Systems and methods for determining cardiac contractility based on signals from mechanical circulation support device

Abstract

Methods and apparatus are provided for estimating a measure of cardiac contractility based on a set of features determined from a set of signals associated with a mechanical loop support device. The method includes determining, using a computer processor, a set of features based at least in part on the set of signals; providing the set of features as inputs to a machine learning model, the machine learning model trained to output a measure of cardiac contractility; and performing an action based at least in part on the measure of cardiac contractility output by the machine learning model.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to techniques for determining cardiac contractility based on signals from mechanical circulatory support devices. BACKGROUND

[0002] Cardiovascular disease is a leading cause of morbidity, mortality, and global health care burden. Multiple treatment modalities have been developed for heart health ranging from medications to mechanical devices and transplantation. Mechanical heart support devices, such as heart pump systems, provide hemodynamic support and facilitate heart recovery. Some heart pump systems can be inserted percutaneously into the heart and can operate in parallel with the native heart to supplement cardiac output. Examples of such devices include the Impella® series of devices (Abiomed, Inc., Danvers, MA). Such heart pump systems can have sensors that detect blood pressure (or assess transmembrane differential pressure) and can monitor motor current, and can use the sensor data and motor current readings to help identify pump position, among other things.

[0003] Such pumps can be positioned in a heart chamber, such as the left ventricle, to assist the heart. In this case, the pump can be inserted via the femoral artery and introduced upward into the left ventricle of the patient's heart by means of a hollow catheter. From this position, the pump inlet can draw in blood and the pump outlet can expel blood into the aorta. In this way, the function of the heart can be replaced or at least assisted by the operation of the pump.

[0004] Heart pumps are typically connected to a respective heart pump controller that controls the heart pump (e.g., motor speed), and collects and displays operational data about the blood pump, such as heart signal levels, battery temperature, blood flow rate, and tubing integrity. An exemplary heart pump controller can be the Automated Impella Controller ® from Abiomed, Inc. In some cases, the controller can issue an alert when an operational data value is outside of a predetermined value or range, such as in the event of a detected leak, suction, and / or pump malfunction. The controller can include a video display screen on which a graphical user interface is displayed that is configured to display operational data and / or alerts. SUMMARY

[0005] Mechanical heart support devices, such as heart pump systems, provide hemodynamic support and facilitate the recovery of native heart function (e.g., native systolic function of the heart). As such, a measure of the current systolic function of a patient can be useful in facilitating clinical decision making regarding the care of the patient while using such a device. For example, as the native systolic function of a patient recovers, the patient can be weaned off a heart pump system. The present inventors have recognized and appreciated that a heart pump of a heart pump system is positioned in the heart of a patient in a manner that provides a unique set of signals that can be used to more accurately estimate the systolic function of the patient as compared to existing techniques for estimating systolic function that do not take into account such signals. To this end, some embodiments of the present disclosure relate to improved techniques for estimating systolic function using a machine learning model trained to output a measure of systolic function as a function of a set of features determined based at least in part on signals associated with a heart pump system.

[0006] In one aspect, a computer-implemented method is provided. The computer- implemented method includes receiving a set of signals from a mechanical circulatory support device; determining, using a computer processor, a set of features based at least in part on the set of signals; providing the set of features as input to a machine learning model trained to output a measure of systolic function; and performing an action based at least in part on the measure of systolic function output by the machine learning model.

[0007] In another aspect, the set of signals includes at least one pressure signal and / or at least one pump function signal. In another aspect, determining the set of features includes determining at least one feature using a combination of a first pressure signal of the at least one pressure signal and a first pump function signal of the at least one pump function signal. In another aspect, the at least one feature using a combination of a first pressure signal of the at least one pressure signal and a first pump function signal of the at least one pump function signal includes a maximum upstroke rate of left ventricular pressure during systole. In another aspect, determining a set of features includes determining one or more interbeat features based at least in part on the set of signals. In another aspect, determining the one or more interbeat features includes determining one or more of mean pump flow, mean aortic pressure, end diastolic left ventricular pressure, maximum upstroke rate of left ventricular pressure during systole, mean motor current. In some embodiments, the interbeat features can further include a maximum upstroke rate of a transient signal during systole, where the transient signal can include, but is not limited to, an aortic pressure signal, a motor current signal, or a combination or derivation of one or more of the foregoing signals. In another aspect, the measure of systolic function is an estimate of a preload-restricted stroke work index. In another aspect, the machine learning model includes a feedforward dense neural network.

[0008] In another aspect, performing an action based at least in part on the measure of cardiac contractility output by the machine learning model comprises displaying an indication of the measure of cardiac contractility on a user interface associated with the mechanical circulatory support device. In another aspect, the indication of the measure of cardiac contractility is a trend in cardiac contractility over a particular time range. In another aspect, performing an action based at least in part on the measure of cardiac contractility output by the machine learning model comprises determining a stability score for a patient in which the mechanical circulatory support device is implanted, the stability score based at least in part on the measure of cardiac contractility, and displaying the stability score on a user interface associated with the mechanical circulatory support device. In another aspect, performing an action based at least in part on the measure of cardiac contractility output by the machine learning model comprises providing a therapy recommendation for a patient in which the mechanical circulatory support device is implanted, wherein the therapy recommendation is determined based at least in part on the measure of cardiac contractility. In another aspect, performing an action based at least in part on the measure of cardiac contractility output by the machine learning model comprises adjusting an operating condition of the mechanical circulatory support device. In another aspect, adjusting an operating condition of the mechanical circulatory support device comprises adjusting a pump speed of the mechanical circulatory support device.

