Deep learning for outcome prediction of postanoxic coma

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

3 Citations (Scopus)

Abstract

Electroencephalography (EEG) is increasingly used to assist in outcome prediction for patients with a postanoxic coma after cardiac arrest. Current literature shows that neurological outcome is invariably poor if the EEG remains iso-electric or low-voltage at 24 h after cardiac arrest or if it shows burst-suppression with identical bursts; such patterns are observed in approximately 30-50% of patients. Return of continuous EEG rhythms within 12 h after cardiac arrest predicts good neurological outcome with sensitivities in the range of 30 to 50% at specificities near 100%. In previous work, we reported on the Cerebral Recovery Index to assist in the visual assessment of the EEG. In this paper, we explore a deep learning approach, using a convolutional neural network for outcome prediction in patients with a postanoxic encephalopathy. Using EEGs from 287 patients at 12 h after cardiac arrest and 399 patients at 24 h after cardiac arrest, we trained and validated a convolutional neural network with raw EEG data (18 channels, longitudinal bipolar montage). As the outcome measure, we used the Cerebral Performance Category scale (CPC), dichotomized between good (CPC score 1-2) and poor outcome (CPC score 3-5). Using 5 minute artifact-free epochs from the continuous EEG recordings partitioned into 10 s snippets, we trained the convolutional neural network using 80% of the patients. Validation was performed with EEGs from the remaining 20% of patients. Outcome prediction was most accurate at 12 h after cardiac arrest, with a sensitivity of 58% at a specificity of 100% for the prediction of poor outcome. Good neurological outcome could be predicted at 12 h after cardiac arrest with a sensitivity of 58% at a specificity of 97%. In conclusion, we present a classifier for the prediction of neurological outcome after cardiac arrest, based on a convolutional neural network, providing reliable and objective prognostic information.

Original languageEnglish
Title of host publicationEMBEC and NBC 2017
Subtitle of host publication Joint Conference of the European Medical and Biological Engineering Conference EMBEC 2017 and the Nordic-Baltic Conference on Biomedical Engineering and Medical Physics, NBC 2017
PublisherSpringer
Pages506-509
Number of pages4
ISBN (Electronic)978-981-10-5122-7
ISBN (Print)9789811051210
DOIs
Publication statusPublished - 2017

Fingerprint

Electroencephalography
Neural networks
Deep learning
Classifiers
Recovery
Electric potential

Keywords

  • Cardiac arrest
  • Cerebral Recovery Index
  • EEG monitoring
  • Electroencephalography
  • ICU
  • Postanoxic encephalopathy
  • Prognostication

