Noninvasive diagnostic system for coronary artery disease

• US 5,036,857 A
• Filed: 10/26/1989
• Issued: 08/06/1991
• Est. Priority Date: 10/26/1989
• Status: Expired due to Term

• Current Assignee: Rutgers State University of New Jersey, University of Medicine and Dentistry of New Jersey (State of New Jersey)
• Original Assignee: Rutgers University
• Inventors: Akay, Metin, Semmlow, John L., Welkowitz, Walter, Kostis, John
• Primary Examiner(s): Jaworski, Francis
• Assistant Examiner(s): MANUEL, GEORGE C
• Application Number:
  US07/426,770
• Time in Patent Office: 649 Days
• Field of Search: 128/715, 128/710, 128/661.04, 128/670, 128/668, 128/680
• US Class Current: 600/528
• CPC Class Codes: A61B 5/024 Detecting, measuring or recording pulse rate or heart rate A61B5/0205, A61B5/021 take precedence
• CPC Class Codes: A61B 5/0285 Measuring or recording phase velocity of blood waves
• CPC Class Codes: A61B 7/04 Electric stethoscopes microphones, acoustic transducers therefor H04R
• CPC Class Codes: G01P 15/0922 of the bending or flexing mode type
• CPC Class Codes: G01P 2015/0831 the mass being of the paddle type having the pivot axis between the longitudinal ends of the mass, e.g. see-saw configuration

Abstract

A method and system for non-invasively detecting Coronary Artery Disease. The method comprises analyzing the diastolic heart sounds detected from a patient'"'"'s chest cavity during the diastolic portion of the heart cycle in order to identify a low level auditory component associated with turbulent blood flow in partially occluded coronary arteries. These diastolic heart sounds are modeled using advanced signal processing techniques such as Autoregressive (AR), Autoregressive Moving Averaging (ARMA) and Eigenvector methods, so that the presence of such an auditory component may be reliably indicated even under high noise conditions. The system includes an acoustic transducer, pulse sensor device, signal processor means and a diagnostic display. Additionally, the system includes a controller for automatically sequencing data collection, analysis and display stages, therefore requiring a minimum of operator interaction. The system further discloses a piezoelectric type low mass accelerometer for use as an acoustic transducer providing higher sensitivity at the high frequencies necessary for detecting coronary artery disease, specifically between 300 to 1200 Hz. In addition, the system discloses a method for automatically identifying and isolating a diastolic segment of a heart sound recording through a "window" placement technique. A window is placed relative to the second heart sound and various criteria are implemented to evaluate the recording for artifacts, i.e., data due to breath sounds, stomach growls, external ambient noise.

Claims

1. An apparatus for noninvasive detection of coronary artery disease comprising:

an acoustic transducer for detecting heart sounds from a patient'"'"'s chest cavity and providing a heart sound signal representative of said detected heart sounds;

frequency analyzer means for providing a parametric model for parametrically modeling said detected heart sounds based on said heart sound signal;

turbulent flow detector means for detecting the presence of auditory component in said parametric model, said component being associated with turbulent blood flow in a partially occluded coronary artery, said turbulent flow detector means including means for applying a decision criterion based on one or more parameters of said parametric model, said decision criterion being indicative of the presence of said auditory component in the range of about 300 to about 1200 Hz; and

display means for presenting diagnostic information based on the detected presence of said auditory component.

2. The apparatus of claim 1 which further comprises time windowing means for isolating within said heart sound signal the detected heart sounds which occur during a selected segment of the cardiac cycle.

3. The apparatus of claim 2 wherein said time windowing means comprises means for delaying from a pulse signal for a predetermined amount of time, to a time just after the last of the prominent heart sounds, and additional means for isolating a predetermined length of time of said detected heart sound signal following said delay for analysis by said frequency analyzer means.

4. The apparatus of claim 3 wherein said delay from said pulse signal varies between 100 to 300 msec., depending on the heart rate, and said duration of the time window is approximately 256 msec.

5. The apparatus of claim 2 wherein said selected segment of the cardiac cycle comprises a diastolic segment of the cardiac cycle.

