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.
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
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
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
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
Patel DA, Lavie CJ, Milani RV, Shah S, Gilliland Y (2009) Clinical implications of left atrial enlargement: a review. Ochsner J 9:191–196
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
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
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
Taskin V, Bates MC, Chillag SA (1991) Tracheal carinal angle and left atrial size. Arch Intern Med 151:307–308
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
Ngo LH (2019) Using a deep learning network to diagnose congestive heart failure. Radiology 290:523–524
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
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
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
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
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
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
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
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
Lim S, Kim I, Kim T, Kim C, Kim S (2019) Fast autoaugment. Advances in Neural Information Processing Systems, pp 6665–6675
Simonyan K, Zisserman A (2014) Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:14091556
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
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
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
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
Youden WJ (1950) Index for rating diagnostic tests. Cancer 3:32–35
Bartko JJ (1966) The intraclass correlation coefficient as a measure of reliability. Psychol Rep 19:3–11
Glover L, Baxley WA, Dodge HT (1973) A quantitative evaluation of heart size measurements from chest roentgenograms. Circulation 47:1289–1296
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
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
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
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
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
Hoffman RB, Rigler LG (1965) Evaluation of left ventricular enlargement in the lateral projection of the chest. Radiology 85:93–100
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
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
Henry WL, Morganroth J, Pearlman AS et al (1976) Relation between echocardiographically determined left atrial size and atrial fibrillation. Circulation 53:273–279
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
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
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
Authors and Affiliations
Corresponding author
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
About this article
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
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00330-021-07963-1