Toward Better Risk Stratification for Implantable Cardioverter-Defibrillator Recipients: Implications of Explainable Machine Learning Models
Abstract
:1. Introduction
2. Materials and Methods
2.1. Patient Population
2.2. Outcome Definition
2.3. Data Collection and Preprocessing
2.4. Modeling Strategies
2.5. Model Interpretability
2.6. Statistical Analysis
3. Results
3.1. Patient Characteristics
3.2. Outcome Events
3.3. Prediction of All-Cause Death and Appropriate Shock
3.4. Explainability Based on SHAP Values
3.5. Establishment of Bi-Dimensional Risk Profiles
3.6. Sensitivity Analysis
4. Discussion
5. Conclusions
Supplementary Materials
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Conflicts of Interest
References
- Kiguchi, T.; Okubo, M.; Nishiyama, C.; Maconochie, I.; Ong, M.E.H.; Kern, K.B.; Wyckoff, M.H.; McNally, B.; Christensen, E.F.; Tjelmeland, I.; et al. Out-of-hospital cardiac arrest across the World: First report from the International Liaison Committee on Resuscitation (ILCOR). Resuscitation 2020, 152, 39–49. [Google Scholar] [CrossRef] [PubMed]
- Tsao, C.W.; Aday, A.W.; Almarzooq, Z.I.; Alonso, A.; Beaton, A.Z.; Bittencourt, M.S.; Boehme, A.K.; Buxton, A.E.; Carson, A.P.; Commodore-Mensah, Y.; et al. Heart Disease and Stroke Statistics-2022 Update: A Report From the American Heart Association. Circulation 2022, 145, e153–e639. [Google Scholar] [CrossRef] [PubMed]
- Al-Khatib, S.M.; Stevenson, W.G.; Ackerman, M.J.; Bryant, W.J.; Callans, D.J.; Curtis, A.B.; Deal, B.J.; Dickfeld, T.; Field, M.E.; Fonarow, G.C. 2017 AHA/ACC/HRS guideline for management of patients with ventricular arrhythmias and the prevention of sudden cardiac death: A report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Rhythm Society. J. Am. Coll. Cardiol. 2018, 72, e91–e220. [Google Scholar] [CrossRef]
- Merchant, F.M.; Jones, P.; Wehrenberg, S.; Lloyd, M.S.; Saxon, L.A. Incidence of defibrillator shocks after elective generator exchange following uneventful first battery life. J. Am. Heart Assoc. 2014, 3, e001289. [Google Scholar] [CrossRef] [PubMed]
- Zabel, M.; Willems, R.; Lubinski, A.; Bauer, A.; Brugada, J.; Conen, D.; Flevari, P.; Hasenfuss, G.; Svetlosak, M.; Huikuri, H.V.; et al. Clinical effectiveness of primary prevention implantable cardioverter-defibrillators: Results of the EU-CERT-ICD controlled multicentre cohort study. Eur. Heart J. 2020, 41, 3437–3447. [Google Scholar] [CrossRef] [PubMed]
- Briongos-Figuero, S.; Garcia-Alberola, A.; Rubio, J.; Segura, J.M.; Rodriguez, A.; Peinado, R.; Alzueta, J.; Martinez-Ferrer, J.B.; Vinolas, X.; Fernandez de la Concha, J.; et al. Long-Term Outcomes Among a Nationwide Cohort of Patients Using an Implantable Cardioverter-Defibrillator: UMBRELLA Study Final Results. J. Am. Heart Assoc. 2021, 10, e018108. [Google Scholar] [CrossRef] [PubMed]