[0009] In another aspect, performing an action based at least in part on the measure of cardiac contractility output by the machine learning model comprises determining a contractility reserve based at least in part on the measure of cardiac contractility. In another aspect, the measure of cardiac contractility includes a first cardiac contractility output from the machine learning model based on signals from the mechanical circulatory support device at a first time and a second cardiac contractility output from the machine learning model based on signals from the mechanical circulatory support device at a second time after the first time, and determining a contractility reserve comprises determining the contractility reserve based on the first cardiac contractility and the second cardiac contractility. In another aspect, determining the contractility reserve based on the first cardiac contractility and the second cardiac contractility comprises determining the contractility reserve based on a difference between the first cardiac contractility and the second cardiac contractility.

[0010] In one aspect, a controller for a mechanical circulatory support device is provided. The controller includes at least one hardware processor configured to: determine a set of features based at least in part on a set of signals received from a mechanical circulatory support device; provide the set of features as input to a machine learning model trained to output a measure of cardiac contractility; and perform an action based at least in part on the measure of cardiac contractility output by the machine learning model.

[0011] In another aspect, the set of signals includes at least one pressure signal and / or at least one pump function signal. In another aspect, determining the set of features includes determining at least one feature using a combination of a first pressure signal of the at least one pressure signal and a first pump function signal of the at least one pump function signal. In another aspect, the at least one feature using a combination of a first pressure signal of the at least one pressure signal and a first pump function signal of the at least one pump function signal includes a maximum rate of rise of left ventricular pressure during systole. In another aspect, the at least one hardware processor is further configured to determine the set of features by determining one or more interbeat features based at least in part on the set of signals. In another aspect, determining the one or more interbeat features includes determining one or more of mean pump flow, mean aortic pressure, end diastolic left ventricular pressure, maximum rate of rise of left ventricular pressure during systole, mean motor current. In another aspect, the measure of cardiac contractility is an estimate of preload-restricted stroke work index. In another aspect, the machine learning model includes a feedforward dense neural network.

[0012] In another aspect, the at least one hardware processor is configured to perform an action based at least in part on the measure of cardiac contractility output by the machine learning model by displaying an indication of the measure of cardiac contractility on a user interface associated with the mechanical circulatory support device. In another aspect, the indication of the measure of cardiac contractility is a trend in cardiac contractility over a particular time range. In another aspect, the at least one hardware processor is configured to perform an action based at least in part on the measure of cardiac contractility output by the machine learning model by determining a stability score for a patient in which the mechanical circulatory support device is implanted, the stability score being based at least in part on the measure of cardiac contractility, and displaying the stability score on a user interface associated with the mechanical circulatory support device. In another aspect, the at least one hardware processor is configured to perform an action based at least in part on the measure of cardiac contractility output by the machine learning model by providing a treatment recommendation for a patient in which the mechanical circulatory support device is implanted, the treatment recommendation being determined based at least in part on the measure of cardiac contractility. In another aspect, performing an action based at least in part on the measure of cardiac contractility output by the machine learning model includes adjusting an operating condition of the mechanical circulatory support device based on the measure of cardiac contractility. In another aspect, adjusting an operating condition of the mechanical circulatory support device includes adjusting a pump speed of the mechanical circulatory support device.

[0013] In another aspect, the at least one hardware processor is configured to determine a contractility reserve based at least in part on the measure of cardiac contractility, perform an action based at least in part on the measure of cardiac contractility output by the machine learning model. In another aspect, the measure of cardiac contractility includes a first cardiac contractility output from the machine learning model based on signals from the mechanical circulatory support device at a first time and a second cardiac contractility output from the machine learning model based on signals from the mechanical circulatory support device at a second time after the first time, and determining a contractility reserve comprises determining the contractility reserve based on the first cardiac contractility and the second cardiac contractility. In one aspect, determining the contractility reserve based on the first cardiac contractility and the second cardiac contractility comprises determining the contractility reserve based on a difference between the first cardiac contractility and the second cardiac contractility.

[0014] In one aspect, a heart pump system is provided. The heart pump system includes a heart pump including at least one pressure sensor configured to sense a pressure within a portion of a heart of a patient, and a controller. The controller is configured to determine a set of features based at least in part on a set of signals received from the heart pump, the set of features including a first feature based on the sensed pressure, provide the set of features as input to a machine learning model trained to output a measure of cardiac contractility, and perform an action based at least in part on the measure of cardiac contractility output by the machine learning model.

