Probabilistic biomedical parameter estimation apparatus and method of operation therefor

• US 9,375,171 B2
• Filed: 04/28/2011
• Issued: 06/28/2016
• Est. Priority Date: 04/22/2009
• Status: Active Grant

• Current Assignee: Vital Metrix Inc.
• Original Assignee: Streamline Automation LLC
• Inventors: Teixeira, Rodrigo E.
• Primary Examiner(s): Kahelin, Michael
• Assistant Examiner(s): Alter, Mitchell E
• Application Number:
  US13/096,876
• Publication Number:
  US 20120277545A1
• Time in Patent Office: 1,888 Days
• Field of Search: 600/301
• US Class Current: 1/1
• CPC Class Codes: A61B 5/0075 by spectroscopy, i.e. measuring spectra, e.g. Raman spectroscopy, infrared absorption spectroscopy A61B5/0071 takes precedence
• 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 5/14551 for measuring blood gases
• CPC Class Codes: A61B 5/441 Skin evaluation, e.g. for skin disorder diagnosis
• CPC Class Codes: A61B 5/7275 Determining trends in physiological measurement data; Predicting development of a medical condition based on physiological measurements, e.g. determining a risk factor
• CPC Class Codes: G16H 50/30 for calculating health indices; for individual health risk assessment
• CPC Class Codes: G16H 50/50 for simulation or modelling of medical disorders

Abstract

A probabilistic digital signal processor for medical function is described. Initial probability distribution functions are input to a dynamic state-space model, which operates on state and/or model probability distribution functions to generate a prior probability distribution function, which is input to a probabilistic updater. The probabilistic updater integrates sensor data with the prior to generate a posterior probability distribution function passed to a probabilistic sampler, which estimates one or more parameters using the posterior, which is output or re-sampled in an iterative algorithm. For example, the probabilistic processor operates using a physical model on data from a medical meter, where the medical meter uses a first physical parameter, such as blood oxygen saturation levels from a pulse oximeter, to generate a second physical parameter not output by the medical meter, such as a heart stroke volume, a cardiac output flow rate, and/or a blood pressure.

Claims

1. An apparatus for determination of a left-ventricle stroke volume of a heart of a patient, comprising:

a medical monitoring device comprising;

a pulse oximeter;

a blood pressure meter;

a non-physiological sensor; and

a system processor, said system processor configured to process sensor data representing a first physical parameter, said processor comprising;

a dynamic state-space model, said dynamic state-space model comprising a hemodynamic process model configured to provide a prior probability distribution function using a dual estimation algorithm, said dual estimation algorithm using;

(1) input from said pulse oximeter, wherein a blood oxygen saturation and a heart rate are generated using said pulse oximeter;

(2) input from said blood pressure meter; and

(3) input from said non-physiological sensor, said non-physiological sensor comprising an accelerometer configured to measure state of said pulse oximeter; and

a probabilistic processor configured to operate on the prior probability distribution function to produce a posterior probability distribution function,

wherein output correlated with the posterior probability distribution function relates to a second physical parameter, wherein said second physical parameter comprises a physical parameter other than said first physical parameter, wherein said second physical parameter comprises a left ventricle stroke volume of the heart,

wherein said medical monitoring device outputs the left ventricle stroke volume.

2. The apparatus of claim 1, wherein said second physical parameter further comprises a cardiac output flow rate.

3. The apparatus of claim 1, said medical monitoring device further comprising an auscultation sensor, and said dynamic state-space model further comprising:

a hemodynamic dynamic state-space model comprising at least one equation related to blood movement.

4. The apparatus of claim 3, wherein a hemodynamics state parameter input to said hemodynamics dynamic state-space model from said pressure meter comprises an aortic pressure.

5. The apparatus of claim 1, said hemodynamics process model configured to represent state of the patient over time in terms of said second physical parameter;

and

a hemodynamics observation model configured to use a light reading to represent processes involved in collecting the sensor data, wherein said hemodynamics observation model mathematically represents processes involved in use of said biological sensor.

