Wei-Long Zheng1, Edilberto Amorim2, Jin Jing3, Wendong Ge3, Shenda Hong4, Ona Wu5, Mohammad Ghassemi6, Jong Woo Lee7, Adithya Sivaraju8, Trudy Pang9, Susan T Herman10, Nicolas Gaspard11, Barry J Ruijter12, Jimeng Sun4, Marleen C Tjepkema-Cloostermans13, Jeannette Hofmeijer14, Michel J A M van Putten15, M Brandon Westover16. 1. Department of Neurology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA. Electronic address: weilonglive@gmail.com. 2. Department of Neurology, University of California, San Francisco, San Francisco, CA, USA. 3. Department of Neurology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA. 4. Department of Computer Science, University of Illinois at Urbana Champaign, Champaign, IL, USA. 5. Department of Radiology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA. 6. Department of Computer Science and Engineering, Michigan State University, East Lansing, MI, USA; Department of Electrical Engineering and Computer Science, Massachusetts Institute of Technology, Cambridge, MA, USA. 7. Department of Neurology, Brigham and Women's Hospital, Boston, MA, USA. 8. Department of Neurology, Yale School of Medicine, New Haven, CT, USA. 9. Department of Neurology, Beth Israel Deaconess Medical Center, Boston, MA, USA. 10. Barrow Neurological Institute, Phoenix, AZ, USA. 11. Department of Neurology, Université Libre de Bruxelles, Brussels, Belgium. 12. Department of Clinical Neurophysiology, University of Twente, Enschede, the Netherlands. 13. Departments of Neurology and Clinical Neurophysiology, Medisch Spectrum Twente, Enschede, the Netherlands; Department of Neurology, Rijnstate Hospital, Arnhem, the Netherlands. 14. Department of Clinical Neurophysiology, University of Twente, Enschede, the Netherlands; Department of Neurology, Rijnstate Hospital, Arnhem, the Netherlands. 15. Department of Clinical Neurophysiology, University of Twente, Enschede, the Netherlands; Departments of Neurology and Clinical Neurophysiology, Medisch Spectrum Twente, Enschede, the Netherlands. 16. Department of Neurology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA. Electronic address: mwestover@mgh.harvard.edu.
Abstract
OBJECTIVE: Electroencephalography (EEG) is an important tool for neurological outcome prediction after cardiac arrest. However, the complexity of continuous EEG data limits timely and accurate interpretation by clinicians. We develop a deep neural network (DNN) model to leverage complex EEG trends for early and accurate assessment of cardiac arrest coma recovery likelihood. METHODS: We developed a multiscale DNN combining convolutional neural networks (CNN) and recurrent neural networks (long short-term memory [LSTM]) using EEG and demographic information (age, gender, shockable rhythm) from a multicenter cohort of 1,038 cardiac arrest patients. The CNN learns EEG feature representations while the multiscale LSTM captures short-term and long-term EEG dynamics on multiple time scales. Poor outcome is defined as a Cerebral Performance Category (CPC) score of 3-5 and good outcome as CPC score 1-2 at 3-6 months after cardiac arrest. Performance is evaluated using area under the receiver operating characteristic curve (AUC) and calibration error. RESULTS: Model performance increased with EEG duration, with AUC increasing from 0.83 (95% Confidence Interval [CI] 0.79-0.87 at 12h to 0.91 (95%CI 0.88-0.93) at 66h. Sensitivity of good and poor outcome prediction was 77% and 75% at a specificity of 90%, respectively. Sensitivity of poor outcome was 50% at a specificity of 99%. Predicted probability was well matched to the observation frequency of poor outcomes, with a calibration error of 0.11 [0.09-0.14]. CONCLUSIONS: These results demonstrate that incorporating EEG evolution over time improves the accuracy of neurologic outcome prediction for patients with coma after cardiac arrest.
