Search

EP-4736779-A1 - HEMODYNAMIC MONITORING SYSTEM

EP4736779A1EP 4736779 A1EP4736779 A1EP 4736779A1EP-4736779-A1

Abstract

A method for evaluating changes in patient hemodynamic status, e.g. hemodynamic responsiveness, between before and after clinical intervention events. Waveform features extracted from an ensemble average blood velocity pulse waveform for the patient are fed as input to an analysis algorithm, for example comprising a classifier, to generate an output indicative of one or more hemodynamic characteristics, for example indicative of hemodynamic responsiveness. This can be used to evaluate changes in hemodynamic responsiveness before and after a clinical intervention event.

Inventors

  • LAU, Kevin Daniel Seng Hung
  • JOSHI, ROHAN
  • DINIS FERNANDES-KUILBOER, Catarina

Assignees

  • Koninklijke Philips N.V.

Dates

Publication Date
20260506
Application Date
20241105

Claims (15)

  1. A computer-implemented method, comprising: - receiving (12) first ultrasound data of a blood vessel of a subject, the first ultrasound data acquired over a first epoch, the first epoch occurring prior to a clinical intervention event; - receiving (14) second ultrasound data of a blood vessel of a subject, the second ultrasound data acquired over a second epoch, the second epoch occurring following the clinical intervention event; - deriving (16) from the received first ultrasound data a first blood velocity-time signal representative of blood velocity over a time window contained in the first epoch, and computing (20) a first ensemble average blood velocity waveform from a plurality of blood velocity waveforms present in the first blood velocity-time signal; - deriving (18) from the second ultrasound data a second blood velocity-time signal representative of blood velocity over a time window contained in the second epoch, and computing (22) a second ensemble average blood velocity waveform from a plurality of blood velocity waveforms present in the second blood velocity-time signal; - determining (24) a first set of parameters, the determining the first set of parameters including extracting from the first ensemble average blood velocity waveform one or more pre-defined blood velocity waveform features; - determining (26) a second set of parameters, the determining the second set of parameters including extracting from the second ensemble average blood velocity waveform the one or more pre-defined blood velocity waveform features; - retrieving and applying a pre-defined analysis algorithm, the analysis algorithm configured to receive said first and second sets of parameters as inputs and to generate (28) a hemodynamic response indicator as an output, the hemodynamic response indicator indicative of characteristics of a patient's hemodynamic responsiveness to the clinical intervention event.
  2. The method of claim 1, wherein the analysis algorithm comprises a classifier algorithm and wherein the hemodynamic response indicator output from the analysis algorithm is a classification of patient hemodynamic responsiveness into one of a pre-defined set of possible classifications.
  3. The method of claim 1 or claim 2, wherein the analysis algorithm comprises a machine learning algorithm.
  4. The method of claim 1, further comprising: - deriving from the first ultrasound data a first arterial diameter signal representative of a diameter of the blood vessel over said time window contained in the first epoch; - deriving from the second ultrasound data a second arterial diameter signal representative of a diameter of the blood vessel over said time window contained in the second epoch; and - wherein the computing the first set of parameters further comprises extracting from the first arterial diameter signal one or more pre-defined arterial diameter waveform features; - wherein the determining the second set of parameters further comprises extracting from the second arterial diameter signal the one or more pre-defined arterial diameter waveform features.
  5. The method of any preceding claim, wherein the clinical intervention event is one or more of: end-expiratory occlusion pressure, inspiratory hold positive end-expiratory pressure 10cm H2O, inspiratory hold positive end-expiratory pressure 15cm H2O, a passive leg raise, and intra-arterial administration of a bolus of liquid to the subject.
  6. The method of any preceding claim, wherein each of the first ultrasound data and second ultrasound data comprise both B-mode ultrasound data and pulsed wave Doppler ultrasound data, and preferably wherein the blood velocity-time signal is derived using the pulsed wave Doppler data.
  7. The method of any of claims 1-6, further comprising obtaining first and second blood pressure measurement information for the subject corresponding to the first and second epochs and wherein the analysis algorithm is further configured to receive the first and second blood pressure measurement information as inputs.
  8. The method of claim 7, wherein the obtaining the first and second blood pressure measurement information comprises obtaining first and second blood pressure signals for the subject indicative of a blood pressure waveform for the subject over the time window contained in the first and second epochs respectively.
  9. The method of claim 8, wherein the method comprises computing a first ensemble average blood pressure waveform from a plurality of blood pressure waveforms present in the first blood pressure signal and further computing a second ensemble average blood pressure waveform from a plurality of blood pressure waveforms present in the second blood pressure signal; and wherein the first and second blood pressure measurement information is derived from the first and second ensemble average blood pressure waveforms.
  10. The method of claim 9, wherein the determining the first set of parameters includes extracting from the first ensemble average blood pressure waveform one or more pre-defined blood pressure waveform features, and wherein the determining the second set of parameters includes computing from the second ensemble average blood pressure waveform the same one or more pre-defined blood pressure waveform features; wherein the blood pressure waveform features extracted from the first and second ensemble average blood pressure waveforms provide the blood pressure measurement information.
  11. The method of any of claims 1-10, further comprising: - deriving from the first ultrasound data a first arterial diameter signal representative of a diameter of the blood vessel over said time window contained in the first epoch, and computing a first ensemble average arterial diameter waveform from a plurality of arterial diameter waveforms present in the first arterial diameter signal; - deriving from the second ultrasound data a second arterial diameter signal representative of a diameter of the blood vessel over said time window contained in the second epoch and computing a second ensemble average arterial diameter waveform from a plurality of arterial diameter waveforms present in the second arterial diameter signal; and - wherein the computing the first set of parameters comprises computing from the first ensemble average diameter waveform one or more pre-defined arterial diameter waveform features and computing from the second ensemble average diameter waveform the same one or more pre-defined arterial diameter waveform features.
  12. The method of any of claims 1-11, wherein each of the first and second set of parameters comprise a same group of waveform features which include at least one of: skewness of the ensemble average velocity waveform; kurtosis of the ensemble average velocity waveform; Full Width Half Maximum, FWHM, of the ensemble average velocity waveform; and an acceleration of the ensemble average velocity waveform indicative of a steepness of a leading edge of the ensemble average velocity waveform.
  13. A computer program product comprising computer code configured to cause a processor to perform a method in accordance with any of claims 1-12.
  14. A processing device comprising one or more processors configured to perform a method in accordance with any of claims 1-12.
  15. A system (30), comprising; the processing device (32) of claim 14; and an ultrasound acquisition apparatus (54) for acquiring the first and/or second ultrasound data.

