DOI QR코드

DOI QR Code

Automated Measurement of Native T1 and Extracellular Volume Fraction in Cardiac Magnetic Resonance Imaging Using a Commercially Available Deep Learning Algorithm

  • Suyon Chang (Department of Radiology, Seoul St. Mary's Hospital, College of Medicine, The Catholic University of Korea) ;
  • Kyunghwa Han (Department of Radiology, Research Institute of Radiological Science, Center for Clinical Imaging Data Science, Yonsei University College of Medicine) ;
  • Suji Lee (Department of Radiology, Research Institute of Radiological Science, Center for Clinical Imaging Data Science, Yonsei University College of Medicine) ;
  • Young Joong Yang (Phantomics, Inc.) ;
  • Pan Ki Kim (Phantomics, Inc.) ;
  • Byoung Wook Choi (Department of Radiology, Research Institute of Radiological Science, Center for Clinical Imaging Data Science, Yonsei University College of Medicine) ;
  • Young Joo Suh (Department of Radiology, Research Institute of Radiological Science, Center for Clinical Imaging Data Science, Yonsei University College of Medicine)
  • Received : 2022.07.20
  • Accepted : 2022.10.06
  • Published : 2022.12.01

Abstract

Objective: T1 mapping provides valuable information regarding cardiomyopathies. Manual drawing is time consuming and prone to subjective errors. Therefore, this study aimed to test a DL algorithm for the automated measurement of native T1 and extracellular volume (ECV) fractions in cardiac magnetic resonance (CMR) imaging with a temporally separated dataset. Materials and Methods: CMR images obtained for 95 participants (mean age ± standard deviation, 54.5 ± 15.2 years), including 36 left ventricular hypertrophy (12 hypertrophic cardiomyopathy, 12 Fabry disease, and 12 amyloidosis), 32 dilated cardiomyopathy, and 27 healthy volunteers, were included. A commercial deep learning (DL) algorithm based on 2D U-net (Myomics-T1 software, version 1.0.0) was used for the automated analysis of T1 maps. Four radiologists, as study readers, performed manual analysis. The reference standard was the consensus result of the manual analysis by two additional expert readers. The segmentation performance of the DL algorithm and the correlation and agreement between the automated measurement and the reference standard were assessed. Interobserver agreement among the four radiologists was analyzed. Results: DL successfully segmented the myocardium in 99.3% of slices in the native T1 map and 89.8% of slices in the post-T1 map with Dice similarity coefficients of 0.86 ± 0.05 and 0.74 ± 0.17, respectively. Native T1 and ECV showed strong correlation and agreement between DL and the reference: for T1, r = 0.967 (95% confidence interval [CI], 0.951-0.978) and bias of 9.5 msec (95% limits of agreement [LOA], -23.6-42.6 msec); for ECV, r = 0.987 (95% CI, 0.980-0.991) and bias of 0.7% (95% LOA, -2.8%-4.2%) on per-subject basis. Agreements between DL and each of the four radiologists were excellent (intraclass correlation coefficient [ICC] of 0.98-0.99 for both native T1 and ECV), comparable to the pairwise agreement between the radiologists (ICC of 0.97-1.00 and 0.99-1.00 for native T1 and ECV, respectively). Conclusion: The DL algorithm allowed automated T1 and ECV measurements comparable to those of radiologists.

Keywords

Acknowledgement

This research was supported by the Technology development Program (S3033533) funded by the Ministry of SMEs and Startups (MSS, Korea).

