Method and system for analyzing respiratory tract sounds

[0001] This invention relates to medical devices and methods, and more particularly to such devices and methods for analyzing body sounds.

[0002] Body sounds are routinely used by physicians in the diagnosis of various disorders. A physician may place a stethoscope on a person's chest or back and monitor the patient's breathing in order to detect adventitious (i.e. abnormal or unexpected) lung sounds. The identification and classification of adventitious lung sounds often provides important information about pulmonary abnormalities.

[0003] It is also known to fix one or more microphones onto a subject's chest or back and to record lung sounds. U.S. Pat. No. 6,139,505 discloses a system in which a plurality of microphones are placed around a patient's chest. The recordings of the microphones during inhalation and expiration are displayed on a screen, or printed on paper. The recordings are then visually examined by a physician in order to detect a pulmonary disorder in the patent. Kompis et al. (Chest, 120(4), 2001) disclose a system in which M microphones are placed on a patient's chest, and lung sounds are recorded. The recordings generate M linear equations that are solved using a least-squares fit. The solution of the system is used to determine the location in the lungs of the source of a sound detected in the recordings.

[0004] In the following description and set of claims, two explicitly described, calculable, or measurable variables are considered equivalent to each other when the two variables are proportional to one another.

[0005] The present invention provides, in one of its embodiments, a system and method for recording and analyzing respiratory tract sounds produced in the respiratory tract. The system includes a plurality of N transducers (microphones) configured to be attached to an essentially planar region R of the individual's back or chest over the individual's thorax. Positions in the region R are indicated by two-dimensional position vectors x(x¹,x²) in a two-dimensional coordinate system defined in the planar region R. The ith transducer, for i1 to N, is fixed at a position x_i in the region R and generates a signal, denoted herein by P(x_i,t), indicative of pressure waves in the body arriving at x_i.

[0006] The transducers are typically embedded in a matrix that permits to affix them easily onto the individual's skin. Such a matrix may typically be in the form of a vest or garment for easily placing over the individual's thorax. As may be appreciated, different matrices may be used for differently sized individuals, for different ages, sexes, etc.

[0007] The N signals P(x_i, t) are processed by signal processing circuitry. In accordance with the invention, the processing involves determining from the N signals an average acoustic energy, denoted herein by {tilde over (P)}(x,t₁,t₂), at least one position x in the region R over a time interval from t₁ to t₂. The term acoustic energy at a location is used herein to refer to a parameter indicative of or approximating the product of the pressure and the mass propagation velocity at that location.

[0008] In one embodiment, an average acoustic energy over a time interval from t₁ to t₂ is obtained at a position of one of the microphones using the algebraic expression $\begin{array}{cc}\tilde{P}\left(x_i,t_1,t_2\right)={}_{t_1}^{t_2}{P}^2\left(x_i,t\right)t& \left(1\right)\end{array}$

[0009] where x_i is the position of the microphone.

[0010] In a more preferred embodiment, an average acoustic energy {tilde over (P)}(x_i, t_i, t₂) over a time interval from t₁ to t₂ is obtained at a plurality of positions x_i of the microphones, for example using Equation (1), and then calculating {tilde over (P)}(x, t₁,t₂) at other locations x by interpolation of the {tilde over (P)}(x_i,t₁,t₂) using any known interpolation method.

[0011] In a most preferred embodiment, the interpolation is performed to obtain an average acoustic energy {tilde over (P)}(x,t₁,t₂) at a position x(x¹,x²) in the surface R using the algebraic expression: $\begin{array}{cc}\tilde{P}\left(x,t_1,t_2\right)=\underset{i=1}{\overset{N}{}}\tilde{P}\left(x_i,t_1,t_2\right)g\left(x,x_i,\right)& \left(2\right)\end{array}$

[0012] where g(x,x_i,) is a kernel satisfying $\begin{array}{cc}{}^{2}g=\frac{g}{}& \left(3\right)\\ \underset{i=1}{\overset{N}{}}g\left(x,x_i,\right)\mathrm{is}\mathrm{approximately}\mathrm{equal}\mathrm{to}1& \left(4\right)\end{array}$

[0013] and where x_i(x_i¹,x_i²) is the position of the ith microphone and is a selectable parameter.

[0014] For example, the kernel $\begin{array}{cc}g\left(x,x_i,\right)=\mathrm{Exp}-\left(\frac{{\left(x^1-x_i^1\sqrt{}\right)}^2}{2}\right)\mathrm{Exp}-\left(\frac{{\left(x^2-x_i^2\sqrt{}\right)}^2}{2}\right)& \left(5\right)\end{array}$

[0015] may be used.