Description

FIELD OF THE INVENTION The present invention pertains to the field of medical technology, specifically to systems and methods for monitoring a patient's hemodynamic status using ultrasound data. BACKGROUND OF THE INVENTION Hemodynamic monitoring is an important aspect of patient monitoring. Hemodynamics refers to the physical principles governing the circulation of blood in the cardiovascular system. It involves inter alia the assessment of blood volume, pressure, and the oxygenation level in the blood. Monitoring these parameters is of significance in the management of patients in intensive care units (ICUs) and during high-risk surgeries. One aspect of hemodynamic monitoring is the assessment of fluid responsiveness. Fluid responsiveness is a term used to describe the likelihood that a patient's cardiac output will increase in response to fluid administration. Accurate assessment of fluid responsiveness is of particular clinical relevance, especially in the management of patients with circulatory failure. A measure of fluid responsiveness can help guide clinicians in deciding whether to administer fluids to improve physiological stability and patient outcome. Monitoring of fluid responsiveness in the current state of the art involves invasive monitoring methods such as an arterial catheter. Non-invasive methods for monitoring fluid responsiveness, and hemodynamic status more generally, would be of value. SUMMARY OF THE INVENTION The invention is defined by the claims. In accordance with an aspect of the invention, a computer-implemented method is provided. The method includes receiving first and second ultrasound data of a blood vessel of a subject, the first ultrasound data acquired over a first epoch occurring prior to a clinical intervention event or patient interaction event, and the second ultrasound data acquired over a second epoch occurring following the clinical intervention event or patient interaction event. The method further includes deriving from the received first and second ultrasound data first and second blood velocity-time signals representative of blood velocity over a time window contained in the first and second epochs respectively. The method further includes computing first and second ensemble average blood velocity waveforms from a plurality of blood velocity waveforms present in the first and second blood velocity-time signals respectively. Each respective one of the plurality of blood velocity waveforms may correspond to a respective single respective heartbeat. The method further includes determining first and second sets of parameters by extracting from the first and second ensemble average blood velocity waveforms one or more pre-defined blood velocity waveform features. The method also includes retrieving and applying a pre-defined analysis algorithm configured to receive the first and second sets of parameters as inputs and to generate a hemodynamic response indicator as an output, the hemodynamic response indicator indicative of characteristics of a patient's hemodynamic responsiveness to the clinical intervention event. In the context of the present disclosure, a "clinical intervention event" may refer to any intervention performed in relation to a patient, for example a medical procedure or treatment, for example an intervention for modifying the patient's physiological status. The clinical intervention events could for example include any event that could potentially affect the patient's hemodynamic status, such as a change in medication, a surgical procedure, or a change in the patient's physical activity level. Examples of clinical intervention events include, but are not limited to, end-expiratory occlusion pressure (test), inspiratory hold positive end-expiratory pressure 10cm H2O, inspiratory hold positive end-expiratory pressure 15cm H2O, a passive leg raise, and intra-arterial administration of a bolus of liquid to the subject. Other examples of clinical intervention events include any kind of surgical procedure. In some embodiments, the method includes processing the blood velocity-time signal to detect and/or extract the said plurality of blood velocity waveforms from the signal. From these, the ensemble average blood velocity waveform can be computed, for example by aligning and then averaging the plurality of extracted blood velocity waveforms. Each blood velocity waveform corresponds to a heart beat and is a pulse in the blood velocity at the measurement site at which the relevant ultrasound data is acquired. The analysis algorithm used in the method can take various forms. For example, it could be a multiparametric regression function that uses the extracted features as input variables to predict the hemodynamic indicator. Alternatively, the analysis algorithm could be a trained machine learning algorithm that has been trained on a dataset of previous patient data to predict the patient's hemodynamic response based on the extracted features. In some embodim