Cite this

van Putten, M. J. A. M., Hofmeijer, J., Ruijter, B. J., & Tjepkema-Cloostermans, M. C. (2017). Deep learning for outcome prediction of postanoxic coma. In EMBEC and NBC 2017: Joint Conference of the European Medical and Biological Engineering Conference EMBEC 2017 and the Nordic-Baltic Conference on Biomedical Engineering and Medical Physics, NBC 2017 (pp. 506-509). Springer. https://doi.org/10.1007/978-981-10-5122-7_127
van Putten, Michel J.A.M. ; Hofmeijer, Jeannette ; Ruijter, B. J. ; Tjepkema-Cloostermans, Marleen C. / Deep learning for outcome prediction of postanoxic coma. EMBEC and NBC 2017: Joint Conference of the European Medical and Biological Engineering Conference EMBEC 2017 and the Nordic-Baltic Conference on Biomedical Engineering and Medical Physics, NBC 2017. Springer, 2017. pp. 506-509
@inproceedings{0ac6f1a58bcd4f799b406d34854eed2f,
title = "Deep learning for outcome prediction of postanoxic coma",
abstract = "Electroencephalography (EEG) is increasingly used to assist in outcome prediction for patients with a postanoxic coma after cardiac arrest. Current literature shows that neurological outcome is invariably poor if the EEG remains iso-electric or low-voltage at 24 h after cardiac arrest or if it shows burst-suppression with identical bursts; such patterns are observed in approximately 30-50{\%} of patients. Return of continuous EEG rhythms within 12 h after cardiac arrest predicts good neurological outcome with sensitivities in the range of 30 to 50{\%} at specificities near 100{\%}. In previous work, we reported on the Cerebral Recovery Index to assist in the visual assessment of the EEG. In this paper, we explore a deep learning approach, using a convolutional neural network for outcome prediction in patients with a postanoxic encephalopathy. Using EEGs from 287 patients at 12 h after cardiac arrest and 399 patients at 24 h after cardiac arrest, we trained and validated a convolutional neural network with raw EEG data (18 channels, longitudinal bipolar montage). As the outcome measure, we used the Cerebral Performance Category scale (CPC), dichotomized between good (CPC score 1-2) and poor outcome (CPC score 3-5). Using 5 minute artifact-free epochs from the continuous EEG recordings partitioned into 10 s snippets, we trained the convolutional neural network using 80{\%} of the patients. Validation was performed with EEGs from the remaining 20{\%} of patients. Outcome prediction was most accurate at 12 h after cardiac arrest, with a sensitivity of 58{\%} at a specificity of 100{\%} for the prediction of poor outcome. Good neurological outcome could be predicted at 12 h after cardiac arrest with a sensitivity of 58{\%} at a specificity of 97{\%}. In conclusion, we present a classifier for the prediction of neurological outcome after cardiac arrest, based on a convolutional neural network, providing reliable and objective prognostic information.",
keywords = "Cardiac arrest, Cerebral Recovery Index, EEG monitoring, Electroencephalography, ICU, Postanoxic encephalopathy, Prognostication",
author = "{van Putten}, {Michel J.A.M.} and Jeannette Hofmeijer and Ruijter, {B. J.} and Tjepkema-Cloostermans, {Marleen C.}",
year = "2017",
doi = "10.1007/978-981-10-5122-7_127",
language = "English",
isbn = "9789811051210",
pages = "506--509",
booktitle = "EMBEC and NBC 2017",
publisher = "Springer",

}

van Putten, MJAM, Hofmeijer, J, Ruijter, BJ & Tjepkema-Cloostermans, MC 2017, Deep learning for outcome prediction of postanoxic coma. in EMBEC and NBC 2017: Joint Conference of the European Medical and Biological Engineering Conference EMBEC 2017 and the Nordic-Baltic Conference on Biomedical Engineering and Medical Physics, NBC 2017. Springer, pp. 506-509. https://doi.org/10.1007/978-981-10-5122-7_127

Deep learning for outcome prediction of postanoxic coma. / van Putten, Michel J.A.M.; Hofmeijer, Jeannette ; Ruijter, B. J.; Tjepkema-Cloostermans, Marleen C.

EMBEC and NBC 2017: Joint Conference of the European Medical and Biological Engineering Conference EMBEC 2017 and the Nordic-Baltic Conference on Biomedical Engineering and Medical Physics, NBC 2017. Springer, 2017. p. 506-509.

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

TY - GEN

T1 - Deep learning for outcome prediction of postanoxic coma

AU - van Putten, Michel J.A.M.

AU - Hofmeijer, Jeannette

AU - Ruijter, B. J.

AU - Tjepkema-Cloostermans, Marleen C.