- Nielsen, J.C.; Lin, Y.-J.; de Oliveira Figueiredo, M.J.; Sepehri Shamloo, A.; Alfie, A.; Boveda, S.; Dagres, N.; Di Toro, D.; Eckhardt, L.L.; Ellenbogen, K. European Heart Rhythm Association (EHRA)/Heart Rhythm Society (HRS)/Asia Pacific Heart Rhythm Society (APHRS)/Latin American Heart Rhythm Society (LAHRS) expert consensus on risk assessment in cardiac arrhythmias: Use the right tool for the right outcome, in the right population. EP Europace 2020, 22, 1147–1148. [Google Scholar] [CrossRef]
- Levy, W.C.; Lee, K.L.; Hellkamp, A.S.; Poole, J.E.; Mozaffarian, D.; Linker, D.T.; Maggioni, A.P.; Anand, I.; Poole-Wilson, P.A.; Fishbein, D.P.; et al. Maximizing survival benefit with primary prevention implantable cardioverter-defibrillator therapy in a heart failure population. Circulation 2009, 120, 835–842. [Google Scholar] [CrossRef]
- Shadman, R.; Poole, J.E.; Dardas, T.F.; Mozaffarian, D.; Cleland, J.G.; Swedberg, K.; Maggioni, A.P.; Anand, I.S.; Carson, P.E.; Miller, A.B.; et al. A novel method to predict the proportional risk of sudden cardiac death in heart failure: Derivation of the Seattle Proportional Risk Model. Heart Rhythm 2015, 12, 2069–2077. [Google Scholar] [CrossRef]
- Levy, W.C.; Li, Y.; Reed, S.D.; Zile, M.R.; Shadman, R.; Dardas, T.; Whellan, D.J.; Schulman, K.A.; Ellis, S.J.; Neilson, M.; et al. Does the Implantable Cardioverter-Defibrillator Benefit Vary With the Estimated Proportional Risk of Sudden Death in Heart Failure Patients? JACC Clin. Electrophysiol. 2017, 3, 291–298. [Google Scholar] [CrossRef]
- Lee, D.S.; Hardy, J.; Yee, R.; Healey, J.S.; Birnie, D.; Simpson, C.S.; Crystal, E.; Mangat, I.; Nanthakumar, K.; Wang, X.; et al. Clinical Risk Stratification for Primary Prevention Implantable Cardioverter Defibrillators. Circ. Heart Fail. 2015, 8, 927–937. [Google Scholar] [CrossRef] [PubMed]
- Bilchick, K.C.; Wang, Y.; Cheng, A.; Curtis, J.P.; Dharmarajan, K.; Stukenborg, G.J.; Shadman, R.; Anand, I.; Lund, L.H.; Dahlstrom, U.; et al. Seattle Heart Failure and Proportional Risk Models Predict Benefit From Implantable Cardioverter-Defibrillators. J. Am. Coll. Cardiol. 2017, 69, 2606–2618. [Google Scholar] [CrossRef] [PubMed]
- Bergau, L.; Willems, R.; Sprenkeler, D.J.; Fischer, T.H.; Flevari, P.; Hasenfuss, G.; Katsaras, D.; Kirova, A.; Lehnart, S.E.; Luthje, L.; et al. Differential multivariable risk prediction of appropriate shock versus competing mortality-A prospective cohort study to estimate benefits from ICD therapy. Int. J. Cardiol. 2018, 272, 102–107. [Google Scholar] [CrossRef]
