Predicting delayed cerebral ischemia after subarachnoid hemorrhage using physiological time series data

• Published in: Journal of clinical monitoring and computing Vol. 33; no. 1; pp. 95 - 105
• Main Authors: Park, Soojin, Megjhani, Murad, Frey, Hans-Peter, Grave, Edouard, Wiggins, Chris, Terilli, Kalijah L., Roh, David J., Velazquez, Angela, Agarwal, Sachin, Connolly, E. Sander, Schmidt, J. Michael, Claassen, Jan, Elhadad, Noemie
• Format: Journal Article
• Language: English
• Published: Dordrecht Springer Netherlands 01.02.2019; Springer Nature B.V
• Subjects: Aged; Anesthesiology; Aneurysms; Area Under Curve; Blood pressure; Brain Ischemia - diagnostic imaging; Classifiers; Critical Care; Critical Care Medicine; Datasets; Demographics; Diagnosis, Computer-Assisted - methods; Evaluation; False Positive Reactions; Feature extraction; Female; Glasgow Coma Scale; Health Sciences; Heart rate; Hemorrhage; Humans; Intensive; Ischemia; Least-Squares Analysis; Male; Mathematical models; Medicine; Medicine & Public Health; Middle Aged; Model accuracy; Original Research; Oxygen content; Patient Admission; Physiology; Prediction models; Predictive Value of Tests; Reproducibility of Results; Respiratory rate; Risk Factors; Severity of Illness Index; Statistics for Life Sciences; Stroke; Subarachnoid Hemorrhage - diagnostic imaging; Support Vector Machine; Support vector machines; Tertiary Care Centers; Time Factors; Time series; Critical care; Random kernels; Subarachnoid hemorrhage; Time series; Machine learning
• ISSN: 1387-1307; 1573-2614; 1573-2614
• DOI: 10.1007/s10877-018-0132-5

To develop and validate a prediction model for delayed cerebral ischemia (DCI) after subarachnoid hemorrhage (SAH) using a temporal unsupervised feature engineering approach, demonstrating improved precision over standard features. 488 consecutive SAH admissions from 2006 to 2014 to a tertiary care hospital were included. Models were trained on 80%, while 20% were set aside for validation testing. Baseline information and standard grading scales were evaluated: age, sex, Hunt Hess grade, modified Fisher Scale (mFS), and Glasgow Coma Scale (GCS). An unsupervised approach applying random kernels was used to extract features from physiological time series (systolic and diastolic blood pressure, heart rate, respiratory rate, and oxygen saturation). Classifiers (Partial Least Squares, linear and kernel Support Vector Machines) were trained on feature subsets of the derivation dataset. Models were applied to the validation dataset. The performances of the best classifiers on the validation dataset are reported by feature subset. Standard grading scale (mFS): AUC 0.58. Combined demographics and grading scales: AUC 0.60. Random kernel derived physiologic features: AUC 0.74. Combined baseline and physiologic features with redundant feature reduction: AUC 0.77. Current DCI prediction tools rely on admission imaging and are advantageously simple to employ. However, using an agnostic and computationally inexpensive learning approach for high-frequency physiologic time series data, we demonstrated that our models achieve higher classification accuracy.