Method and system to assess disease using multi-sensor signals
The exemplified methods and systems (e.g., machine-learned systems) facilitate the acquisition of ballistocardiographic signals and the determination and use of ballistocardiographic signal related features, or parameters, in a model or classifier to estimate metrics associated with the physiological state of a subject, including for the presence or non-presence of a disease, medical condition, or indication of either. The estimated metric may be used to assist a physician or other healthcare provider in diagnosing the presence or non-presence and/or severity and/or localization of diseases, medical condition, or indication of either or in the treatment of said diseases or indicating conditions. In some embodiments, certain ballistocardiographic signals can also be used to remove motion artifacts from biophysical signals used for the estimation.
1 . A method to remove motion-associated artifacts from an acquired measurement signal, wherein the acquired measurement signal is used to non-invasively assess a cardiac disease state or abnormal cardiac condition of a subject, the method comprising:
obtaining, by one or more processors, a first biophysical signal data set of the subject comprising a first photoplethysmographic signal and a second photoplethysmographic signal-or a cardiac signal;
obtaining, by the one or more processors, a second biophysical signal data set of the subject associated with a ballistocardiogram signal, wherein the ballistocardiogram signal are temporally and spatially acquired with respect to the first photoplethysmographic signal, the second photoplethysmographic signal, or the cardiac signal;
determining, by the one or more processors, a filtered biophysical-signal data set of the first biophysical signal data set by removing an estimated motion signal determined using the ballistocardiogram signal;
determining, by the one or more processors, one or more dynamical features including a first dynamical feature and a second dynamical feature, wherein the first and second dynamical features each characterize, via a statistical-or dynamical-analysis assessment, one or more dynamical properties across multiple heart cycles of the second biophysical signal data set associated with the ballistocardiogram signal;
determining, by the one or more processors, via a trained classifier model and based, at least in part on the determined one or more dynamical features, an estimated value related to a presence or non-presence of the cardiac disease state or abnormal condition; and
in response to determining the presence of the cardiac disease state or abnormal condition, outputting, by the one or more processors and via a report and/or display, a recommended treatment for the cardiac disease state or abnormal condition, wherein the treatment is selected based on the estimated value determined via the trained classifier model.
2 . The method of claim 1 , wherein the estimated motion signal comprises an estimated gross movement of the subject.
3 . The method of claim 2 , further comprising:
determining, by the one or more processors, estimated gross subject's movements during an acquisition of the first biophysical signal data set, wherein the estimated gross subject's movement is used to reject a portion of the acquired first biophysical signal data set having the estimated gross subject's movements above a pre-defined threshold value.
4 . The method of claim 1 , wherein the estimated motion signal comprises assessed vibrations associated with heartbeats of the subject.
5 . The method of claim 4 , wherein the estimated motion signal is determined by generating a template-signal vector data set characteristic of a representative motion signal pattern of the ballistocardiogram signal.
6 . The method of claim 1 , wherein the estimated motion signal is used to assess a change in cell potential of respective electrodes used to acquire the cardiac signal.
7 . The method of claim 1 , wherein a subsequent analysis to determine the estimated value for the presence of the cardiac disease state or abnormal cardiac condition comprises:
determining, by the one or more processor, one or more synchronicity dynamical features including a first synchronicity feature and a second synchronicity feature, wherein the first and second synchronicity features each characterize, via the statistical- or dynamical-analysis assessment, the one or more synchronicity dynamical features across multiple heart cycles between (i) the first biophysical signal data set associated with the first and second photoplethysmographic signals or cardiac signal and (ii) the second biophysical signal data set associated with the ballistocardiogram signal.
8 . The method of claim 7 , wherein the statistical-or dynamical- analysis assessment is selected from the group consisting of:
a statistical- or dynamical-analysis assessment of values of the ballistocardiogram signal at a registration point defined by one or both the first photoplethysmographic signal and the second photoplethysmographic signal;
a statistical- or dynamical-analysis assessment of values of the ballistocardiogram signal at a registration point defined by the cardiac signal;
a statistical- or dynamical-analysis assessment of values of one of the first photoplethysmographic signal or the second photoplethysmographic signal at a landmark defined in the ballistocardiogram signal;
a statistical- or dynamical-analysis assessment of values of the cardiac signal at a landmark defined in the ballistocardiogram signal;
a statistical- or dynamical-analysis assessment of time intervals between (a) a first set of landmarks defined in the ballistocardiogram signal and (b) a second set of landmarks defined in the cardiac signal;
a statistical- or dynamical-analysis assessment of time intervals between (a) a first set of landmarks defined between the first photoplethysmographic signal and the second photoplethysmographic signal and (b) a second set of landmarks defined in the ballistocardiogram signal;
a statistical- or dynamical-analysis assessment of phase relations between (i) periods of the ballistocardiogram signal and (ii) periods of the cardiac signal; and
a statistical- or dynamical-analysis assessment of phase relations between (i) periods of one of the first or second photoplethysmographic signals and (ii) periods of the ballistocardiogram signal.