[0016] The system may optionally contain a display device for displaying the function {tilde over (P)}. The function {tilde over (P)} may be displayed on the display, for example using a gray level scale, as demonstrated in the examples below. A two dimensional graphical representation of the function {tilde over (P)} produces an image of the body region that may be analyzed for the detection of a disorder in the body region similar to the analysis of images obtained by other imaging methods such as X-ray or ultrasound imaging.

[0017] A region or regions in a displayed image that are suspected of including a pathological condition, may de identified in the image, and this may be in a number of ways, for example, by different colors, by different patterns, by way of a written text, and many other ways. The term pathological condition refers to any deviation from the normal, healthy condition of the respiratory tract. This includes infection, inflammation, tumor, pleural effusion, pneumonia, narrowing of the airways, and other space containing lesions in the respiratory tract, etc.

[0018] Additionally, a time interval can be divided into a plurality of sub intervals, and an average acoustic energy {tilde over (P)} determined over the region R for two or more of the sub intervals. An image of {tilde over (P)} for each of these sub intervals may then be determined and displayed sequentially on the display device. This generates a movie showing dynamic changes occurring in the acoustic energy in the body region, over the time interval. For example, transducers may be placed on a person's chest and an average acoustic energy {tilde over (P)} determined in accordance with the invention for a plurality of sub intervals over a breathing cycle. An image can be obtained for each of these sub intervals and displayed sequentially so as to generate a movie showing changes in the acoustic energy of the lungs over the breathing cycle.

[0019] The signals P(x_i,t) may also be subjected to band pass filtering before being analyzed by the method of the invention, so that an average acoustic energy is produced for one or more frequency bands of interest. The functions may be superimposed on the display device by representing each average acoustic energy function with a different color. Since respiratory sounds known as wheezes and crackles have different characteristic frequency ranges, band pass filtering can be used to identify these respiratory sounds. A region or regions in a displayed image of wheezes or crackles may be identified in the image, for example, by a characteristic color, pattern, by way of a written text.

[0020] The present invention thus provides a system for analyzing sounds in at least portion of an individual's respiratory tract comprising:

[0021] (a) a plurality of N transducers, each transducer configured to be fixed on a surface of the individual over the thorax, the ith transducer being fixed at a location x_i and generating a signal P(x_i, t) indicative of pressure waves at the location x_i; for i1 to N; and

[0022] (Two) a processor configured to receive the signals P(x_i,t) and determine an average acoustic energy {tilde over (P)}(x,t₁,t₂) at at least one position x over a time interval from a first time t₁ to a second time t₂, {tilde over (P)} being determined in an algorithm involving at least one of the signals P(x_i, t).

[0023] The present invention further provides a method for analyzing sounds in at least a portion of an individual's thorax, comprising:

[0024] (One) obtaining N signals P(x_i,t) for i1 to N, the signal P(x_i,t) being indicative of pressure waves at the location x_i; on a surface of the body over the thorax;

[0025] (Two) determining an average acoustic energy {tilde over (P)}(x,t₁,t₂) at at least one position x over a time interval from a first time t₁ to a second time t₂, {tilde over (P)} determined in an algorithm involving at least one of the signals.

[0026] The present invention also provides a program storage device readable by machine, tangibly embodying a program of instructions executable by the machine to perform method steps for determining for at least one time interval, an average acoustic energy function {tilde over (P)} using an algorithm involving at least one signal P(xi, t) indicative of pressure waves at a location x_i on a body surface.

[0027] The present invention still further provides a computer program product comprising a computer useable medium having computer readable program code embodied therein analyzing sounds in at least a portion of an individual's body, the computer program product comprising:

[0028] computer readable program code for causing the computer to determine, for at least one time interval, an acoustic energy function {tilde over (P)}, {tilde over (P)} being determined in algorithm involving at least one signal P(xi,t) indicative of pressure waves at a location x_i on a body surface.

[0029] In order to understand the invention and to see how it may be carried out in practice, a preferred embodiment will now be described, by way of non-limiting example only, with reference to the accompanying drawings, in which:

[0030] FIG. 1 shows a system for obtaining an analyzing body sound in accordance with one embodiment of the invention;

[0031] FIG. 2 shows a flow chart for carrying out a method of obtaining analyzing body sounds in accordance with one embodiment of the invention;

[0032] FIG. 3 shows recording and analysis of signals over an inspiratory phase of a respiratory cycle; and

[0033] FIG. 4 shows recording and analysis of signals over an expiratory phase of a respiratory cycle.

