1
|
Yu D, Williams GW, Aguilar D, Yamal JM, Maroufy V, Wang X, Zhang C, Huang Y, Gu Y, Talebi Y, Wu H. Machine learning prediction of the adverse outcome for nontraumatic subarachnoid hemorrhage patients. Ann Clin Transl Neurol 2020; 7:2178-2185. [PMID: 32990362 PMCID: PMC7664270 DOI: 10.1002/acn3.51208] [Citation(s) in RCA: 14] [Impact Index Per Article: 3.5] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/15/2020] [Revised: 08/13/2020] [Accepted: 09/01/2020] [Indexed: 01/25/2023] Open
Abstract
Objective Subarachnoid hemorrhage (SAH) is often devastating with increased early mortality, particularly in those with presumed delayed cerebral ischemia (DCI). The ability to accurately predict survival for SAH patients during the hospital course would provide valuable information for healthcare providers, patients, and families. This study aims to utilize electronic health record (EHR) data and machine learning approaches to predict the adverse outcome for nontraumatic SAH adult patients. Methods The cohort included nontraumatic SAH patients treated with vasopressors for presumed DCI from a large EHR database, the Cerner Health Facts® EMR database (2000–2014). The outcome of interest was the adverse outcome, defined as death in hospital or discharged to hospice. Machine learning‐based models were developed and primarily assessed by area under the receiver operating characteristic curve (AUC). Results A total of 2467 nontraumatic SAH patients (64% female; median age [interquartile range]: 56 [47–66]) who were treated with vasopressors for presumed DCI were included in the study. 934 (38%) patients died or were discharged to hospice. The model achieved an AUC of 0.88 (95% CI, 0.84–0.92) with only the initial 24 h EHR data, and 0.94 (95% CI, 0.92–0.96) after the next 24 h. Interpretation EHR data and machine learning models can accurately predict the risk of the adverse outcome for critically ill nontraumatic SAH patients. It is possible to use EHR data and machine learning techniques to help with clinical decision‐making.
Collapse
Affiliation(s)
- Duo Yu
- Department of Biostatistics & Data Science, School of Public Health, The University of Texas Health Science Center at Houston (UTHealth), Houston, Texas, USA
| | - George W Williams
- Department of Anesthesiology, McGovern Medical School, The University of Texas Health Science Center at Houston (UTHealth), Houston, Texas, USA
| | - David Aguilar
- Department of Medicine, McGovern Medical School, The University of Texas Health Science Center at Houston (UTHealth), Houston, Texas, USA.,Department of Epidemiology, Human Genetics, and Environmental Sciences, School of Public Health, University of Texas Health Science Center at Houston, Houston, Texas, USA
| | - José-Miguel Yamal
- Department of Biostatistics & Data Science, School of Public Health, The University of Texas Health Science Center at Houston (UTHealth), Houston, Texas, USA
| | - Vahed Maroufy
- Department of Biostatistics & Data Science, School of Public Health, The University of Texas Health Science Center at Houston (UTHealth), Houston, Texas, USA
| | - Xueying Wang
- Department of Biostatistics & Data Science, School of Public Health, The University of Texas Health Science Center at Houston (UTHealth), Houston, Texas, USA
| | - Chenguang Zhang
- Department of Biostatistics & Data Science, School of Public Health, The University of Texas Health Science Center at Houston (UTHealth), Houston, Texas, USA
| | - Yuefan Huang
- Department of Biostatistics & Data Science, School of Public Health, The University of Texas Health Science Center at Houston (UTHealth), Houston, Texas, USA
| | - Yuxuan Gu
- Department of Biostatistics & Data Science, School of Public Health, The University of Texas Health Science Center at Houston (UTHealth), Houston, Texas, USA
| | - Yashar Talebi
- Department of Biostatistics & Data Science, School of Public Health, The University of Texas Health Science Center at Houston (UTHealth), Houston, Texas, USA
| | - Hulin Wu
- Department of Biostatistics & Data Science, School of Public Health, The University of Texas Health Science Center at Houston (UTHealth), Houston, Texas, USA
| |
Collapse
|
2
|
Aggarwal V, Sharma A, Sinha VD. Role of Diffusion-weighted Imaging in Detecting Early Ischemic Brain Injury Following Aneurysmal Subarachnoid Hemorrhage. Asian J Neurosurg 2018; 13:1074-1077. [PMID: 30459871 PMCID: PMC6208208 DOI: 10.4103/ajns.ajns_73_17] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 12/02/2022] Open