OBJECTIVE: Electroencephalography (EEG) is an important tool for neurological outcome prediction after cardiac arrest. However, the complexity of continuous EEG data limits timely and accurate interpretation by clinicians. We develop a deep neural network (DNN) model to leverage complex EEG trends for early and accurate assessment of cardiac arrest coma recovery likelihood. METHODS: We developed a multiscale DNN combining convolutional neural networks (CNN) and recurrent neural networks (long short-term memory [LSTM]) using EEG and demographic information (age, gender, shockable rhythm) from a multicenter cohort of 1,038 cardiac arrest patients. The CNN learns EEG feature representations while the multiscale LSTM captures short-term and long-term EEG dynamics on multiple time scales. Poor outcome is defined as a Cerebral Performance Category (CPC) score of 3-5 and good outcome as CPC score 1-2 at 3-6 months after cardiac arrest. Performance is evaluated using area under the receiver operating characteristic curve (AUC) and calibration error. RESULTS: Model performance increased with EEG duration, with AUC increasing from 0.83 (95% Confidence Interval [CI] 0.79-0.87 at 12h to 0.91 (95%CI 0.88-0.93) at 66h. Sensitivity of good and poor outcome prediction was 77% and 75% at a specificity of 90%, respectively. Sensitivity of poor outcome was 50% at a specificity of 99%. Predicted probability was well matched to the observation frequency of poor outcomes, with a calibration error of 0.11 [0.09-0.14]. CONCLUSIONS: These results demonstrate that incorporating EEG evolution over time improves the accuracy of neurologic outcome prediction for patients with coma after cardiac arrest.
Authors: Sunil B Nagaraj; Marleen C Tjepkema-Cloostermans; Barry J Ruijter; Jeannette Hofmeijer; Michel J A M van Putten Journal: Clin Neurophysiol Date: 2018-10-27 Impact factor: 3.708
Authors: Clifton W Callaway; Michael W Donnino; Ericka L Fink; Romergryko G Geocadin; Eyal Golan; Karl B Kern; Marion Leary; William J Meurer; Mary Ann Peberdy; Trevonne M Thompson; Janice L Zimmerman Journal: Circulation Date: 2015-11-03 Impact factor: 29.690
Authors: Teresa L May; Christine W Lary; Richard R Riker; Hans Friberg; Nainesh Patel; Eldar Søreide; John A McPherson; Johan Undén; Robert Hand; Kjetil Sunde; Pascal Stammet; Stein Rubertsson; Jan Belohlvaek; Allison Dupont; Karen G Hirsch; Felix Valsson; Karl Kern; Farid Sadaka; Johan Israelsson; Josef Dankiewicz; Niklas Nielsen; David B Seder; Sachin Agarwal Journal: Intensive Care Med Date: 2019-03-08 Impact factor: 17.440
Authors: Jin Jing; Haoqi Sun; Jennifer A Kim; Aline Herlopian; Ioannis Karakis; Marcus Ng; Jonathan J Halford; Douglas Maus; Fonda Chan; Marjan Dolatshahi; Carlos Muniz; Catherine Chu; Valeria Sacca; Jay Pathmanathan; Wendong Ge; Justin Dauwels; Alice Lam; Andrew J Cole; Sydney S Cash; M Brandon Westover Journal: JAMA Neurol Date: 2020-01-01 Impact factor: 18.302
Authors: Marleen C Tjepkema-Cloostermans; Catarina da Silva Lourenço; Barry J Ruijter; Selma C Tromp; Gea Drost; Francois H M Kornips; Albertus Beishuizen; Frank H Bosch; Jeannette Hofmeijer; Michel J A M van Putten Journal: Crit Care Med Date: 2019-10 Impact factor: 7.598
Authors: Marleen C Tjepkema-Cloostermans; Jeannette Hofmeijer; Albertus Beishuizen; Harold W Hom; Michiel J Blans; Frank H Bosch; Michel J A M van Putten Journal: Crit Care Med Date: 2017-08 Impact factor: 7.598
Authors: M Brandon Westover; Mouhsin M Shafi; Shinung Ching; Jessica J Chemali; Patrick L Purdon; Sydney S Cash; Emery N Brown Journal: J Neurosci Methods Date: 2013-07-23 Impact factor: 2.390
Authors: Claudio Sandroni; Alain Cariou; Fabio Cavallaro; Tobias Cronberg; Hans Friberg; Cornelia Hoedemaekers; Janneke Horn; Jerry P Nolan; Andrea O Rossetti; Jasmeet Soar Journal: Intensive Care Med Date: 2014-11-15 Impact factor: 17.440
Authors: Barry J Ruijter; Marleen C Tjepkema-Cloostermans; Selma C Tromp; Walter M van den Bergh; Norbert A Foudraine; Francois H M Kornips; Gea Drost; Erik Scholten; Frank H Bosch; Albertus Beishuizen; Michel J A M van Putten; Jeannette Hofmeijer Journal: Ann Neurol Date: 2019-06-24 Impact factor: 10.422