[0034] FIG. 5a shows an image obtained on a healthy individual in accordance with the invention, and FIG. 5b shows a chest X-ray of the same individual;

[0035] FIG. 6 shows successive frames from a movie of the respiratory tract of a healthy individual;

[0036] FIG. 7a shows an image obtained on an individual with pleural effusion in accordance with the invention, and FIG. 7b shows a chest X-ray of the same individual;

[0037] FIG. 8 shows successive frames from a movie of the respiratory tract of an individual with pleural effusion;

[0038] FIG. 9a shows an image obtained on an individual with pneumonia in accordance with the invention, and FIG. 9b shows a chest X-ray of the same individual; and

[0039] FIG. 10 shows successive frames from a movie of the respiratory tract of an individual with pleural effusion.

[0040] FIG. 1 shows a system generally indicated by 100 for analyzing body sounds in a three-dimensional region of an individual's body in accordance with one embodiment of the invention. A plurality of N sound transducers 105, of which four are shown, are applied to a planar region of the chest or back skin of individual 110. The transducers 105 may be applied to the subject by any means known in the art, for example using an adhesive, suction, or fastening straps. Each transducer 105 produces an analog voltage signal 115 indicative of pressure waves arriving to the transducer. The analog signals 115 are digitized by a multichannel analog to digital converter 120. The digital data signals P(x_i,t) 125, represent the pressure wave at the location x_i of the ith transducer (i1 to N) at time t. The data signals 125 are input to a memory 130. Data input to the memory 130 are accessed by a processor 135 configured to process the data signals 125. The signals 125 may be denoised by filtering components having frequencies outside of the range of body sounds in the body region, for example, vibrations due to movement of the individual. Each signal 125 may also be subject to band pass filtering so that only components in the signal within a range of interest are analyzed.

[0041] An input device such as a computer keyboard 140 or mouse 145 is used to input relevant information relating to the examination such as personal details of the individual 110. The input device 140 may also be used to input values of the times t₁ and t₂. Alternatively, the times t₁ and t₂ may be determined automatically in a respiratory phase analysis of the signals P(x_i,t) performed by the processor 135. The processor 135 determines an average acoustic energy {tilde over (P)}(x,t₁,t₂) over the time interval from t₁ to t₂ at least one locations in the region R in a calculation involving at least one of the signals P(x_i,t).

[0042] The average acoustic energies are stored in the memory 130 and may be displayed on a display device 150 such as a CRT screen for diagnosis by a physician.

[0043] The processor 135 may also perform an automatic differential diagnosis by comparing the function {tilde over (P)} to functions stored in the memory and known to be indicative of various disorders in the body region.

[0044] FIG. 2 shows a flow chart diagram for carrying out the method of the invention in accordance with one embodiment. In step 200 the signals P(x_i,t) are obtained from N transducers placed at predetermined locations x_i for i from 1 to N in a region R on the body surface. In step 205 values of t₁ and t₂ are either input to the processor 135 using the input devices 140 or 145, or are determined by the processor. In step 210, an average acoustic energy {tilde over (P)}(x,t₁,t₂) is determined at least one location x in the region R over the time interval t₁ to t₂. In step 220 the average acoustic energy is displayed on the display 150 for at least one value of x. In step 230, it is determined whether a function {tilde over (P)} is to be determined over another time interval. If yes, the process returns to step 205. If not, the process terminates.

[0045] It will also be understood that the system according to the invention may be a suitably programmed computer. Likewise, the invention contemplates a computer program being readable by a computer for executing the method of the invention. The invention further contemplates a machine-readable memory tangibly embodying a program of instructions executable by the machine for executing the method of the invention.

[0046] The system and method of the invention were used to analyze lower respiratory tract sounds in an individual.

[0047] FIG. 3 shows recording and analysis of signals over an inspiratory phase of a respiratory cycle in an individual. A two-dimensional coordinate system was defined on the individual's back. As shown in FIG. 3a, 48 transducers were placed on the individual's back over the lungs at the locations indicated by the circles 300. The curves 305 show the presumed contours of the lungs. As can be seen, the transducers were arranged in a regular orthogonal lattice with a spacing between the transducers in the horizontal and vertical directions of 5 cm. The signals P(x_i,t) were then recorded over one inspiratory phase of a breathing cycle (t₁ and t₂ are the beginning and end respectively of the inspiratory phase). Each signal was filtered using a low-pass filter having a cut-off of 150 Hz. The average value of each filtered function P(x_i,t) over the inspiratory phase is indicated in FIG. 3a by means of gray level shading of each circle 300 with reference to the gray level scale 310. {tilde over (P)}(x,t₁,t₂) was obtained using Equations (1) and (2) above with the kernel g of Equation (5) with 36 pixels. FIG. 3b shows a 512 pixel512 pixel graphical representation of the function {tilde over (P)}(x,t₁,t₂) over the inspiratory phase also in reference to the gray level scale 310. In the graphical representation of the function {tilde over (P)}(x,t₁,t₂) shown in FIG. 3b, the contours of the lungs and heart are easily discernable.

