ASSESSMENT AND PREDICTION OF CARDIOVASCULAR STATUS DURING CARDIAC ARREST AND THE POST-RESUSCITATION PERIOD USING SIGNAL PROCESSING AND MACHINE LEARNING

• US 20150065815A1
• Filed: 05/25/2012
• Published: 03/05/2015
• Est. Priority Date: 05/27/2011
• Status: Active Grant

• Current Assignee: Virginia Commonwealth University
• Original Assignee: Virginia Commonwealth University
• Inventors: Najarian, Kayvan, Shandilya, Sharad, Ward, Kevin R.
• Granted Patent: US 9,339,241 B2
• Time in Patent Office: Days
• US Class Current: 600/301
• CPC Class Codes: A61B 2505/01 Emergency care
• CPC Class Codes: A61B 5/0205 Simultaneously evaluating both cardiovascular conditions and different types of body conditions, e.g. heart and respiratory condition
• CPC Class Codes: A61B 5/053 Measuring electrical impedance or conductance of a portion of the body
• CPC Class Codes: A61B 5/0836 Measuring rate of CO2 production
• CPC Class Codes: A61B 5/14551 for measuring blood gases
• CPC Class Codes: A61B 5/316 Modalities, i.e. specific diagnostic methods
• CPC Class Codes: A61B 5/318 Heart-related electrical modalities, e.g. electrocardiography [ECG]
• CPC Class Codes: A61B 5/349 Detecting specific parameters of the electrocardiograph cycle
• CPC Class Codes: A61B 5/4836 Diagnosis combined with treatment in closed-loop systems or methods A61B5/0036 takes precedence
• CPC Class Codes: A61B 5/486 Bio-feedback A61B5/375 takes precedence
• CPC Class Codes: A61B 5/7203 for noise prevention, reduction or removal
• CPC Class Codes: A61B 5/725 using specific filters therefor, e.g. Kalman or adaptive filters A61B5/316 take precedence
• CPC Class Codes: A61B 5/726 using Wavelet transforms
• CPC Class Codes: A61B 5/7267 involving training the classification device
• CPC Class Codes: A61N 1/3925 Monitoring; Protecting
• CPC Class Codes: G16H 40/63 for local operation
• CPC Class Codes: G16H 50/50 for simulation or modelling of medical disorders

Abstract

Real-time, short-term analysis of ECG, by using multiple signal processing and machine learning techniques, is used to determine counter shock success in defibrillation. Combinations of measures when used with machine learning algorithms readily predict successful resuscitation, guide therapy and predict complications. In terms of guiding resuscitation, they may serve as indicators and when to provide counter shocks and at what energy levels they should be provided as well as to serve as indicators of when certain drugs should be provided (in addition to their doses). For cardiac arrest, the system is meant to run in real time during all current resuscitation procedures including post-resuscitation care to detect deterioration for guiding care such as therapeutic hypothermia.

Claims

1. A method for automated monitoring and online assessment of chances of survival for a patient in cardiac arrest comprising:

obtaining an ECG signal from the patient;

preprocessing the ECG signal to remove high frequency noise and baseline jumps caused by noise and interference;

performing non-linear characterization of the preprocessed ECG signal and calculating the prototype distance;

performing feature extraction of the preprocessed ECG signal with complex wavelet transform;

performing attribute extraction from the preprocessed ECG signal;

performing attribute extraction from ETCO₂ signal;

receiving distance values from non-linear characterization of the preprocessed ECG signal, extracted features of the preprocessed time-series ECG signal and attributes extracted from Dual-Tree Complex Wavelet Decomposition of the pre-processed ECG signal, and performing a feature selection with a predictive model;

using machine learning to classify results of the feature selection process;

generating a shock success prediction, which results in return of spontaneous circulation (ROSC);

generating decompensation and re-arrest prediction; and

recommending therapeutic alternatives and medications, thereby guiding therapy.

2. The method of claim 1, further comprising the step of integrating categorical data such as demographics, medical history and medication information into the feature database.

3. The method of claim 1, further comprising the step of performing attribute extraction from other continuous physiologic signals such as tissue impedance, vascular waveform data from piezoelectric sensors or other devices capable of obtaining vascular waveform data, tissue oxygenation signals from near infrared spectroscopy and other tissue oxygenation devices and wherein the step of receiving includes receiving attributes extracted from other physiologic signals and categorical variables.

4. A system for automated monitoring and online assessment of chances of survival for a patient in cardiac arrest comprising:

an ECG device for providing an ECG signal from the patient;

a filter for preprocessing the ECG signal to remove high frequency noise and baseline jumps caused by noise and interference;

a first signal processor for performing non-linear characterization of the preprocessed ECG signal and calculating the prototype distance;

a second signal processor for performing feature extraction of the preprocessed ECG signal with complex wavelet transform;

a third signal processor for performing attribute extraction from the preprocessed ECG signal;

a fourth signal processor for performing attribute extraction from ETCO₂ signal;

feature extraction means receiving distance values from non-linear characterization of the preprocessed ECG signal, extracted features of the preprocessed time-series ECG signal and the attributes extracted from Dual-Tree Complex Wavelet Decomposition of the pre-processed ECG signal and performing a feature selection with a predictive model;

a machine learning system for classifying results of the feature selection process, said machine learning system generating a shock success prediction, which results in return of spontaneous circulation (ROSC), generating decompensation and re-arrest prediction, and recommending therapeutic alternatives and medications, thereby guiding therapy.

5. The system of claim 4, further comprising means for inputting categorical data such as demographics, medical history and medication information into the feature database of the machine learning system.

6. The system of claim 4, further comprising means for performing attribute extraction from other continuous physiologic signals such as tissue impedance, vascular waveform data from piezoelectric sensors or other devices capable of obtaining vascular waveform data, tissue oxygenation signals from near infrared spectroscopy and other tissue oxygenation devices and wherein the extracted attributes from other continuous physiologic signals are input to the machine learning system.