6. The apparatus of claim 5, said hemodynamics process model further comprising:

a heart model configured to relate cardiac output to heart rate; and

a vascular model using a time-varying Windkessel compliance function.

7. The apparatus of claim 1, said dynamic state-space model further comprising:

a physics based probabilistic vascular model configured to represent movement of blood in arteries.

8. The apparatus of claim 1, said medical monitoring device comprising a biomedical monitoring device, wherein the sensor data comprises discrete non-probabilistic input data to said probabilistic processor, said probabilistic processor configured to generate a probabilistic output function.

9. The apparatus of claim 1, said probabilistic processor configured to iteratively circulate a hemodynamics probability distribution function through at least two of:

said dynamic state-space model;

a probabilistic updater; and

a probabilistic sampler.

10. A method for state parameter estimation of a heart of a patient, comprising the steps of:

providing a biomedical device comprising a first monitoring device and a second monitoring device;

providing first sensor data representative of a first physical parameter from said first monitoring device, said first monitoring device comprising a pulse oximeter;

providing second sensor data representative of a second physical parameter from said second monitoring device, said second monitoring device comprising a blood pressure meter;

processing the sensor data using a computer containing a probabilistic digital signal processor configured to use a dual estimation dynamic state-space model, said dynamic state-space model comprising a hemodynamic process model comprising;

(1) a process model receiving first input from said pulse oximeter, wherein a blood oxygen saturation and a heart rate are generated using said pulse oximeter, and input from said blood pressure meter and (2) a noise observation model receiving second input from a non-physiological sensor monitoring physical movement of said first monitoring device;

generating a prior probability distribution function using said dynamic state-space model;

operating on the prior probability distribution function to produce a posterior probability distribution function, wherein output correlated with the posterior probability distribution function relates to a third physical parameter, wherein the third physical parameter comprises a physical parameter other than the first physical parameter, wherein the third physical parameter comprises a physical parameter other than the second physical parameter, wherein said third physical parameter comprises a left ventricle stroke volume of the heart; and

outputting said left ventricle stroke volume.

11. The method of claim 10, wherein the first sensor data comprises spectral readings of skin.

12. The method of claim 10, said hemodynamic process model further comprising third sensor data of a fourth physical parameter from a third monitoring device, wherein said third monitoring device comprises an electrocardiogram configured to generate an electrocardiograph, wherein the third sensor data comprises electrical pulse readings through a lung of the patient.

13. The method of claim 10, said step of operating further comprising the steps of:

using a probabilistic updater to iteratively receive updated sensor data, the sensor data updated in successive iterations of use of said hemodynamics process model; and

combining the sensor data with input from a medical database of an individual in said dual estimation dynamic state-space model to form a patient specific model.

14. The method of claim 10, further comprising the step of:

iteratively converting output of said prior probability distribution function using iterative data of the sensor data to generate iterative output of the posterior probability distribution function.

15. The method of claim 10, further comprising the step of:

processing said prior probability distribution function, wherein said prior probability distribution function comprises a first distribution of a parameter,

wherein said step of processing generates a second distribution of said parameter, wherein said second distribution comprises a narrower distribution compared to said first distribution.

16. The method of claim 15, wherein said second model parameter distribution comprises a distribution of pressure values.

17. The method of claim 10, said probabilistic digital signal processor further configured with at least one of:

a light scattering model configured to relate scattered light readings to the left ventricle stroke volume; and

an absorbance model configured to relate absorbed light readings to the left ventricle stroke volume.

18. The method of claim 10, said dynamic state-space model further comprising:

using the hemodynamic process model to represent state of a patient over time in terms of state parameters; and

using the hemodynamic observation model to represent processes involved in collecting the first sensor data.

19. The method of claim 18, said observation model further comprising the step of:

using a sensor noise model to add noise as a noise probability distribution function to a function representing light values of the sensor data, wherein the added noise comprises movement of sensor noise.