9 . The method of claim 1 , further comprising:
determining, by the one or more processors, a Poincaré map of the obtained biophysical signal data set;
determining, by the one or more processors, a geometric shape object of the Poincaré map; and
determining, by the one or more processors, one or more geometric properties of the geometric shape object, wherein the one or more determined geometric properties are used in the determination of the estimated value for the presence, non-presence, localization, and/or severity of the cardiac disease or condition.
10 . The method of claim 1 , wherein the ballistocardiogram signal is acquired via an accelerometer co-located on a sensor associated with the cardiac signal.
11 . The method of claim 1 , wherein the ballistocardiogram signal is acquired via an acoustic sensor co-located on a sensor associated with the cardiac signal.
12 . The method of claim 1 , wherein the ballistocardiogram signal is acquired via a pressure sensor co-located on a sensor associated with the cardiac signal.
13 . The method of claim 1 , further comprising:
obtaining, by the one or more processors, a third biophysical signal data set of the subject associated with a second ballistocardiogram signal, wherein the second ballistocardiogram signal are temporally and spatially acquired with respect to cardiac signal.
14 . The method of claim 1 , wherein the determined estimated value for the presence of the cardiac disease state or abnormal condition comprises an assessed indication or estimate of at least one of presence, non-presence, and severity of elevated or abnormal left ventricular end-diastolic pressure (LVEDP).
15 . The method of claim 1 , wherein the cardiac disease state or condition is selected from the group consisting of coronary artery disease, pulmonary hypertension, pulmonary arterial hypertension, pulmonary hypertension due to left heart disease, rare disorders that lead to pulmonary hypertension, left ventricular heart failure or left-sided heart failure, right ventricular heart failure or right-sided heart failure, systolic heart failure, diastolic heart failure, ischemic heart disease, and arrhythmia.
16 . The method of claim 1 , wherein determining the estimated value for the presence or non-presence of the cardiac disease state or abnormal cardiac condition comprises determining a ballistocardiographic feature or parameter that includes a quantification of dynamical characteristic including at least one of Lyapunov exponent, correlation dimension, entropy, mutual information, and correlation in relation to another biophysical signal.
17 . The method of claim 1 , wherein determining the estimated value for the presence or non-presence of the cardiac disease state or abnormal cardiac condition comprises determining a ballistocardiographic feature or parameter that includes a quantification of linear characteristic including at least one of peak amplitudes, peak-to-peak distances, and angles between registrations points in the ballistocardiogram signal.
18 . The method of claim 1 , wherein determining the estimated value for the presence or non-presence of the cardiac disease state or abnormal cardiac condition comprises determining a ballistocardiographic feature or parameter that includes a quantification of linear characteristic comprising 3D vector among the ballistocardiogram signal, a first derivative of the ballistocardiogram signal, and a second derivative of the ballistocardiogram signal.
19 . The method of claim 1 , wherein determining the estimated value for the presence or non-presence of the cardiac disease state or abnormal cardiac condition comprises determining a ballistocardiographic feature or parameter that includes a quantification of at least one of a power spectrum, frequency content, or coherence analysis of the ballistocardiogram signal.
20 . The method of claim 1 , wherein the ballistocardiogram signal is acquired from a plurality of multi-axis accelerometers, wherein each of the plurality of multi-axis accelerometers is co-located to a respective first sensor of a plurality of first sensors.
21 . The method of claim 20 , wherein the plurality of multi-axis accelerometers and respective first sensor are integrated into a snap lead.
22 . The method of claim 21 , wherein the snap lead includes a wireless communication transceiver.
23 . The method of claim 22 , wherein the wireless communication transceiver is configured for ultra-wide-band operation.
24 . A system comprising:
a processor; and
a memory having instructions stored thereon, wherein the instructions, when executed by the processor, cause the processor to: .
obtain a first biophysical signal data set of a subject comprising a first photoplethysmographic signal and a second photoplethysmographic signal or a cardiac signal;
obtain a second biophysical signal data set of the subject associated with a ballistocardiogram signal, wherein the ballistocardiogram signal are temporally and spatially acquired with respect to the first photoplethysmographic signal, the second photoplethysmographic signal, or the cardiac signal;
determine a filtered biophysical-signal data set of the first biophysical signal data set by removing an estimated motion signal determined using the ballistocardiogram signal;
determine one or more dynamical features including a first dynamical feature and a second dynamical feature, wherein the first and second dynamical features each characterize, via a statistical- or dynamical-analysis assessment, one or more dynamical properties across multiple heart cycles of the second biophysical signal data set associated with the ballistocardiogram signal;
determine, via a trained classifier model and based, at least in part on the determined one or more dynamical features, an estimated value related to a presence or non-presence of a cardiac disease state or abnormal condition; and
in response to determining the presence of the cardiac disease state or abnormal condition, output, via a report and/or display, a recommended treatment for the cardiac disease state or abnormal condition, wherein the treatment is selected based on the estimated value determined via the trained classifier model.