- Kristensen, S.L.; Levy, W.C.; Shadman, R.; Nielsen, J.C.; Haarbo, J.; Videbaek, L.; Bruun, N.E.; Eiskjaer, H.; Wiggers, H.; Brandes, A.; et al. Risk Models for Prediction of Implantable Cardioverter-Defibrillator Benefit: Insights From the DANISH Trial. JACC Heart Fail. 2019, 7, 717–724. [Google Scholar] [CrossRef]
- Younis, A.; Goldberger, J.J.; Kutyifa, V.; Zareba, W.; Polonsky, B.; Klein, H.; Aktas, M.K.; Huang, D.; Daubert, J.; Estes, M.; et al. Predicted benefit of an implantable cardioverter-defibrillator: The MADIT-ICD benefit score. Eur. Heart J. 2021, 42, 1676–1684. [Google Scholar] [CrossRef] [PubMed]
- Deng, Y.; Zhang, N.; Hua, W.; Cheng, S.; Niu, H.; Chen, X.; Gu, M.; Cai, C.; Liu, X.; Huang, H.; et al. Nomogram predicting death and heart transplantation before appropriate ICD shock in dilated cardiomyopathy. ESC Heart Fail. 2022, 9, 1269–1278. [Google Scholar] [CrossRef] [PubMed]
- Reeder, H.T.; Shen, C.; Buxton, A.E.; Haneuse, S.J.; Kramer, D.B. Joint Shock/Death Risk Prediction Model for Patients Considering Implantable Cardioverter-Defibrillators. Circ. Cardiovasc. Qual. Outcomes 2019, 12, e005675. [Google Scholar] [CrossRef]
- Feeny, A.K.; Chung, M.K.; Madabhushi, A.; Attia, Z.I.; Cikes, M.; Firouznia, M.; Friedman, P.A.; Kalscheur, M.M.; Kapa, S.; Narayan, S.M.; et al. Artificial Intelligence and Machine Learning in Arrhythmias and Cardiac Electrophysiology. Circ. Arrhythm. Electrophysiol. 2020, 13, e007952. [Google Scholar] [CrossRef]
- Lundberg, S.M.; Lee, S.-I. A unified approach to interpreting model predictions. Adv. Neural Inf. Processing Syst. 2017, 30, 4768–4777. [Google Scholar]
- Pölsterl, S. scikit-survival: A Library for Time-to-Event Analysis Built on Top of scikit-learn. J. Mach. Learn. Res. 2020, 21, 1–6. [Google Scholar]
- Chen, T.; Guestrin, C. Xgboost: A scalable tree boosting system. In Proceedings of the 22nd Acm Sigkdd International Conference on Knowledge Discovery and Data Mining, San Francisco, CA, USA, 13–17 August 2016; pp. 785–794. [Google Scholar]
- Stolfo, D.; Ceschia, N.; Zecchin, M.; De Luca, A.; Gobbo, M.; Barbati, G.; Gigli, M.; Mase, M.; Pinamonti, B.; Pivetta, A.; et al. Arrhythmic Risk Stratification in Patients With Idiopathic Dilated Cardiomyopathy. Am. J. Cardiol. 2018, 121, 1601–1609. [Google Scholar] [CrossRef] [PubMed]
- Di Marco, A.; Brown, P.F.; Bradley, J.; Nucifora, G.; Claver, E.; de Frutos, F.; Dallaglio, P.D.; Comin-Colet, J.; Anguera, I.; Miller, C.A.; et al. Improved Risk Stratification for Ventricular Arrhythmias and Sudden Death in Patients With Nonischemic Dilated Cardiomyopathy. J. Am. Coll. Cardiol. 2021, 77, 2890–2905. [Google Scholar] [CrossRef] [PubMed]