[0048] FIG. 4 shows recording and analysis of signals over an expiratory phase of a respiratory cycle. As shown in FIG. 4a, 48 transducers were placed on an individual's back at the same locations x_i used in FIG. 3, as indicated by the circles 400. The curves 405 show the presumed contours of the individual's lungs. The signals P(x_i,t) were then recorded over one expiratory phase of a breathing cycle (t₁ and t₂ are the beginning and end respectively of the expiratory phase). Each signal was filtered using a low-pass filter having a cut-off of 150 Hz. The average value of each function P(x_i, t) over the expiratory phase is indicated in FIG. 4a by means of gray level shading of each circle 400 with reference to the gray level scale 410. {tilde over (P)}(x, t₁,t₂) was obtained using Equations (1) and (2) above. FIG. 3b shows the function {tilde over (P)}(x,t₁,t₂) over the expiratory phase also in reference to the gray level scale 410. Comparison of FIGS. 3b and 4b shows the volume change in acoustic energy between the inspiratory and expiratory phase of the respiratory cycle.

[0049] Movies showing changes in the lungs during a respiratory cycle were obtained using the method and system of the invention. Signals 125 were obtained and divided into time segments of 0.5 sec duration. Each segment was analyzed by the method of the invention and an image was generated. The images were displayed on the display device 150 in rapid succession so as to produce a movie of the respiratory tract over the respiratory cycle.

[0050] FIG. 5a shows an image of the respiratory tract of a healthy individual obtained over an entire respiratory cycle in accordance with the invention, and FIG. 5b shows a chest X-ray of the same individual. FIG. 6 shows 11 successive images obtained over successive 0.4 sec time intervals during a respiratory cycle of the individual. Each frame represents the processing of the recorded signals over a time interval of 0.4 sec. Frames 01 to 05 (obtained at times 0.0 to 1.6 sec) were obtained during the inspiratory phase of the respiratory cycle, while frames 06 to 11 (obtained at times 1.8 to 3.6 sec) were obtained during the expiratory phase. The sequence of images shown in FIG. 6 can be displayed in succession on a display device so as to create a movie of the respiratory tract over a respiratory cycle. The sequence of images shown in FIG. 6 shows complete filling and emptying of the lungs during the respiratory cycle, as would be expected in a healthy individual not having any space-filling lesions.

[0051] FIG. 7a shows an image of the respiratory tract of an individual with pleural effusion obtained over an entire respiratory cycle in accordance with the invention, and FIG. 7b shows a chest X-ray of the same individual. FIG. 8 shows 16 successive images obtained over successive 0.4 sec time intervals during a respiratory cycle of the individual. Each frame represents the processing of the recorded signals over a time interval of 0.4 sec. Frames 01 to 06 (obtained at times 0.0 to 2.0 sec) were obtained during the inspiratory phase of the respiratory cycle, while frames 07 to 16 (obtained at times 2.4 to 4.0 sec) were obtained during the expiratory phase. The sequence of images shown in FIG. 8 can be displayed in succession on a display device so as to create a movie of the respiratory tract over a respiratory cycle. In the sequence of images shown in FIG. 8, lung tissue in the lower right lobe is not visualized indicating the absence of air flow in the lower right lung, as would be expected in an individual having a space-filling lesion as occurs in pleural effusion. Airflow in the upper portion of the right lung is also observed to be impaired.

[0052] FIG. 9a shows an image of the respiratory tract of an individual with pneumonia obtained over an entire respiratory cycle in accordance with the invention, and FIG. 9b shows a chest X-ray of the same individual. FIG. 10 shows 12 successive images obtained over successive 0.4 sec time intervals during a respiratory cycle of the individual. The sequence of images shown in FIG. 10 can be displayed in succession on a display device so as to create a movie of the respiratory tract over a respiratory cycle. Each frame represents the processing of the recorded signals over a time interval of 0.4 sec. Frames 01 to 06 (obtained at times 0.0 to 2.0 sec) were obtained during the inspiratory phase of the respiratory cycle, while frames 07 to 16 (obtained at times 2.4 to 4.0 sec) were obtained during the expiratory phase. In the sequence of images shown in FIG. 10, lung tissue in the lower left lobe is not visualized indicating the absence of air flow in the lower left lung, as would be expected in an individual having a space-filling lesion as occurs in pneumonia. Airflow in the upper portion of the left lung is observed to be normal.