PY - 2017

Y1 - 2017

N2 - Electroencephalography (EEG) is increasingly used to assist in outcome prediction for patients with a postanoxic coma after cardiac arrest. Current literature shows that neurological outcome is invariably poor if the EEG remains iso-electric or low-voltage at 24 h after cardiac arrest or if it shows burst-suppression with identical bursts; such patterns are observed in approximately 30-50% of patients. Return of continuous EEG rhythms within 12 h after cardiac arrest predicts good neurological outcome with sensitivities in the range of 30 to 50% at specificities near 100%. In previous work, we reported on the Cerebral Recovery Index to assist in the visual assessment of the EEG. In this paper, we explore a deep learning approach, using a convolutional neural network for outcome prediction in patients with a postanoxic encephalopathy. Using EEGs from 287 patients at 12 h after cardiac arrest and 399 patients at 24 h after cardiac arrest, we trained and validated a convolutional neural network with raw EEG data (18 channels, longitudinal bipolar montage). As the outcome measure, we used the Cerebral Performance Category scale (CPC), dichotomized between good (CPC score 1-2) and poor outcome (CPC score 3-5). Using 5 minute artifact-free epochs from the continuous EEG recordings partitioned into 10 s snippets, we trained the convolutional neural network using 80% of the patients. Validation was performed with EEGs from the remaining 20% of patients. Outcome prediction was most accurate at 12 h after cardiac arrest, with a sensitivity of 58% at a specificity of 100% for the prediction of poor outcome. Good neurological outcome could be predicted at 12 h after cardiac arrest with a sensitivity of 58% at a specificity of 97%. In conclusion, we present a classifier for the prediction of neurological outcome after cardiac arrest, based on a convolutional neural network, providing reliable and objective prognostic information.

AB - Electroencephalography (EEG) is increasingly used to assist in outcome prediction for patients with a postanoxic coma after cardiac arrest. Current literature shows that neurological outcome is invariably poor if the EEG remains iso-electric or low-voltage at 24 h after cardiac arrest or if it shows burst-suppression with identical bursts; such patterns are observed in approximately 30-50% of patients. Return of continuous EEG rhythms within 12 h after cardiac arrest predicts good neurological outcome with sensitivities in the range of 30 to 50% at specificities near 100%. In previous work, we reported on the Cerebral Recovery Index to assist in the visual assessment of the EEG. In this paper, we explore a deep learning approach, using a convolutional neural network for outcome prediction in patients with a postanoxic encephalopathy. Using EEGs from 287 patients at 12 h after cardiac arrest and 399 patients at 24 h after cardiac arrest, we trained and validated a convolutional neural network with raw EEG data (18 channels, longitudinal bipolar montage). As the outcome measure, we used the Cerebral Performance Category scale (CPC), dichotomized between good (CPC score 1-2) and poor outcome (CPC score 3-5). Using 5 minute artifact-free epochs from the continuous EEG recordings partitioned into 10 s snippets, we trained the convolutional neural network using 80% of the patients. Validation was performed with EEGs from the remaining 20% of patients. Outcome prediction was most accurate at 12 h after cardiac arrest, with a sensitivity of 58% at a specificity of 100% for the prediction of poor outcome. Good neurological outcome could be predicted at 12 h after cardiac arrest with a sensitivity of 58% at a specificity of 97%. In conclusion, we present a classifier for the prediction of neurological outcome after cardiac arrest, based on a convolutional neural network, providing reliable and objective prognostic information.

KW - Cardiac arrest

KW - Cerebral Recovery Index

KW - EEG monitoring

KW - Electroencephalography

KW - ICU

KW - Postanoxic encephalopathy

KW - Prognostication

UR - http://www.scopus.com/inward/record.url?scp=85021757747&partnerID=8YFLogxK

U2 - 10.1007/978-981-10-5122-7_127

DO - 10.1007/978-981-10-5122-7_127

M3 - Conference contribution

SN - 9789811051210

SP - 506

EP - 509

BT - EMBEC and NBC 2017

PB - Springer

ER -

van Putten MJAM, Hofmeijer J, Ruijter BJ, Tjepkema-Cloostermans MC. Deep learning for outcome prediction of postanoxic coma. In EMBEC and NBC 2017: Joint Conference of the European Medical and Biological Engineering Conference EMBEC 2017 and the Nordic-Baltic Conference on Biomedical Engineering and Medical Physics, NBC 2017. Springer. 2017. p. 506-509 https://doi.org/10.1007/978-981-10-5122-7_127