- Ambale-Venkatesh, B.; Yang, X.; Wu, C.O.; Liu, K.; Hundley, W.G.; McClelland, R.; Gomes, A.S.; Folsom, A.R.; Shea, S.; Guallar, E.; et al. Cardiovascular Event Prediction by Machine Learning: The Multi-Ethnic Study of Atherosclerosis. Circ. Res. 2017, 121, 1092–1101. [Google Scholar] [CrossRef]
- Levy, W.C.; Mozaffarian, D.; Linker, D.T.; Sutradhar, S.C.; Anker, S.D.; Cropp, A.B.; Anand, I.; Maggioni, A.; Burton, P.; Sullivan, M.D.; et al. The Seattle Heart Failure Model: Prediction of survival in heart failure. Circulation 2006, 113, 1424–1433. [Google Scholar] [CrossRef] [PubMed]
- Deng, Y.; Cheng, S.J.; Hua, W.; Cai, M.S.; Zhang, N.X.; Niu, H.X.; Chen, X.H.; Gu, M.; Cai, C.; Liu, X.; et al. N-Terminal Pro-B-Type Natriuretic Peptide in Risk Stratification of Heart Failure Patients With Implantable Cardioverter-Defibrillator. Front. Cardiovasc. Med. 2022, 9, 823076. [Google Scholar] [CrossRef]
- Gulati, A.; Jabbour, A.; Ismail, T.F.; Guha, K.; Khwaja, J.; Raza, S.; Morarji, K.; Brown, T.D.; Ismail, N.A.; Dweck, M.R.; et al. Association of fibrosis with mortality and sudden cardiac death in patients with nonischemic dilated cardiomyopathy. JAMA 2013, 309, 896–908. [Google Scholar] [CrossRef] [PubMed]
- Cheng, A.; Zhang, Y.; Blasco-Colmenares, E.; Dalal, D.; Butcher, B.; Norgard, S.; Eldadah, Z.; Ellenbogen, K.A.; Dickfeld, T.; Spragg, D.D.; et al. Protein biomarkers identify patients unlikely to benefit from primary prevention implantable cardioverter defibrillators: Findings from the Prospective Observational Study of Implantable Cardioverter Defibrillators (PROSE-ICD). Circ. Arrhythm. Electrophysiol. 2014, 7, 1084–1091. [Google Scholar] [CrossRef] [PubMed]
- Zegard, A.; Okafor, O.; De Bono, J.; Kalla, M.; Lencioni, M.; Marshall, H.; Hudsmith, L.; Qiu, T.; Steeds, R.; Stegemann, B. Myocardial fibrosis as a predictor of sudden death in patients with coronary artery disease. J. Am. Coll. Cardiol. 2021, 77, 29–41. [Google Scholar] [CrossRef]
- Packer, M. What causes sudden death in patients with chronic heart failure and a reduced ejection fraction? Eur. Heart J. 2020, 41, 1757–1763. [Google Scholar] [CrossRef]
- Rohde, L.E.; Vaduganathan, M.; Claggett, B.L.; Polanczyk, C.A.; Dorbala, P.; Packer, M.; Desai, A.S.; Zile, M.; Rouleau, J.; Swedberg, K.; et al. Dynamic changes in cardiovascular and systemic parameters prior to sudden cardiac death in heart failure with reduced ejection fraction: A PARADIGM-HF analysis. Eur. J. Heart Fail. 2021, 23, 1346–1356. [Google Scholar] [CrossRef]