6. The apparatus of claim 5 wherein said time windowing means comprises:

means for establishing a preliminary data window of a predetermined width upon an indication of the onset of diastole in said cardiac cycle, said preliminary data window comprising three zones, the first of which (zone

     1) includes the expected trailing edge of a second heart sound, the third of which (zone

     3) includes the expected leading edge of a succeeding first heart sound, and the second of which (zone

     2) is approximately centered in the diastolic segment of said heart sound;

means for establishing a threshold level within said preliminary data window for discriminating the occurrence of said second or first heart sound;

means for establishing a test window within said preliminary data window, said test window width being significantly smaller than said preliminary data window width;

means for evaluating a RMS value of said heart sound signal within said test window for a plurality of test window locations within said zones of the preliminary data window;

means for locating a lower diastolic boundary which is defined by the latest in time location of said test window in zone 1 where said RMS value exceeds the value of said threshold level;

means for locating an upper diastolic boundary which is defined by the earliest location in time of said test window in zone 3 where said RMS value exceeds said threshold level;

means for positioning a diastolic window of a predetermined width at the approximate center between said lower and upper diastolic boundaries, said diastolic window thereby isolating the heart sound signal within for subsequent analysis by said frequency analyzer means; and

means for rejecting said heart sound signal data within said diastolic window if a window defined by said lower and upper diastolic boundaries is smaller in width than said predetermined width of the diastolic window.

7. The apparatus of claim 6, said apparatus further including:

means for rejecting said heart sound signal data within said diastolic window if any of said RMS values which exceed said threshold level are located within zone 2.

8. The apparatus of claim 1 wherein said parametric modeling means are Autoregressive (AR) modeling means.

9. The apparatus of claim 1 wherein said parametric modeling means are Eigenvector modeling means.

10. The apparatus of claim 1 wherein said parametric modeling means are Autoregressive Moving Average modeling means.

11. The apparatus of claim 1 wherein said means for applying a decision criterion comprises means for evaluating the power spectral density function of said detected heart sounds model.

12. The apparatus of claim 1 wherein said means for applying a decision criterion comprises means for evaluating the location of the zeros of the linear prediction coefficients of said detected heart sounds model.

13. The apparatus of claim 1 wherein said means for applying a decision criterion comprises means for evaluating the peak power levels of said detected heart sounds model.

14. The apparatus of claim 1 wherein said means for applying a decision criterion comprises means for evaluating the reflection coefficients of said detected heart sounds model.

15. An apparatus for noninvasive detection of coronary artery disease comprising:

an acoustic transducer for detecting heart sounds from a patient'"'"'s chest cavity and providing a heart sound signal representative of said detected heart sounds, said transducer comprising;

a double cantilever beam configuration, the unsupported end of each beam having an equal mass attached thereto;

a bending moment sensor affixed to one side of each beam which produces an electrical output signal in relation to its bending moment;

support means for supporting said beam;

a contact plate affixed to said support means, for mounting against a patient'"'"'s chest such that chest vibrations couple through said contact plate and support means causing said beam to bend in accordance with such vibrations; and

the transducer having a resonant frequency slightly about 1200 Hz, and increased sensitivity in the frequency band of 300 to 1200 Hz;

frequency analyzer means for parametrically modeling said detected heart sounds based on said heart sound signal;

turbulent flow detector means for detecting the presence of an auditory component in said parametric model, said component being associated with turbulent blood flow in a partially occluded coronary artery, said detector means applying a decision criterion based on one or more parameters of said parametric model;

display means for presenting diagnostic information based on the detected presence of said auditory component.

16. A method for noninvasive detection of coronary artery disease, said method comprising the steps of:

detecting heart sounds from a patient'"'"'s chest cavity;

providing a parametric model for parametrically modeling said detected heart sounds;

detecting the presence of an auditory component in said parametric model by applying a decision criteria based on one or more parameters of the parametric model being indicative of said component being in the range of about 300 to about 1200 Hz, said component being associated with turbulent blood flow in a partially occluded coronary artery; and

presenting diagnostic information based on the detected presence of said auditory component.

17. The method of claim 16 which further comprises the step of isolating the detected heart sounds which occur within a selected segment of the cardiac cycle.

18. The method of claim 17 wherein said isolating step comprises the steps of delaying from a pulse signal for a predetermined amount of time, to a time just after the last of the prominent heart sounds, and isolating a predetermined length of time of the detected heart sounds succeeding said delay.

19. The method of claim 18 wherein said delay from the said pulse signal varies between 100 to 300 msec., depending on the heart rate, and wherein said predetermined length of time comprises a time window of approximately 256 msec duration.

20. The method of claim 17 wherein said selected segment of the cardiac cycle comprises a diastolic segment of the cardiac cycle.

