2016) Deep residual learning for image recognition. Proceedings of the IEEE conference on computer vision and pattern recognition, pp 770–778

  24. Goff DC, Lloyd-Jones DM, Bennett G et al (2014) 2013 ACC/AHA guideline on the assessment of cardiovascular risk: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. JACC 63:2935–2959

    Article  Google Scholar 

  25. Youden WJ (1950) Index for rating diagnostic tests. Cancer 3:32–35

    Article  CAS  Google Scholar 

  26. Bartko JJ (1966) The intraclass correlation coefficient as a measure of reliability. Psychol Rep 19:3–11

    Article  CAS  Google Scholar 

  27. Glover L, Baxley WA, Dodge HT (1973) A quantitative evaluation of heart size measurements from chest roentgenograms. Circulation 47:1289–1296

    Article  CAS  Google Scholar 

  28. Konstam MA, Cohen SR, Weiland DS et al (1986) Relative contribution of inotropic and vasodilator effects to amrinone-induced hemodynamic improvement in congestive heart failure. Am J Cardiol 57:242–248

    Article  CAS  Google Scholar 

  29. McGee WT (2009) A simple physiologic algorithm for managing hemodynamics using stroke volume and stroke volume variation: physiologic optimization program. J Intensive Care Med 24:352–360

    Article  Google Scholar 

  30. Philbin EF, Garg R, Danisa K et al (1998) The relationship between cardiothoracic ratio and left ventricular ejection fraction in congestive heart failure. Arch Intern Med 158:501–506

    Article  CAS  Google Scholar 

  31. Freeman V, Mutatiri C, Pretorius M, Doubell A (2003) Evaluation of left ventricular enlargement in the lateral position of the chest using the Hoffman and Rigler sign: cardiovascular topics. Cardiovasc J S Afr 14:134–137

    CAS  PubMed  Google Scholar 

  32. Eyler WR, Wayne DL, Rhodenbaugh JE (1959) The importance of the lateral view in the evaluation of left ventricular enlargement in rheumatic heart disease. Radiology 73:56–61

    Article  CAS  Google Scholar 

  33. Hoffman RB, Rigler LG (1965) Evaluation of left ventricular enlargement in the lateral projection of the chest. Radiology 85:93–100

    Article  CAS  Google Scholar 

  34. Patel DA, Lavie CJ, Milani RV, Ventura HO (2008) Left atrial volume index and left ventricular geometry independently predict mortality in 47,865 patients with preserved ejection fraction. Circulation 118:S_843

    Google Scholar 

  35. Geske JB, Sorajja P, Nishimura RA, Ommen SR (2009) The relationship of left atrial volume and left atrial pressure in patients with hypertrophic cardiomyopathy: an echocardiographic and cardiac catheterization study. J Am Soc Echocardiogr 22:961–966

    Article  Google Scholar 

  36. Henry WL, Morganroth J, Pearlman AS et al (1976) Relation between echocardiographically determined left atrial size and atrial fibrillation. Circulation 53:273–279

    Article  CAS  Google Scholar 

  37. Bayes-Genis A, Vazquez R, Puig T et al (2007) Left atrial enlargement and NT-proBNP as predictors of sudden cardiac death in patients with heart failure. Eur J Heart Fail 9:802–807

    Article  CAS  Google Scholar 

  38. Schuster A, Backhaus SJ, Stiermaier T et al (2019) Left atrial function with MRI enables prediction of cardiovascular events after myocardial infarction: insights from the AIDA STEMI and TATORT NSTEMI trials. Radiology 293:292–302

    Article  Google Scholar 

Download references

Funding

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Ministry of Science and ICT (MSIT) (grant number: NRF-2018R1A5A1060031) and the Seoul National University Hospital Research Fund (grant number: 03-2019-0190).

Author information

Author notes

  1. Ju Gang Nam and Jinwook Kim contributed equally to this work.

    Min‐Soo Kim and Chang Min Park contributed equally to this work.

Authors and Affiliations

  1. Radiology, Seoul National University Hospital, Seoul, Republic of Korea

    Ju Gang Nam, Hyewon Choi, Eui Jin Hwang, Jin Mo Goo, Eun-Ah Park & Chang Min Park

  2. College of Medicine, Seoul National University, Seoul, 03080, Republic of Korea

    Ju Gang Nam, Hyewon Choi, Eui Jin Hwang, Jin Mo Goo, Eun-Ah Park & Chang Min Park

  3. Information and Communication Engineering, Daegu Gyeongbuk Institute of Science & Technology (DGIST), Daegu, 42988, Republic of Korea

    Jinwook Kim, Keonwoo Noh & Hyun-Lim Yang

  4. Radiology, Busan Paik Hospital and Inje University College of Medicine, Busan, 47392, Republic of Korea

    Da Som Kim

  5. Radiology, Hanyang University Medical Center and College of Medicine, Seoul, 04763, Republic of Korea

    Seung-Jin Yoo

  6. Institute of Radiation Medicine, Seoul National University Medical Research Center, Seoul, 03080, Republic of Korea

    Jin Mo Goo, Eun-Ah Park & Chang Min Park

  7. Radiology, Seoul National University Hospital Healthcare System Gangnam Center, Seoul, 06236, Republic of Korea

    Hye Young Sun

  8. School of Computing, Korea Advanced Institute of Science and Technology (KAIST), Daejeon, 34141, Republic of Korea

    Min-Soo Kim

  9. Institute of Medical and Biological Engineering, Seoul National University Medical Research Center, 101 Daehak-ro, Jongno-gu, Seoul, 03080, Republic of Korea

    Chang Min Park

Authors
  1. Ju Gang Nam
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Jinwook Kim
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Keonwoo Noh
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Hyewon Choi
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Da Som Kim
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Seung-Jin Yoo
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Hyun-Lim Yang
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Eui Jin Hwang
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. Jin Mo Goo
    View author publications

    You can also search for this author in PubMed Google Scholar

  10. Eun-Ah Park
    View author publications

    You can also search for this author in PubMed Google Scholar

  11. Hye Young Sun
    View author publications

    You can also search for this author in PubMed Google Scholar

  12. Min-Soo Kim
    View author publications

    You can also search for this author in PubMed Google Scholar

  13. Chang Min Park
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Chang Min Park.

Ethics declarations

Guarantor

The scientific guarantor of this publication is Chang Min Park.

Conflict of interest

All 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 waived by the Institutional Review Board.

Methodology

• retrospective

• observational study

• performed at one institution

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 3342 kb)

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Nam, J.G., Kim, J., Noh, K. et al. Automatic prediction of left cardiac chamber enlargement from chest radiographs using convolutional neural network. Eur Radiol 31, 8130–8140 (2021). https://doi.org/10.1007/s00330-021-07963-1

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00330-021-07963-1

Keywords

Search

Navigation