Non-invasive method and system for characterizing cardiovascular systems

• US 9,289,150 B1
• Filed: 08/19/2013
• Issued: 03/22/2016
• Est. Priority Date: 08/17/2012
• Status: Active Grant

• Current Assignee: Analytics For Life
• Original Assignee: Analytics For Life
• Inventors: Gupta, Sunny, Yazdi, Mohsen Najafi, Burton, Timothy William Fawcett, Ramchandani, Shyamlal, Exner, Derek Vincent
• Primary Examiner(s): LaVert, Nicole F.
• Assistant Examiner(s): MAHMOOD, NADIA AHMAD
• Application Number:
  US13/970,580
• Time in Patent Office: 946 Days
• Field of Search: 600/512, 600/523
• US Class Current: 1/1
• CPC Class Codes: A61B 2576/023 for the heart
• 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 5/021 Measuring pressure in heart or blood vessels A61B5/0205 takes precedence
• 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 5/339 Displays specially adapted therefor
• CPC Class Codes: A61B 5/341 Vectorcardiography [VCG]
• CPC Class Codes: A61B 5/7235 Details of waveform analysis A61B5/349 takes precedence
• CPC Class Codes: G16H 30/40 for processing medical images, e.g. editing

Abstract

The present disclosure uses physiological data, ECG signals as an example, to evaluate cardiac structure and function in mammals. Two approaches are presented, e.g., a model-based analysis and a space-time analysis. The first method uses a modified Matching Pursuit (MMP) algorithm to find a noiseless model of the ECG data that is sparse and does not assume periodicity of the signal. After the model is derived, various metrics and subspaces are extracted to image and characterize cardiovascular tissues using complex-sub-harmonic-frequencies (CSF) quasi-periodic and other mathematical methods. In the second method, space-time domain is divided into a number of regions, the density of the ECG signal is computed in each region and inputted into a learning algorithm to image and characterize the tissues.

Claims

1. A method for localizing and characterizing both the architectural features and function of cardiovascular tissues, comprising the steps of

obtaining ECG data for the heart;

processing the ECG data to localize, image, and characterize architectural features and function of tissues without use of other measuring devices or invasive procedures, the processing including inputting N-dimensional ECG data to a modified moving average filter to filter the N-dimensional ECG data and performing a phase space transformation process on the filtered N-dimensional ECG data in which a dynamically rich system is synchronized with a physiological signal; and

using phase information to determine a location of an architectural feature or function of the tissues to display an abnormality associated with the tissues in a 3-D image.

2. The method of claim 1, wherein processing the ECG data further comprises creating a phase space diagram, wherein the phase space diagram is used to localize, image, and characterize the architectural features and the function of the tissues.

3. The method of claim 2, further comprising modeling phase space information on a reconstructed 3-D model to localize and to characterize the architectural features and the function of the tissues.

4. The method of claim 1, wherein the ECG data comprises at least one heart beat cycle and wherein the abnormality in the heart can be detected within a predetermined time interval and wherein the predetermined time interval is used to localize and to characterize the architectural features and the function of the tissues.

5. The method of claim 4, wherein the cardiac cycle corresponds to a vector sum electrical activation pathway through the heart, the method further comprising using the pathway with at least 24 dimensional dynamical phase space density information to determine the location and to characterize the architectural features and the function of the tissues.

6. The method of claim 4, further comprising locating and index the abnormality or pathologic cardiovascular tissue to a library of imaging modalities based on mining and linking ECG'"'"'s dynamical phase space density information or other cardiovascular physiological data to associated imaging modalities.

7. The method of claim 4, wherein the predetermined time period is at least 50 seconds and typically from 100 and 700 seconds.

8. The method of claim 1, wherein the phase space transformation process magnifies dynamic features of the filtered N-dimensional ECG data.

9. The method of claim 1, further comprising synchronizing the filtered N-dimensional ECG data with a Rossler system or other quasi-periodic dynamically rich differential equations.

10. The method of claim 1, further comprising determining a dictionary that spans the input ECG data, wherein the dictionary is a family of waveforms that is used to decompose the input ECG data.

11. The method of claim 10, further comprising determining:

a sparse model from the dictionary for the physiological signal under study, the determining being an iterative process that, at each step, chooses a dictionary atom that best correlates with the signal;

continuing the iterative process until a pre-defined stopping condition occurs; and

determining coefficients of selected atoms.

12. The method of claim 11, further comprising correlating the ECG data with all atoms to select a best atom in each iteration in order to minimize a pre-defined cost function that quantifies a distance between the model and the target waveform.

13. The method of claim 11, further comprising representing an output of the iterative process as a 3-D phase space plot, wherein the 3-D phase space plot illustrates cardiac electrical conduction patterns, and associated alterations in tissue architecture, structure and function.

14. The method of claim 11, further comprising:

extracting different subspaces from the derived model; and

extracting predetermined subspaces that represent a dynamical abnormality in the tissue architecture, structure and function.

15. The method of claim 1, further comprising substituting continuous blood pressure, pulse oximetry, or other physiologic data for the ECG data.

16. The method of claim 1, further comprising associating an effectiveness of patient therapies that includes using the model results to guide a treatment or intervention.

17. The method of claim 1, further comprising modeling the ECG data to link 17 myocardial segments in the ventricle to all imaging modalities of the heart.

20. The method of claim 1, further comprising modeling the ECG data to link 17 myocardial segments in the ventricle to all imaging modalities of the heart.

18. A method for localizing and characterizing both the architectural features and function of cardiovascular tissues, comprising the steps of

obtaining ECG data for the heart;

processing the ECG data to localize, image, and characterize architectural features and function of tissues without use of other measuring devices or invasive procedures;

using phase information to determine a location of an architectural feature or function of the tissues to display an abnormality associated with the tissues in a 3-D image;

using at least 24 variables corresponding to the at least 24 dimensional dynamical space density as terms that are selected in nonlinear combinations selected from a list comprising sin, cos, cos h, sin h, Rossler functions, product, division, addition, subtraction, Gaussian, exponential functions and become candidates based on the genetic operators selected from a second list comprising inheritance, mutation, selection, and crossover;

generating offspring function combinations that are evaluated and optimized by freezing all but one variable;

optimizing an unfrozen variable to reduce an absolute error of a model;

optimizing in a sequence, the other variables until all at least 24 variable have a lowest error;

using a fitness function to determine a solution having a lowest absolute error; and

continuing until a highest-ranking solution'"'"'s fitness has reached a plateau such that successive iterations no longer produce better results.

19. The method of claim 18, further comprising:

using the at least 24 variables in a genetic algorithm; and

modeling, using the at least 24 variables to link 17 myocardial segments in the ventricle to tomographic imaging modalities of the heart.