- Wu, K.C.; Wongvibulsin, S.; Tao, S.; Ashikaga, H.; Stillabower, M.; Dickfeld, T.M.; Marine, J.E.; Weiss, R.G.; Tomaselli, G.F.; Zeger, S.L. Baseline and Dynamic Risk Predictors of Appropriate Implantable Cardioverter Defibrillator Therapy. J. Am. Heart Assoc. 2020, 9, e017002. [Google Scholar] [CrossRef] [PubMed]
- Halliday, B.P.; Cleland, J.G.F.; Goldberger, J.J.; Prasad, S.K. Personalizing Risk Stratification for Sudden Death in Dilated Cardiomyopathy: The Past, Present, and Future. Circulation 2017, 136, 215–231. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- Hammersley, D.J.; Halliday, B.P. Sudden Cardiac Death Prediction in Non-ischemic Dilated Cardiomyopathy: A Multiparametric and Dynamic Approach. Curr. Cardiol. Rep. 2020, 22, 85. [Google Scholar] [CrossRef] [PubMed]
- Borleffs, C.J.; van Erven, L.; Schotman, M.; Boersma, E.; Kies, P.; van der Burg, A.E.; Zeppenfeld, K.; Bootsma, M.; van der Wall, E.E.; Bax, J.J.; et al. Recurrence of ventricular arrhythmias in ischaemic secondary prevention implantable cardioverter defibrillator recipients: Long-term follow-up of the Leiden out-of-hospital cardiac arrest study (LOHCAT). Eur. Heart J. 2009, 30, 1621–1626. [Google Scholar] [CrossRef] [PubMed]
- Schaer, B.; Kuhne, M.; Reichlin, T.; Osswald, S.; Sticherling, C. Incidence of and predictors for appropriate implantable cardioverter-defibrillator therapy in patients with a secondary preventive implantable cardioverter-defibrillator indication. Europace 2016, 18, 227–231. [Google Scholar] [CrossRef]
- Klem, I.; Klein, M.; Khan, M.; Yang, E.Y.; Nabi, F.; Ivanov, A.; Bhatti, L.; Hayes, B.; Graviss, E.A.; Nguyen, D.T.; et al. Relationship of LVEF and Myocardial Scar to Long-Term Mortality Risk and Mode of Death in Patients With Nonischemic Cardiomyopathy. Circulation 2021, 143, 1343–1358. [Google Scholar] [CrossRef]
- Buxton, A.E.; Lee, K.L.; Hafley, G.E.; Pires, L.A.; Fisher, J.D.; Gold, M.R.; Josephson, M.E.; Lehmann, M.H.; Prystowsky, E.N.; Investigators, M. Limitations of ejection fraction for prediction of sudden death risk in patients with coronary artery disease: Lessons from the MUSTT study. J. Am. Coll. Cardiol. 2007, 50, 1150–1157. [Google Scholar] [CrossRef]
- Oscar, O.; Enrique, R.; Andres, B. Subanalyses of secondary prevention implantable cardioverter-defibrillator trials: Antiarrhythmics versus implantable defibrillators (AVID), Canadian Implantable Defibrillator Study (CIDS), and Cardiac Arrest Study Hamburg (CASH). Curr. Opin. Cardiol. 2004, 19, 26–30. [Google Scholar] [CrossRef]
- Moss, A.J.; Schuger, C.; Beck, C.A.; Brown, M.W.; Cannom, D.S.; Daubert, J.P.; Estes, N.A., III; Greenberg, H.; Hall, W.J.; Huang, D.T.; et al. Reduction in inappropriate therapy and mortality through ICD programming. N. Engl. J. Med. 2012, 367, 2275–2283. [Google Scholar] [CrossRef]