Abstract
Background: Aneurysmal SAH is the significant cause of morbidity and mortality in stroke patients. Early brain injury and delayed cerebral ischemia are the two main responsible pathophysiologic processes. Cerebral ischemia needs to be detected early so that early aggressive therapy could be started. Although Diffusion weighted imaging (DWI) has often been utilized for the measurement of acute ischemic strokes, its role in the detection of early cerebral ischemia due to aneurysmal subarachnoid hemorrhage has not been extensively investigated. This study is being carried out to describe the role of DWI in detecting early ischemic brain injury and outcome after aneurysmal SAH. Aim: Efficacy of DWI in detecting ischemic injury and predicting outcome after aneurysmal SAH. Material and Methods: In this prospective study 44 consecutive patients who had aneurysmal SAH; admitted within 7 days of their ictus were included. Hunt and Hess grade on admission and modified Fisher grade of SAH were noted. Plain CT brain and MR DWI was done on day before surgery. Diffusion restriction on DWI was correlated with postoperative neurological deficit, postoperative CT finding and outcome of the patient at 1 month follow-up. Results: DWI revealed restricted diffusion in 12 patients, out of which 1 patient was having infarction in preoperative CT scan, 6 patients were having postoperative deficit in the form of disorientation, hemiparesis and aphasia, and all patients were having infarction in postoperative CT scan. When DWI findings were compared on the basis of postoperative neurological deficit, postoperative CT finding and modified Rankin outcome score at 1month follow-up, results were statistically significant. Conclusion: DWI shows cerebral ischemia much earlier than CT scan in cases of aneurysmal SAH. It has significant correlation with postoperative neurological status and outcome of the patient.
Collapse
Affiliation(s)
- Varun Aggarwal
- Department of Neurosurgery, SMS Hospital, Jaipur, Rajasthan, India
| | - Achal Sharma
- Department of Neurosurgery, SMS Hospital, Jaipur, Rajasthan, India
| | - V D Sinha
- Department of Neurosurgery, SMS Hospital, Jaipur, Rajasthan, India
| |
Collapse
|
3
|
Uryga A, Burzyńska M, Tabakow P, Kasprowicz M, Budohoski KP, Kazimierska A, Smielewski P, Czosnyka M, Goździk W. Baroreflex sensitivity and heart rate variability are predictors of mortality in patients with aneurysmal subarachnoid haemorrhage. J Neurol Sci 2018; 394:112-119. [PMID: 30245190 DOI: 10.1016/j.jns.2018.09.014] [Citation(s) in RCA: 10] [Impact Index Per Article: 1.7] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/18/2018] [Revised: 08/24/2018] [Accepted: 09/11/2018] [Indexed: 01/09/2023]
Abstract
OBJECT We aimed to investigate the link between the autonomic nervous system (ANS) impairment, assessed using baroreflex sensitivity (BRS) and heart rate variability (HRV) indices, and mortality after aneurysmal subarachnoid haemorrhage (aSAH). METHODS A total of 57 patients (56 ± 18 years) diagnosed with aSAH were retrospectively enrolled in the study, where 25% of patients died in the hospital. BRS was calculated using a modified cross-correlation method. Time- and frequency-domain HRV indices were calculated from a time-series of systolic peak intervals of arterial blood pressure signals. Additionally, cerebral autoregulation (CA) was assessed using the mean velocity index (Mxa), where Mxa > 0 indicates impaired CA. RESULTS Both BRS and HRV indices were lower in non-survivors than in survivors. The patients with disturbed BRS and HRV had more extensive haemorrhage in the H-H scale (p = .040) and were more likely to die (p = .013) when compared to patients with the intact ANS. The logistic regression model for mortality included: the APACHE II score (p = .002; OR 0.794) and the normalised high frequency power of the HRV (p < <.001; OR 0.636). A positive relationship was found between the Mxa and BRS (R = 0.48, p = .003), which suggests that increasing BRS is moderately strongly associated with worsening CA. CONCLUSION Our results indicated that lower values of HRV indices and BRS correlate with mortality and that there is a link between cerebral dysautoregulation and the analysed estimates of the ANS in aSAH patients.