21. The method of claim 20 wherein said step of isolating said diastolic segment comprises the steps of:

establishing a preliminary data window of a predetermined width upon an indication of the onset of diastole in said cardiac cycle, said preliminary data window comprising three zones, the first of which (zone

     1) includes the expected trailing edge of a second heart sound, the third of which (zone

     3) includes the expected leading edge of a succeeding first heart sound, and the second of which (zone

     2) is approximately centered in the diastolic segment of said heart sound;

establishing a threshold level within said preliminary data window for discriminating the occurrence of said second or first heart sound;

establishing a test window within said preliminary data window, said test window width being significantly smaller than said preliminary data window width;

evaluating a RMS value of said heart sound signal within said test window for a plurality of test window locations within said zones of the preliminary data window;

locating a lower diastolic boundary which is defined by the latest in time location of said test window in zone 1 where said RMS value exceeds the value of said threshold level;

locating an upper diastolic boundary which is defined by the earliest location in time of said test window in zone 3 where said RMS value exceeds said threshold level;

positioning a diastolic window of a predetermined width at the approximate center between said lower and upper diastolic boundaries, said diastolic window thereby isolating the heart sound signal within for subsequent analysis; and

rejecting said heart sound signal data within said diastolic window if a window defined by said lower and upper diastolic boundaries is smaller in width than said predetermined width of the diastolic window.

22. The method of claim 16 wherein said step of parametric modeling comprises Autoregressive (AR) modeling methods.

23. The method of claim 16 wherein said step of parametric modeling comprises Eigenvector modeling methods.

24. The method of claim 16 wherein said step of parametric modeling comprises Autoregressive Moving Average modeling methods.

25. The method of claim 16 wherein said step of applying a decision criterion comprises the step of evaluating the power spectral density function of said detected heart sounds model.

26. The method of claim 16 wherein said step of applying a decision criterion comprises the step of evaluating the location of the zeros of the linear prediction coefficients of said detected heart sounds model.

27. The method of claim 16 wherein said step of applying a decision criterion comprises the step of evaluating the peak power levels of said detected heart sounds model.

28. The method of claim 16 wherein said step of applying a decision criterion comprises the step of evaluating the reflection coefficients of said detected heart sounds model.

29. A method for automatically isolating the diastolic segment of a heart sound signal and editing the selected segment to determine if it is substantially free of artifacts resulting from other body sounds, said method comprising the steps of:

establishing a preliminary data window of a predetermined width upon an indication of the onset of diastole in a cardiac cycle, said preliminary data window comprising three zones, the first of which (zone

     1) includes the expected trailing edge of a second heart sound, the third of which (zone

     3) includes the expected leading edge of a succeeding first heart sound, and the second of which (zone

     2) is approximately centered in the diastolic segment of said heart sound;

establishing a threshold level within said preliminary data window for discriminating the occurrence of said second or first heart sound;

establishing a test window within said preliminary data window, said test window width being significantly smaller than said preliminary data window width;

evaluating a RMS value of said heart sound signal within said test window for a plurality of test window locations within said zones of the preliminary data window;

locating a lower diastolic boundary which is defined by the latest in time location of said test window in zone 1 where said RMS value exceeds the value of said threshold level;

locating an upper diastolic boundary which is defined by the earliest location in time of said test window in zone 3 where said RMS value exceeds said threshold level;

positioning a diastolic window of a predetermined width at the approximate center between said lower and upper diastolic boundaries, said diastolic window thereby isolating the heart sound signal within for subsequent analysis; and

rejecting said heart sound signal data within said diastolic window if a window defined by said lower and upper diastolic boundaries is smaller in width than said predetermined width of the diastolic window.

30. The method of claim 29, said method further comprising the step of:

rejecting said heart sound signal data within said diastolic window if any of said RMS values which exceed said threshold level are located in zone 2.

31. The method of claim 29, said method further comprising the step that said heart sound signal data within said diastolic window is rejected if within said zone 2 a signal level of said heart sound signal equals or exceeds said threshold level.

32. The method of claim 29, said method further comprising the step that said heart sound signal data is rejected if any RMS values within said diastolic window indicate microphone saturation.

33. The method of claim 29, said method further comprising the step that said heart sound signal data is rejected if an F-distribution test on the data within said record fails, said F-distribution test comprising the steps of evaluating the data based on predetermined degrees of freedom and a predetermined significance level.