- Aktas, M.K.; Bennett, A.L.; Younis, A.; Kutyifa, V.; Polonsky, B.; McNitt, S.; Zareba, W.; Rosero, S.; Goldenberg, I. Implantable cardioverter-defibrillator programming after first occurrence of ventricular tachycardia in the Multicenter Automatic Defibrillator Implantation Trial-Reduce Inappropriate Therapy (MADIT-RIT). Heart Rhythm O2 2020, 1, 77–82. [Google Scholar] [CrossRef]
Characteristics | Datasets of All-Cause Death | p-Value | Datasets of First Appropriate Shock | p-Value | |||
---|---|---|---|---|---|---|---|
All Patients (n = 887) | Training Set (n = 665) | Test Set (n = 222) | Training Set (n = 665) | Test Set (n = 222) | |||
Demographics | |||||||
Age (years) | 59.0 ± 13.0 | 59.3 ± 12.8 | 58.3 ± 13.7 | 0.361 | 59.0 ± 13.1 | 59.1 ± 13.0 | 0.894 |
Male sex | 667 (75.2%) | 504 (75.8%) | 163 (73.4%) | 0.537 | 498 (74.9%) | 169 (76.1%) | 0.779 |
Body mass index (kg/m2) | 24.7 ± 3.6 | 24.8 ± 3.5 | 24.5 ± 3.8 | 0.284 | 24.8 ± 3.7 | 24.6 ± 3.3 | 0.631 |
Ischemic etiology | 433 (48.8%) | 324 (48.7%) | 109 (49.1%) | 0.984 | 317 (47.7%) | 116 (52.3%) | 0.269 |
Family history of sudden death | 25 (2.8%) | 20 (3.0%) | 5 (2.3%) | 0.723 | 16 (2.4%) | 9 (4.1%) | 0.293 |
Clinical characteristics | |||||||
Smoking | 416 (46.9%) | 316 (47.5%) | 100 (45.0%) | 0.574 | 314 (47.2%) | 102 (45.9%) | 0.802 |
Primary prevention | 240 (27.1%) | 185 (27.8%) | 55 (24.8%) | 0.425 | 179 (26.9%) | 61 (27.5%) | 0.94 |
Dual-chamber ICD | 303 (34.2%) | 240 (36.1%) | 63 (28.4%) | 0.044 | 230 (34.6%) | 73 (32.9%) | 0.703 |
Systolic BP (mmHg) | 120.5 ± 16.6 | 120.9 ± 16.4 | 119.5 ± 17.2 | 0.298 | 120.8 ± 16.9 | 119.7 ± 15.7 | 0.377 |
Diastolic BP (mmHg) | 73.5 ± 10.3 | 73.8 ± 10.1 | 72.9 ± 10.9 | 0.292 | 73.5 ± 10.5 | 73.7 ± 9.8 | 0.736 |
NYHA class | 0.396 | 0.498 | |||||
I | 239 (26.9%) | 176 (26.5%) | 63 (28.4%) | 184 (27.7%) | 55 (24.8%) | ||
II | 326 (36.8%) | 255 (38.3%) | 71 (32.0%) | 249 (37.4%) | 77 (34.7%) | ||
III | 260 (29.3%) | 189 (28.4%) | 71 (32.0%) | 188 (28.3%) | 72 (32.4%) | ||
IV | 62 (7.0%) | 45 (6.8%) | 17 (7.7%) | 44 (6.6%) | 18 (8.1%) | ||
Echocardiogram | |||||||
LVEDD (mm) | 60.3 ± 10.9 | 60.4 ± 10.8 | 59.9 ± 11.2 | 0.606 | 60.0 ± 11.0 | 61.2 ± 10.5 | 0.158 |
LVEF (%) | 43.1 ± 14.6 | 43.2 ± 14.3 | 43.0 ± 15.3 | 0.897 | 43.4 ± 14.7 | 42.2 ± 14.1 | 0.282 |
LAD (mm) | 42.5 ± 8.1 | 42.5 ± 7.9 | 42.5 ± 8.8 | 0.941 | 42.3 ± 8.0 | 43.0 ± 8.4 | 0.306 |
IVS (mm) | 9.4 ± 2.3 | 9.4 ± 2.2 | 9.5 ± 2.7 | 0.555 | 9.4 ± 2.4 | 9.4 ± 2.1 | 0.747 |
RVD (mm) | 22.6 ± 4.3 | 22.5 ± 4.2 | 22.9 ± 4.5 | 0.281 | 22.7 ± 4.5 | 22.6 ± 3.7 | 0.780 |
