Background: Cardiac resynchronization therapy (CRT) has emerged as an effective treatment for heart failure patients with electrical dyssynchrony. However, accurately predicting which patients will respond to CRT remains a challenge. This study explores the application of deep transfer learning techniques to train a predictive model for CRT response. Methods: In this study, the short-time Fourier transform (STFT) technique was employed to transform ECG signals into two-dimensional images. A transfer learning approach was then applied to the MIT-BIT ECG database to pre-train a convolutional neural network (CNN) model. The model was fine-tuned to extract relevant features from the ECG images and then tested on our dataset of CRT patients to predict their response. Results: Seventy-one CRT patients were enrolled in this study. The transfer learning model achieved an accuracy of 72% in distinguishing responders from non-responders in the local dataset. Furthermore, the model showed good sensitivity (0.78) and specificity (0.79) in identifying CRT responders. The performance of our model outperformed clinic guidelines and traditional machine learning approaches. Conclusion: The utilization of ECG images as input and leveraging the power of transfer learning allows for improved accuracy in identifying CRT responders. This approach offers potential for enhancing patient selection and improving the outcomes of CRT.

1. Engels EB, Mafi-Rad M, van Stipdonk AMW, et al. Why QRS duration should be replaced by better measures of electrical activation to improve patient selection for cardiac resynchronization therapy. Journal of Cardiovascular Translational Research 2016; 9(4): 257–265. doi: 10.1007/s12265-016-9693-1

2. Daubert JC, Saxon L, Adamson PB, et al. 2012 EHRA/HRS expert consensus statement on cardiac resynchronization therapy in heart failure: Implant and follow-up recommendations and management. Europace 2012; 14: 1236–1286. doi: 10.1093/europace/eus222

3. Zareba W, Klein H, Cygankiewicz I, et al. Effectiveness of cardiac resynchronization therapy by QRS morphology in the multicenter automatic defibrillator implantation trial—Cardiac resynchronization therapy (MADIT-CRT). Circulation 2011; 123(10): 1061–1072. doi: 10.1161/CIRCULATIONAHA.110.960898

4. Carità P, Corrado E, Pontone G, et al. Non-responders to cardiac resynchronization therapy: Insights from multimodality imaging and electrocardiography. A brief review. International Journal of Cardiology 2016; 225: 402–407. doi: 10.1016/j.ijcard.2016.09.037

5. Spartalis M, Tzatzaki E, Spartalis E, et al. The role of echocardiography in the optimization of cardiac resynchronization therapy: Current evidence and future perspectives. The Open Cardiovascular Medicine Journal 2017; 11: 133–145. doi: 10.2174/1874192401711010133

6. Hawkins NM, Petrie MC, MacDonald MR, et al. Selecting patients for cardiac resynchronization therapy: Electrical or mechanical dyssynchrony? European Heart Journal 2006; 27(11): 1270–1281. doi: 10.1093/eurheartj/ehi826

7. Delgado V, Bax JJ. Assessment of systolic dyssynchrony for cardiac resynchronization therapy is clinically useful. Circulation 2011; 123(6): 640–655. doi: 10.1161/CIRCULATIONAHA.110.954404

8. Zhou W, Garcia EV. Nuclear image-guided approaches for cardiac resynchronization therapy (CRT). Current Cardiology Reports 2016; 18(1): 7. doi: 10.1007/s11886-015-0687-4

9. He Z, Garcia EV, Zhou W. Nuclear image-guided methods for cardiac resynchronization therapy. In: Mesquita CT, Rezende MF (editors). Nuclear Cardiology: Basic and Advanced Concepts in Clinical Practice. Springer; 2021. pp. 587–608. doi: 10.1007/978-3-030-62195-7_25

10. Jia P, Ramanathan C, Ghanem RN, et al. Electrocardiographic imaging of cardiac resynchronization therapy in heart failure: Observation of variable electrophysiologic responses. Heart Rhythm 2006; 3(3): 296–310. doi: 10.1016/j.hrthm.2005.11.025

11. Vatasescu R, Berruezo A, Mont L, et al. Midterm ‘super-response’ to cardiac resynchronization therapy by biventricular pacing with fusion: Insights from electro-anatomical mapping. EP Europace 2009; 11(12): 1675–1682. doi: 10.1093/europace/eup333

12. Shin HC, Roth HR, Gao M, et al. Deep convolutional neural networks for computer-aided detection: CNN architectures, dataset characteristics and transfer learning. IEEE Transactions on Medical Imaging 2016; 35(5): 1285–1298. doi: 10.1109/TMI.2016.2528162

13. Cui Y, Song Y, Sun C, et al. Large scale fine-grained categorization and domain-specific transfer learning. In: Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition; 18–23 June 2018; Salt Lake City, UT, USA. pp. 4109–4118.