References

  1. Captur G, Manisty C, Moon JC. Cardiac MRI evaluation of myocardial disease. Heart 2016;102:1429-1435 https://doi.org/10.1136/heartjnl-2015-309077
  2. Kim RJ, Wu E, Rafael A, Chen EL, Parker MA, Simonetti O, et al. The use of contrast-enhanced magnetic resonance imaging to identify reversible myocardial dysfunction. N Engl J Med 2000;343:1445-1453 https://doi.org/10.1056/NEJM200011163432003
  3. Ibanez B, Aletras AH, Arai AE, Arheden H, Bax J, Berry C, et al. Cardiac MRI endpoints in myocardial infarction experimental and clinical trials: JACC scientific expert panel. J Am Coll Cardiol 2019;74:238-256 https://doi.org/10.1016/j.jacc.2019.05.024
  4. Kuruvilla S, Adenaw N, Katwal AB, Lipinski MJ, Kramer CM, Salerno M. Late gadolinium enhancement on cardiac magnetic resonance predicts adverse cardiovascular outcomes in nonischemic cardiomyopathy: a systematic review and meta-analysis. Circ Cardiovasc Imaging 2014;7:250-258 https://doi.org/10.1161/CIRCIMAGING.113.001144
  5. Haaf P, Garg P, Messroghli DR, Broadbent DA, Greenwood JP, Plein S. Cardiac T1 mapping and extracellular volume (ECV) in clinical practice: a comprehensive review. J Cardiovasc Magn Reson 2016;18:89
  6. Moon JC, Messroghli DR, Kellman P, Piechnik SK, Robson MD, Ugander M, et al. Myocardial T1 mapping and extracellular volume quantification: a Society for Cardiovascular Magnetic Resonance (SCMR) and CMR working group of the European Society of Cardiology consensus statement. J Cardiovasc Magn Reson 2013;15:92
  7. Fahmy AS, El-Rewaidy H, Nezafat M, Nakamori S, Nezafat R. Automated analysis of cardiovascular magnetic resonance myocardial native T1 mapping images using fully convolutional neural networks. J Cardiovasc Magn Reson 2019;21:7
  8. Puyol-Anton E, Ruijsink B, Baumgartner CF, Masci PG, Sinclair M, Konukoglu E, et al. Automated quantification of myocardial tissue characteristics from native T1 mapping using neural networks with uncertainty-based quality-control. J Cardiovasc Magn Reson 2020;22:60
  9. Farrag NA, Lochbihler A, White JA, Ukwatta E. Evaluation of fully automated myocardial segmentation techniques in native and contrast-enhanced T1-mapping cardiovascular magnetic resonance images using fully convolutional neural networks. Med Phys 2021;48:215-226 https://doi.org/10.1002/mp.14574
  10. Hann E, Popescu IA, Zhang Q, Gonzales RA, Barutcu A, Neubauer S, et al. Deep neural network ensemble for on-the-fly quality control-driven segmentation of cardiac MRI T1 mapping. Med Image Anal 2021;71:102029
  11. Zhu Y, Fahmy AS, Duan C, Nakamori S, Nezafat R. Automated myocardial T2 and extracellular volume quantification in cardiac MRI using transfer learning-based myocardium segmentation. Radiol Artif Intell 2020;2:e190034
  12. Suh YJ, Kim PK, Park J, Park EA, Jung JI, Choi BW. Phantom-based correction for standardization of myocardial native T1 and extracellular volume fraction in healthy subjects at 3-Tesla cardiac magnetic resonance imaging. Eur Radiol 2022 Jun 30 [Epub]. https://doi.org/10.1007/s00330-022-08936-8
  13. Diesbourg LD, Prato FS, Wisenberg G, Drost DJ, Marshall TP, Carroll SE, et al. Quantification of myocardial blood flow and extracellular volumes using a bolus injection of Gd-DTPA: kinetic modeling in canine ischemic disease. Magn Reson Med 1992;23:239-253 https://doi.org/10.1002/mrm.1910230205
  14. Park SH, Choi J, Byeon JS. Key principles of clinical validation, device approval, and insurance coverage decisions of artificial intelligence. Korean J Radiol 2021;22:442-453 https://doi.org/10.3348/kjr.2021.0048
  15. Chen C, Qin C, Qiu H, Tarroni G, Duan J, Bai W, et al. Deep learning for cardiac image segmentation: a review. Front Cardiovasc Med 2020;7:25
  16. Ngo TA, Lu Z, Carneiro G. Combining deep learning and level set for the automated segmentation of the left ventricle of the heart from cardiac cine magnetic resonance. Med Image Anal 2017;35:159-171 https://doi.org/10.1016/j.media.2016.05.009
  17. Isensee F, Jaeger PF, Full PM, Wolf I, Engelhardt S, Maier-Hein KH. Automatic cardiac disease assessment on cine-MRI via time-series segmentation and domain specific features. Proceedings of the Statistical Atlases and Computational Models of the Heart ACDC and MMWHS Challenges; 2017 Sep 10-14; Quebec, Canada: Springer International Publishing; 2018; p. 120-129
  18. Huang HH, Huang CY, Chen CN, Wang YW, Huang TY. Automatic regional analysis of myocardial native T1 values: left ventricle segmentation and AHA parcellations. Int J Cardiovasc Imaging 2018;34:131-140 https://doi.org/10.1007/s10554-017-1216-x
  19. Lee HY, Codella N, Cham M, Prince M, Weinsaft J, Wang Y. Left ventricle segmentation using graph searching on intensity and gradient and a priori knowledge (lvGIGA) for short-axis cardiac magnetic resonance imaging. J Magn Reson Imaging 2008;28:1393-1401 https://doi.org/10.1002/jmri.21586
  20. Maceira AM, Joshi J, Prasad SK, Moon JC, Perugini E, Harding I, et al. Cardiovascular magnetic resonance in cardiac amyloidosis. Circulation 2005;111:186-193 https://doi.org/10.1161/01.CIR.0000152819.97857.9D
  21. Parsai C, O'Hanlon R, Prasad SK, Mohiaddin RH. Diagnostic and prognostic value of cardiovascular magnetic resonance in non-ischaemic cardiomyopathies. J Cardiovasc Magn Reson 2012;14:54