Tricuspid valve regurgitation | 84 (9.5%) | 61 (9.2%) | 23 (10.4%) | 0.696 | 67 (10.1%) | 17 (7.7%) | 0.351 |
Mitral valve regurgitation | 169 (19.1%) | 121 (18.2%) | 48 (21.6%) | 0.305 | 127 (19.1%) | 42 (18.9%) | 1.000 |
Electrocardiogram findings | |||||||
Heart rate (beats per minute) | 69.0 ± 13.6 | 68.6 ± 13.7 | 70.1 ± 13.4 | 0.163 | 68.6 ± 13.8 | 70.1 ± 13.3 | 0.167 |
CLBBB | 48 (5.4%) | 32 (4.8%) | 16 (7.2%) | 0.232 | 38 (5.7%) | 10 (4.5%) | 0.604 |
CRBBB | 53 (6.0%) | 42 (6.3%) | 11 (5.0%) | 0.564 | 39 (5.9%) | 14 (6.3%) | 0.939 |
Frequent PVCs | 371 (41.8%) | 283 (42.6%) | 88 (39.6%) | 0.494 | 281 (42.3%) | 90 (40.5%) | 0.711 |
Pacing indication | 59 (6.7%) | 45 (6.8%) | 14 (6.3%) | 0.934 | 44 (6.6%) | 15 (6.8%) | 1.000 |
Comorbidities | |||||||
Myocardial infarction | 345 (38.9%) | 266 (40.0%) | 79 (35.6%) | 0.276 | 256 (38.5%) | 89 (40.1%) | 0.732 |
Atrial fibrillation | 259 (29.2%) | 190 (28.6%) | 69 (31.1%) | 0.531 | 189 (28.4%) | 70 (31.5%) | 0.425 |
Hypertension | 383 (43.2%) | 291 (43.8%) | 92 (41.4%) | 0.599 | 285 (42.9%) | 98 (44.1%) | 0.797 |
Diabetes | 179 (20.2%) | 134 (20.2%) | 45 (20.3%) | 1.000 | 134 (20.2%) | 45 (20.3%) | 1.000 |
Hyperlipidemia | 431 (48.6%) | 324 (48.7%) | 107 (48.2%) | 0.954 | 322 (48.4%) | 109 (49.1%) | 0.922 |
Stroke | 58 (6.5%) | 42 (6.3%) | 16 (7.2%) | 0.758 | 48 (7.2%) | 10 (4.5%) | 0.208 |
Hyperuricemia | 78 (8.8%) | 64 (9.6%) | 14 (6.3%) | 0.169 | 59 (8.9%) | 19 (8.6%) | 0.995 |
Laboratory tests | |||||||
NT-proBNP (pg/mL) | 788.9 (302.0,1779.0) | 765.8 (299.2,1761.8) | 853.3 (330.8,1794.2) | 0.479 | 743.6 (299.5,1714.0) | 874.8 (316.8,1904.3) | 0.268 |
Hemoglobin (g/L) | 140.3 ± 18.1 | 140.1 ± 17.9 | 141.0 ± 19.0 | 0.530 | 140.8 ± 18.3 | 138.9 ± 17.8 | 0.174 |
Creatinine (μmol/L) | 88.0 (75.2,103.7) | 87.7 (75.3,104.0) | 88.0 (75.0,102.6) | 0.955 | 87.7 (75.0,104.0) | 88.3 (75.7,102.9) | 0.982 |
BUN (mmol/L) | 6.6 (5.3,8.6) | 6.7 (5.4,8.7) | 6.0 (4.9,8.3) | 0.015 | 6.6 (5.3,8.6) | 6.5 (4.9,8.7) | 0.428 |
hs-CRP (mg/L) | 1.9 (0.8,4.6) | 2.0 (0.8,4.7) | 1.9 (0.8,4.2) | 0.962 | 2.0 (0.8,4.3) | 1.7 (0.8,5.6) | 0.856 |
Medications | |||||||
ACEI/ARB/ ARNI | 573 (64.6%) | 426 (64.1%) | 147 (66.2%) | 0.617 | 433 (65.1%) | 140 (63.1%) | 0.637 |
Amiodarone | 461 (52.0%) | 354 (53.2%) | 107 (48.2%) | 0.222 | 344 (51.7%) | 117 (52.7%) | 0.862 |
Beta-blockers | 747 (84.2%) | 567 (85.3%) | 180 (81.1%) | 0.170 | 564 (84.8%) | 183 (82.4%) | 0.462 |
Calcium channel blockers | 94 (10.6%) | 74 (11.1%) | 20 (9.0%) | 0.446 | 72 (10.8%) | 22 (9.9%) | 0.796 |
Diuretics | 564 (63.6%) | 428 (64.4%) | 136 (61.3%) | 0.453 | 422 (63.5%) | 142 (64.0%) | 0.956 |