14. Moody GB, Mark RG. The impact of the MIT-BIH arrhythmia database. IEEE Engineering in Medicine and Biology Magazine 2001; 20(3): 45–50. doi: 10.1109/51.932724

15. Goldberger AL, Amaral LA, Glass L, et al. PhysioBank, PhysioToolkit, and PhysioNet: Components of a new research resource for complex physiologic signals. Circulation 2000; 101(23): e215–e220. doi: 10.1161/01.CIR.101.23.e215

16. Khan FZ, Virdee MS, Palmer CR, et al. Targeted left ventricular lead placement to guide cardiac resynchronization therapy: The TARGET study: A randomized, controlled trial. Journal of the American College of Cardiology 2012; 59(17): 1509–1518. doi: 10.1016/j.jacc.2011.12.030

17. Beela AS, Ünlü S, Duchenne J, et al. Assessment of mechanical dyssynchrony can improve the prognostic value of guideline-based patient selection for cardiac resynchronization therapy. European Heart Journal—Cardiovascular Imaging 2019; 20(1): 66–74. doi: 10.1093/ehjci/jey029

18. Chung ES, Leon AR, Tavazzi L, et al. Results of the predictors of response to CRT (PROSPECT) trial. Circulation 2008; 117(20): 2608–2616. doi: 10.1161/CIRCULATIONAHA.107.743120

19. Engelse WA, Zeelenberg C. A single scan algorithm for QRS-detection and feature extraction. IEEE Computers in Cardiology 1979; 6(1979): 37–42.

20. Elgendi M, Eskofier B, Dokos S, Abbott D. Revisiting QRS detection methodologies for portable, wearable, battery-operated, and wireless ECG systems. PLoS One 2014; 9(1): e84018. doi: 10.1371/journal.pone.0084018

21. Allen J. Short term spectral analysis, synthesis, and modification by discrete Fourier transform. IEEE Transactions on Acoustics, Speech, and Signal Processing 1977; 25(3): 235–238. doi: 10.1109/TASSP.1977.1162950

22. Smith III JO. Spectral Audio Signal Processing. W3K Publishing; 2011. 674p.

23. Haykin S, Veen B. Zobrist B (editor). Signals and Systems. John Willey & Sons; 1999.

24. Huang J, Chen B, Yao B, He W. ECG arrhythmia classification using STFT-based spectrogram and convolutional neural network. IEEE Access 2019; 7: 92871–92880. doi: 10.1109/ACCESS.2019.2928017

25. Virtanen P, Gommers R, Oliphant TE, et al. SciPy 1.0: Fundamental algorithms for scientific computing in Python. Nature Methods 2020; 17(3): 261–272. doi: 10.1038/s41592-019-0686-2

26. He K, Zhang X, Ren S, Sun J. Deep residual learning for image recognition. In: Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR); 27–30 June 2016; Las Vegas, NV, USA. pp. 770–778. doi: 10.1109/CVPR.2016.90

27. Tracy CM, Epstein AE, Darbar D, et al. 2012 ACCF/AHA/HRS focused update of the 2008 guidelines for device-based therapy of cardiac rhythm abnormalities: A report of the American College of Cardiology Foundation/American Heart Association task force on practice guidelines. Journal of the American College of Cardiology 2012; 60(14): 1297–1313. doi: 10.1016/j.jacc.2012.07.009

28. Glikson M, Nielsen JC, Kronborg MB, et al. 2021 ESC Guidelines on cardiac pacing and cardiac resynchronization therapy. European Heart Journal 2021; 42(35): 3427–3520. doi: 10.1093/eurheartj/ehab364

29. Cleland JG, Daubert JC, Erdmann E, et al. The effect of cardiac resynchronization on morbidity and mortality in heart failure. New England Journal of Medicine 2005; 352(15): 1539–1549. doi: 10.1056/NEJMoa050496

30. Abraham WT, Fisher WG, Smith AL, et al. Cardiac resynchronization in chronic heart failure. New England Journal of Medicine 2002; 346(24): 1845–1853. doi: 10.1056/NEJMoa013168

31. Linde C, Abraham WT, Gold MR, et al. Randomized trial of cardiac resynchronization in mildly symptomatic heart failure patients and in asymptomatic patients with left ventricular dysfunction and previous heart failure symptoms. Journal of the American College of Cardiology 2008; 52(23): 1834–1843. doi: 10.1016/j.jacc.2008.08.027

32. Healey JS, Hohnloser SH, Exner DV, et al. Cardiac resynchronization therapy in patients with permanent atrial fibrillation: Results from the Resynchronization for Ambulatory Heart Failure Trial (RAFT). Circulation: Heart Failure 2012; 5(5): 566–570. doi: 10.1161/CIRCHEARTFAILURE.112.968867

33. Ruschitzka F, Abraham WT, Singh JP, et al. Cardiac-resynchronization therapy in heart failure with a narrow QRS complex. New England Journal of Medicine 2013; 369(15): 1395–1405. doi: 10.1056/NEJMoa1306687