Collapse
Affiliation(s)
- Agnieszka Uryga
- Department of Biomedical Engineering, Faculty of Fundamental Problems of Technology, Wroclaw University of Science and Technology, Wroclaw, Poland.
| | - Małgorzata Burzyńska
- Department of Anesthesiology and Intensive Care, Wroclaw Medical University, Wroclaw, Poland
| | - Paweł Tabakow
- Department of Neurosurgery, Wroclaw Medical University, Wroclaw, Poland
| | - Magdalena Kasprowicz
- Department of Biomedical Engineering, Faculty of Fundamental Problems of Technology, Wroclaw University of Science and Technology, Wroclaw, Poland
| | - Karol P Budohoski
- Brain Physics Laboratory, Department of Clinical Neurosciences, Division of Neurosurgery, University of Cambridge, Cambridge, UK
| | - Agnieszka Kazimierska
- Department of Biomedical Engineering, Faculty of Fundamental Problems of Technology, Wroclaw University of Science and Technology, Wroclaw, Poland
| | - Peter Smielewski
- Brain Physics Laboratory, Department of Clinical Neurosciences, Division of Neurosurgery, University of Cambridge, Cambridge, UK
| | - Marek Czosnyka
- Brain Physics Laboratory, Department of Clinical Neurosciences, Division of Neurosurgery, University of Cambridge, Cambridge, UK; Institute of Electronic Systems, Faculty of Electronics and Information Technology, Warsaw University of Technology, Warsaw, Poland
| | - Waldemar Goździk
- Department of Anesthesiology and Intensive Care, Wroclaw Medical University, Wroclaw, Poland
| |
Collapse
|
5
|
Parameter Optimization for Selected Correlation Analysis of Intracranial Pathophysiology. COMPUTATIONAL AND MATHEMATICAL METHODS IN MEDICINE 2015; 2015:652030. [PMID: 26693250 PMCID: PMC4677033 DOI: 10.1155/2015/652030] [Citation(s) in RCA: 4] [Impact Index Per Article: 0.4] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 08/13/2015] [Revised: 10/20/2015] [Accepted: 10/25/2015] [Indexed: 11/17/2022]
Abstract
Recently we proposed a mathematical tool set, called selected correlation analysis, that reliably detects positive and negative correlations between arterial blood pressure (ABP) and intracranial pressure (ICP). Such correlations are associated with severe impairment of the cerebral autoregulation and intracranial compliance, as predicted by a mathematical model. The time resolved selected correlation analysis is based on a windowing technique combined with Fourier-based coherence calculations and therefore depends on several parameters. For real time application of this method at an ICU it is inevitable to adjust this mathematical tool for high sensitivity and distinct reliability. In this study, we will introduce a method to optimize the parameters of the selected correlation analysis by correlating an index, called selected correlation positive (SCP), with the outcome of the patients represented by the Glasgow Outcome Scale (GOS). For that purpose, the data of twenty-five patients were used to calculate the SCP value for each patient and multitude of feasible parameter sets of the selected correlation analysis. It could be shown that an optimized set of parameters is able to improve the sensitivity of the method by a factor greater than four in comparison to our first analyses.
Collapse
|
6
|
The utility of serum procalcitonin in distinguishing systemic inflammatory response syndrome from infection after aneurysmal subarachnoid hemorrhage. Neurocrit Care 2015; 20:375-81. [PMID: 24522762 DOI: 10.1007/s12028-014-9960-4] [Citation(s) in RCA: 18] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 10/25/2022]
Abstract
BACKGROUND Systemic inflammatory response syndrome (SIRS) occurs frequently after aneurysmal subarachnoid hemorrhage (aSAH). It is a clinical challenge to distinguish between SIRS and incipient infection. Procalcitonin (PCT) has been studied among general critical care patients as a biomarker for infection. We hypothesized that PCT could be useful to distinguish SIRS from sepsis in aSAH patients. METHODS Prospective, observational study conducted in the multidisciplinary intensive care unit at Mayo Clinic, Jacksonville, FL between August 2009 and September 2010. Main predictor was serum PCT obtained on admission and with subsequent episodes of SIRS. A level of 0.2 ng/mL or higher was considered as elevated PCT. Main outcome was clinical infection, which was subsequently subcategorized into major (systemic) and minor (localized) infections in the sensitivity analysis. RESULTS Forty consecutive patients were enrolled. Majority (88 %) developed SIRS during the hospitalization. Infection developed in 16 (40 %) patients, with 6 patients meeting criteria for major infection. Overall, PCT was found to be highly specific for all infections and the subcategory of major infections (97 and 93 %, respectively) with related high negative predictive values. Odds ratio for elevated PCT with clinical infections ranged from 25.2 (95 % CI 2.7-233) to 33.3 (95 % CI 4.3-261) for all and major infections, respectively. Related receiver operating characteristic curves for elevated PCT were 0.74 and 0.96 for all and major infections, respectively. CONCLUSIONS Procalcitonin of 0.2 ng/mL or greater was demonstrated to be very specific for sepsis among patients with aSAH. Further studies should validate this result and establish its clinical applicability.
Collapse
|