MRA | 524 (59.1%) | 396 (59.5%) | 128 (57.7%) | 0.676 | 394 (59.2%) | 130 (58.6%) | 0.919 |
Digitalis | 196 (22.1%) | 143 (21.5%) | 53 (23.9%) | 0.520 | 150 (22.6%) | 46 (20.7%) | 0.633 |
Statin | 449 (50.6%) | 330 (49.6%) | 119 (53.6%) | 0.342 | 332 (49.9%) | 117 (52.7%) | 0.523 |
Antiplatelet | 322 (36.3%) | 248 (37.3%) | 74 (33.3%) | 0.326 | 239 (35.9%) | 83 (37.4%) | 0.758 |
Anticoagulants | 163 (18.4%) | 115 (17.3%) | 48 (21.6%) | 0.180 | 120 (18.0%) | 43 (19.4%) | 0.733 |
Algorithms | Parameter | Search Space | Optimal Parameter for Death Prediction | Optimal Parameter for Shock Prediction |
---|---|---|---|---|
EN-Cox | l1 ratio | 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0 | 0.9 | 0.2 |
alpha | Log distribution from 0.0001 to 1 | 0.0233 | 0.0339 | |
RSF | number of trees | 100, 200, 300, 400, 500 | 400 | 500 |
maximum depth | 2, 3, 4, 5, 6, 7 | 4 | 7 | |
minimum samples required to split | 10, 14, 28, 22, 40, 50 | 22 | 40 | |
minimum samples required at leaf nodes | 5, 7, 9, 11, 20, 25 | 5 | 9 | |
SSVM | alpha | 0.1, 1, 10, 100 | 0.1 | 0.1 |
gamma | 1, 0.1, 0.01, 0.001 | 1 | 0.001 | |
kernel | rbf, poly, linear, sigmoid, cosine | poly | rbf | |
degree (poly kernels only) | 2, 3, 4, 5 | 4 | - | |
XGBoost | loss function | CoxPH | - | - |
learning rate | 0.01, 0.05, 0.10 | 0.1 | 0.1 | |
number of trees | 20, 25, 30 | 30 | 30 | |
maximum depth | 1, 2 | 2 | 2 | |
fraction of samples | 0.4, 0.5 | 0.4 | 0.4 | |
fraction of variables | 0.4, 0.5 | 0.5 | 0.4 | |
minimum samples required to split | 1, 2 | 1 | 1 |
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
© 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Deng, Y.; Cheng, S.; Huang, H.; Liu, X.; Yu, Y.; Gu, M.; Cai, C.; Chen, X.; Niu, H.; Hua, W. Toward Better Risk Stratification for Implantable Cardioverter-Defibrillator Recipients: Implications of Explainable Machine Learning Models. J. Cardiovasc. Dev. Dis. 2022, 9, 310. https://doi.org/10.3390/jcdd9090310
Deng Y, Cheng S, Huang H, Liu X, Yu Y, Gu M, Cai C, Chen X, Niu H, Hua W. Toward Better Risk Stratification for Implantable Cardioverter-Defibrillator Recipients: Implications of Explainable Machine Learning Models. Journal of Cardiovascular Development and Disease. 2022; 9(9):310. https://doi.org/10.3390/jcdd9090310
Chicago/Turabian StyleDeng, Yu, Sijing Cheng, Hao Huang, Xi Liu, Yu Yu, Min Gu, Chi Cai, Xuhua Chen, Hongxia Niu, and Wei Hua. 2022. "Toward Better Risk Stratification for Implantable Cardioverter-Defibrillator Recipients: Implications of Explainable Machine Learning Models" Journal of Cardiovascular Development and Disease 9, no. 9: 310. https://doi.org/10.3390/jcdd9090310