Shinichi Goto1, Shinya Goto2, Karen S Pieper3, Jean-Pierre Bassand4, A John Camm5, David A Fitzmaurice6, Samuel Z Goldhaber7, Sylvia Haas8, Alexander Parkhomenko9, Ali Oto10, Frank Misselwitz11, Alexander G G Turpie12, Freek W A Verheugt13, Keith A A Fox14, Bernard J Gersh15, Ajay K Kakkar16. 1. Department of Cardiology, Keio University School of Medicine, Tokyo, Japan. 2. Department of Medicine (Cardiology), Tokai University School of Medicine, Isehara, Japan. 3. Thrombosis Research Institute, London, UK. 4. Thrombosis Research Institute, London, UK and University of Besançon, Besançon, France. 5. Cardiology Clinical Academic Group, Molecular & Clinical Sciences Institute, St. George's University of London, London, UK. 6. Department of Cardio-respiratory Primary Care, Warwick Medical School, University of Warwick, Coventry, UK. 7. Harvard Medical School, Boston, MA, USA. 8. Formerly Klinikum rechts der Isar, Technical University of Munich, Munich, Germany. 9. National Scientific Center, MD Strazhesko Institute of Cardiology, Kiev, Ukraine. 10. Chairman, Department of Cardiology, MHG, Memorial Ankara Hospital, Ankara, Turkey. 11. Bayer AG, Pharmaceuticals Division, Berlin, Germany. 12. McMaster University, Hamilton, Ontario, Canada. 13. Department of Cardiology, Onze Lieve Vrouwe Gasthuis (OLVG), Amsterdam, Netherlands. 14. Edinburgh Centre for Cardiovascular Science, University of Edinburgh, Edinburgh, UK. 15. Department of Cardiovascular Medicine, Mayo Clinic College of Medicine, Rochester, MN, USA. 16. Thrombosis Research Institute and University College London, London, UK.
Abstract
BACKGROUND: Most clinical risk stratification models are based on measurement at a single time-point rather than serial measurements. Artificial intelligence (AI) is able to predict one-dimensional outcomes from multi-dimensional datasets. Using data from GARFIELD-AF registry, a new AI model was developed for predicting clinical outcomes in atrial fibrillation (AF) patients up to 1 year based on sequential measures of PT-INR within 30 days of enrolment. METHODS AND RESULTS: Patients with newly diagnosed AF who were treated with vitamin K antagonists (VKA) and had at least 3 measurements of PT-INR taken over the first 30 days after prescription were analyzed. The AI model was constructed with multilayer neural network including long short-term memory (LSTM) and one-dimensional convolution layers. The neural network was trained using PT-INR measurements within days 0-30 after starting treatment and clinical outcomes over days 31-365 in a derivation cohort (cohorts 1-3; n = 3185). Accuracy of the AI model at predicting major bleed, stroke/SE, and death was assessed in a validation cohort (cohorts 4-5; n = 1523). The model's c-statistic for predicting major bleed, stroke/SE, and all-cause death was 0.75, 0.70, and 0.61, respectively. CONCLUSIONS: Using serial PT-INR values collected within 1 month after starting VKA, the new AI model performed better than time in therapeutic range (TTR) at predicting clinical outcomes occurring up to 12 months thereafter. Serial PT-INR values contain important information that can be analyzed by computer to help predict adverse clinical outcomes.
BACKGROUND: Most clinical risk stratification models are based on measurement at a single time-point rather than serial measurements. Artificial intelligence (AI) is able to predict one-dimensional outcomes from multi-dimensional datasets. Using data from GARFIELD-AF registry, a new AI model was developed for predicting clinical outcomes in atrial fibrillation (AF) patients up to 1 year based on sequential measures of PT-INR within 30 days of enrolment. METHODS AND RESULTS:Patients with newly diagnosed AF who were treated with vitamin K antagonists (VKA) and had at least 3 measurements of PT-INR taken over the first 30 days after prescription were analyzed. The AI model was constructed with multilayer neural network including long short-term memory (LSTM) and one-dimensional convolution layers. The neural network was trained using PT-INR measurements within days 0-30 after starting treatment and clinical outcomes over days 31-365 in a derivation cohort (cohorts 1-3; n = 3185). Accuracy of the AI model at predicting major bleed, stroke/SE, and death was assessed in a validation cohort (cohorts 4-5; n = 1523). The model's c-statistic for predicting major bleed, stroke/SE, and all-cause death was 0.75, 0.70, and 0.61, respectively. CONCLUSIONS: Using serial PT-INR values collected within 1 month after starting VKA, the new AI model performed better than time in therapeutic range (TTR) at predicting clinical outcomes occurring up to 12 months thereafter. Serial PT-INR values contain important information that can be analyzed by computer to help predict adverse clinical outcomes.
Authors: Niraj Varma; Iwona Cygankiewicz; Mintu P Turakhia; Hein Heidbuchel; Yu-Feng Hu; Lin Yee Chen; Jean-Philippe Couderc; Edmond M Cronin; Jerry D Estep; Lars Grieten; Deirdre A Lane; Reena Mehra; Alex Page; Rod Passman; Jonathan P Piccini; Ewa Piotrowicz; Ryszard Piotrowicz; Pyotr G Platonov; Antonio Luiz Ribeiro; Robert E Rich; Andrea M Russo; David Slotwiner; Jonathan S Steinberg; Emma Svennberg Journal: Circ Arrhythm Electrophysiol Date: 2021-02-12
Authors: Niraj Varma; Iwona Cygankiewicz; Mintu P Turakhia; Hein Heidbuchel; Yufeng Hu; Lin Yee Chen; Jean-Philippe Couderc; Edmond M Cronin; Jerry D Estep; Lars Grieten; Deirdre A Lane; Reena Mehra; Alex Page; Rod Passman; Jonathan P Piccini; Ewa Piotrowicz; Ryszard Piotrowicz; Pyotr G Platonov; Antonio Luiz Ribeiro; Robert E Rich; Andrea M Russo; David Slotwiner; Jonathan S Steinberg; Emma Svennberg Journal: Cardiovasc Digit Health J Date: 2021-01-29
Authors: Emelia J Benjamin; Alan S Go; Patrice Desvigne-Nickens; Christopher D Anderson; Barbara Casadei; Lin Y Chen; Harry J G M Crijns; Ben Freedman; Mellanie True Hills; Jeff S Healey; Hooman Kamel; Dong-Yun Kim; Mark S Link; Renato D Lopes; Steven A Lubitz; David D McManus; Peter A Noseworthy; Marco V Perez; Jonathan P Piccini; Renate B Schnabel; Daniel E Singer; Robert G Tieleman; Mintu P Turakhia; Isabelle C Van Gelder; Lawton S Cooper; Sana M Al-Khatib Journal: Circulation Date: 2021-01-25 Impact factor: 29.690
Authors: Amitava Banerjee; Suliang Chen; Ghazaleh Fatemifar; Mohamad Zeina; R Thomas Lumbers; Johanna Mielke; Simrat Gill; Dipak Kotecha; Daniel F Freitag; Spiros Denaxas; Harry Hemingway Journal: BMC Med Date: 2021-04-06 Impact factor: 11.150