Utility of deep learning networks for the generation of artificial cardiac magnetic resonance images in congenital heart disease
Gerhard-Paul Diller, Julius Vahle, Robert Radke, Maria Luisa Benesch Vidal, Alicia Jeanette Fischer, Ulrike M M Bauer, Samir Sarikouch, Felix Berger, Philipp Beerbaum, Helmut Baumgartner, Stefan Orwat, German Competence Network for Congenital Heart Defects Investigators, Gerhard-Paul Diller, Julius Vahle, Robert Radke, Maria Luisa Benesch Vidal, Alicia Jeanette Fischer, Ulrike M M Bauer, Samir Sarikouch, Felix Berger, Philipp Beerbaum, Helmut Baumgartner, Stefan Orwat, German Competence Network for Congenital Heart Defects Investigators
Abstract
Background: Deep learning algorithms are increasingly used for automatic medical imaging analysis and cardiac chamber segmentation. Especially in congenital heart disease, obtaining a sufficient number of training images and data anonymity issues remain of concern.
Methods: Progressive generative adversarial networks (PG-GAN) were trained on cardiac magnetic resonance imaging (MRI) frames from a nationwide prospective study to generate synthetic MRI frames. These synthetic frames were subsequently used to train segmentation networks (U-Net) and the quality of the synthetic training images, as well as the performance of the segmentation network was compared to U-Net-based solutions trained entirely on patient data.
Results: Cardiac MRI data from 303 patients with Tetralogy of Fallot were used for PG-GAN training. Using this model, we generated 100,000 synthetic images with a resolution of 256 × 256 pixels in 4-chamber and 2-chamber views. All synthetic samples were classified as anatomically plausible by human observers. The segmentation performance of the U-Net trained on data from 42 separate patients was statistically significantly better compared to the PG-GAN based training in an external dataset of 50 patients, however, the actual difference in segmentation quality was negligible (< 1% in absolute terms for all models).
Conclusion: We demonstrate the utility of PG-GANs for generating large amounts of realistically looking cardiac MRI images even in rare cardiac conditions. The generated images are not subject to data anonymity and privacy concerns and can be shared freely between institutions. Training supervised deep learning segmentation networks on this synthetic data yielded similar results compared to direct training on original patient data.
Trial registration: ClinicalTrials.gov NCT00266188.
Conflict of interest statement
Not applicable.
Figures
References
- Topol EJ. High-performance medicine: the convergence of human and artificial intelligence. Nat Med. 2019;25(1):44–56. doi: 10.1038/s41591-018-0300-7.
- Hosny A, Parmar C, Quackenbush J, Schwartz LH, Aerts HJWL. Artificial intelligence in radiology. Nat Rev Cancer. 2018;18(8):500–510. doi: 10.1038/s41568-018-0016-5.
- Diller G-P, Babu-Narayan S, Li W, Radojevic J, Kempny A, Uebing A, et al. Utility of machine learning algorithms in assessing patients with a systemic right ventricle. Eur Heart J Cardiovasc Imaging. 2019;20(8):925–931. doi: 10.1093/ehjci/jey211.
- Bai W, Sinclair M, Tarroni G, et al. Automated cardiovascular magnetic resonance image analysis with fully convolutional networks. J Cardiovasc Magn Reson. 2018;20(1):65. doi: 10.1186/s12968-018-0471-x.
- Karras T, Aila T, Laine S, Lehtinen J. Progressive growing of GANs for improved quality, stability, and variation. 2017;arXiv:1710.10196.
- Diller GP, Orwat S, Vahle J, et al. Prediction of prognosis in patients with tetralogy of Fallot based on deep learning imaging analysis. Heart. 2020. 10.1136/heartjnl-2019-315962.
- Volotat AK. GANs library [Internet]. 2020. pp. 1–2. Available from: . Accessed 07 July 2020.
- Gulrajani I, Ahmed F, Arjovsky M, Dumoulin V, Courville AC. Advances in neural information processing systems. 2017. Improved training of wasserstein gans; pp. 5767–5777.
- Diller G-P, Kempny A, Alonso-Gonzalez R, Swan L, Uebing A, Li W, et al. Survival prospects and circumstances of death in contemporary adult congenital heart disease patients under follow-up at a large tertiary Centre. Circulation. 2015;132(22):2118–2125. doi: 10.1161/CIRCULATIONAHA.115.017202.
- Beerbaum P, Barth P, Kropf S, Sarikouch S, Kelter-Kloepping A, Franke D, et al. Cardiac function by MRI in congenital heart disease: impact of consensus training on interinstitutional variance. J Magn Reson Imaging. 2009;30(5):956–966. doi: 10.1002/jmri.21948.
- Sarikouch S, Koerperich H, Dubowy K-O, Boethig D, Boettler P, Mir TS, et al. Impact of gender and age on cardiovascular function late after repair of tetralogy of Fallot: percentiles based on cardiac magnetic resonance. Circ Cardiovasc Imaging. 2011;4(6):703–711. doi: 10.1161/CIRCIMAGING.111.963637.
- Orwat S, Diller G-P, Kempny A, Radke R, Peters B, Kühne T, et al. Myocardial deformation parameters predict outcome in patients with repaired tetralogy of Fallot. Heart. 2016;102(3):209–215. doi: 10.1136/heartjnl-2015-308569.
- Ronneberger O, Fischer P, Brox T. U-net: convolutional networks for biomedical image segmentation. Cham: Springer International Publishing; 2015. pp. 234–241.
- Zhang J, Gajjala S, Agrawal P, Tison GH, Hallock LA, Beussink-Nelson L, et al. Fully automated echocardiogram interpretation in clinical practice: feasibility and diagnostic accuracy. Circulation. 2018;138(16):1623–1635. doi: 10.1161/CIRCULATIONAHA.118.034338.
- Chen C, Qin C, Qiu H, Tarroni G, Duan J, Bai W, Rueckert D. Deep learning for cardiac image segmentation: a review. 2019. arXiv:191103723. arXiv preprint.
- Goodfellow I, Pouget-Abadie J, Mirza M, Xu B, Warde-Farley D, Ozair S, Courville A, Bengio Y. Generative adversarial nets. In: Advances in neural information processing systems. 2014;2014:2672–80.
- Zhao M, Liu X, Liu H, Wong KKL. Super-resolution of cardiac magnetic resonance images using Laplacian pyramid based on generative adversarial networks. Comput Med Imaging Graph. 2020;80:101698. doi: 10.1016/j.compmedimag.2020.101698.
- Diller G-P, Lammers AE, Babu-Narayan S, Li W, Radke RM, Baumgartner H, et al. Denoising and artefact removal for transthoracic echocardiographic imaging in congenital heart disease: utility of diagnosis specific deep learning algorithms. Int J Cardiovasc Imaging. 2019;35(12):2189–2196. doi: 10.1007/s10554-019-01671-0.
- Jin C-B, Kim H, Liu M, Jung W, Joo S, Park E, et al. Deep CT to MR synthesis using paired and unpaired data. Sensors (Basel) 2019;19(10):2361. doi: 10.3390/s19102361.
- Shin HC, Tenenholtz NA, Rogers JK, Schwarz CG, Senjem ML, Gunter JL, et al. Medical image synthesis for data augmentation and anonymization using generative adversarial networks. Cham: Springer International Publishing; 2018. pp. 1–11.
Source: PubMed