34. Delgado V, Ypenburg C, van Bommel RJ, et al. Assessment of left ventricular dyssynchrony by speckle tracking strain imaging: Comparison between longitudinal, circumferential, and radial strain in cardiac resynchronization therapy. Journal of the American College of Cardiology 2008; 51(20): 1944–1952. doi: 10.1016/j.jacc.2008.02.040

35. Brignole M, Pentimalli F, Palmisano P, et al. AV junction ablation and cardiac resynchronization for patients with permanent atrial fibrillation and narrow QRS: The APAF-CRT mortality trial. European Heart Journal 2021; 42(46): 4731–4739. doi: 10.1093/eurheartj/ehab569

36. Beshai JF, Grimm RA, Nagueh SF, et al. Cardiac-resynchronization therapy in heart failure with narrow QRS complexes. New England Journal of Medicine 2007; 357(24): 2461–2471. doi: 10.1056/NEJMoa0706695

37. Stockburger M, Moss AJ, Klein HU, et al. Sustained clinical benefit of cardiac resynchronization therapy in non-LBBB patients with prolonged PR-interval: MADIT-CRT long-term follow-up. Clinical Research in Cardiology 2016; 105(11): 944–952. doi: 10.1007/s00392-016-1003-z

38. Poole JE, Singh JP, Birgersdotter-Green U. QRS duration or QRS morphology: What really matters in cardiac resynchronization therapy? Journal of the American College of Cardiology 2016; 67(9): 1104–1117. doi: 10.1016/j.jacc.2015.12.039

39. Attia ZI, Kapa S, Lopez-Jimenez F, et al. Screening for cardiac contractile dysfunction using an artificial intelligence-enabled electrocardiogram. Nature Medicine 2019; 25(1): 70–74. doi: 10.1038/s41591-018-0240-2

40. Hua X, Han J, Zhao C, et al. A novel method for ECG signal classification via one-dimensional convolutional neural network. Multimedia Systems 2020; 28: 1387–1399. doi: 10.1007/s00530-020-00713-1

41. Baeßler B, Mannil M, Maintz D, et al. Texture analysis and machine learning of non-contrast T1-weighted MR images in patients with hypertrophic cardiomyopathy—Preliminary results. European Journal of Radiology 2018; 102: 61–67. doi: 10.1016/j.ejrad.2018.03.013

42. van Steenkiste G, van Loon G, Crevecoeur G. Transfer learning in ECG classification from human to horse using a novel parallel neural network architecture. Scientific Reports 2020; 10(1): 186. doi: 10.1038/s41598-019-57025-2

43. Weimann K, Conrad TO. Transfer learning for ECG classification. Scientific Reports 2021; 11(1): 5251. doi: 10.1038/s41598-021-84374-8

44. Naz M, Shah JH, Khan MA, et al. From ECG signals to images: A transformation based approach for deep learning. PeerJ Computer Science 2021; 7: e386. doi: 10.7717/peerj-cs.386

45. Abdeldayem SS, Bourlai T. ECG-based human authentication using high-level spectro-temporal signal features. In: Proceedings of the 2018 IEEE International Conference on Big Data (Big Data); 10–13 December 2018; Seattle, WA, USA. pp. 4984–4993. doi: 10.1109/BigData.2018.8622619

46. Gupta V, Mittal M. QRS complex detection using STFT, chaos analysis, and PCA in standard and real-time ECG databases. Journal of The Institution of Engineers (India): Series B 2019; 100(5): 489–497. doi: 10.1007/s40031-019-00398-9

47. Hung GU, Zou J, He Z, et al. Left-ventricular dyssynchrony in viable myocardium by myocardial perfusion SPECT is predictive of mechanical response to CRT. Annals of Nuclear Medicine 2021; 35(8): 947–954. doi: 10.1007/s12149-021-01632-5

48. He Z, Li D, Cui C, et al. Predictive values of left ventricular mechanical dyssynchrony for CRT response in heart failure patients with different pathophysiology. Journal of Nuclear Cardiology 2021; 29(5): 2637–2648. doi: 10.1007/s12350-021-02796-3

49. He Z, Zhang X, Zhao C, et al. A method using deep learning to discover new predictors from left-ventricular mechanical dyssynchrony for CRT response. Journal of Nuclear Cardiology 2022; 30(1): 201–213. doi: 10.1007/s12350-022-03067-5

50. Fudim M, Borges-Neto S. A troubled marriage: When electrical and mechanical dyssynchrony don’t go along. Journal of Nuclear Cardiology 2019; 26(4): 1240–1242. doi: 10.1007/s12350-018-1227-6

51. Zhou Y, He Z, Liao S, et al. Prognostic value of integrative analysis of electrical and mechanical dyssynchrony in patients with acute heart failure. Journal of Nuclear Cardiology 2021; 28(1): 140–149. doi: 10.1007/s12350-020-02429-1