[0015] In another aspect, the set of features further includes a second feature based on a signal corresponding to an operating state of the heart pump. In another aspect, the first feature is further based on a signal corresponding to an operating state of the heart pump. In another aspect, the first feature comprises a maximum upstroke rate of left ventricular pressure during systole. In another aspect, the controller is further configured to determine the set of features by determining one or more interbeat features based at least in part on the set of signals. In another aspect, determining the one or more interbeat features comprises determining one or more of average pump flow, average aortic pressure, end diastolic left ventricular pressure, maximum upstroke rate of left ventricular pressure during systole, average motor current. In another aspect, the measure of cardiac contractility is an estimate of a preload-restricted stroke work index. In another aspect, the machine learning model comprises a feedforward dense neural network.

[0016] In another aspect, the heart pump system further includes a display configured to display a user interface including a representation of one or more signals associated with operation of the heart pump system, and the controller is configured to perform an action based at least in part on the measure of cardiac contractility output by the machine learning model by displaying an indication of the measure of cardiac contractility on the user interface. In another aspect, the indication of the measure of cardiac contractility is a trend in cardiac contractility over a particular time range.

[0017] In another aspect, the heart pump system further includes a display configured to display a user interface including a representation of one or more signals associated with operation of the heart pump system, and the controller is configured to perform an action based at least in part on the measure of cardiac contractility output by determining a stability score for a patient in which the heart pump is implanted, the stability score based at least in part on the measure of cardiac contractility, and displaying the stability score on the user interface.

[0018] In another aspect, the heart pump system further includes a display configured to display a user interface including a representation of one or more signals associated with operation of the heart pump system, and the controller is configured to perform an action based at least in part on the measure of cardiac contractility output by providing a treatment recommendation for a patient in which the heart pump is implanted on the user interface, wherein the treatment recommendation is determined based at least in part on the measure of cardiac contractility. In another aspect, the controller is configured to perform an action based at least in part on the measure of cardiac contractility output by the machine learning model by adjusting an operating condition of the mechanical circulatory support device based on the measure of cardiac contractility. In another aspect, adjusting an operating state of the heart pump includes adjusting a pump speed of the heart pump.

[0019] In another aspect, the controller is configured to perform an action by determining a contractile reserve based at least in part on the measure of cardiac contractility, and at least in part on the measure of cardiac contractility output by the machine learning model. In another aspect, the measure of cardiac contractility includes a first cardiac contractility output from the machine learning model based on a signal from the heart pump at a first time and a second cardiac contractility output from the machine learning model based on a signal from the heart pump at a second time after the first time, and determining the contractile reserve includes determining the contractile reserve based on the first cardiac contractility and the second cardiac contractility. In another aspect, determining the contractile reserve based on the first cardiac contractility and the second cardiac contractility includes determining the contractile reserve based on the difference between the first cardiac contractility and the second cardiac contractility. Attached Figure Description

[0020] Figure 1A An illustrative cardiac support device is shown that can be used in conjunction with some embodiments of this disclosure.

[0021] Figure 1B The display includes Figure 1A An explanatory cardiac support system for cardiac support devices.

[0022] Figure 2A The illustrations illustrate various features that can be used to estimate the natural heart contractility according to some embodiments of the present disclosure.

[0023] Figure 2B The process for estimating a measure of cardiac contractility using a machine learning model, according to some embodiments of the present disclosure, is illustrated schematically.

[0024] Figure 3A is a flowchart of a process for estimating a measure of cardiac contractility according to some embodiments of the present disclosure.

[0025] Figure 3B This is a flowchart of a process for estimating a measure of cardiac contractility using multiple neural networks, according to some embodiments of the present disclosure.

[0026] Figure 4 Example graphs are shown according to some embodiments of the present disclosure for deriving one or more features used to estimate cardiac contractility.

[0027] Figure 5 Example graphs showing different estimated measures of cardiac contractility according to some embodiments of the present disclosure.

[0028] Figure 6A and 6BThe comparison of estimated measures of cardiac contractility under load and contractility according to some embodiments of the present disclosure is described separately.

[0029] Figure 7 This is a flowchart of a process for estimating a measure of cardiac contractility according to some embodiments of the present disclosure.

[0030] Figure 8 This illustration shows a portion of an example user interface for displaying an estimated measure of cardiac contractility according to some embodiments of the present disclosure.

[0031] Figure 9 This image shows a portion of an example user interface, according to some embodiments, for displaying cardiac information of a patient with an implanted heart pump system.

[0032] Figure 10 This is a flowchart of a process for determining a contractile reserve according to some embodiments of the present disclosure. Detailed Implementation

[0033] Contractility is a measure of cardiac function. At the cellular level, contractility can be defined as how much force a cardiomyocyte (muscle cell) can generate. At the chamber level (e.g., the left ventricle), contractility can be defined as the sum of the cellular forces acting in parallel to produce ventricular wall motion and ejection of blood. As a property of the heart, contractility is sometimes referred to as cardiotonicity, cardiac cardiotonicity, cardiotonic effect, or cardiotonic grade.

[0034] Because contractility is a force measurement at the cellular level (e.g., in Newtons), measuring the true sum of these forces can be challenging. Instead, measurable agents have been developed and empirically validated. These measures include preload supplemental stroke work (PRSW), end-systolic pressure-volume relationship (ESPVR), and so on. In each case, it can be assumed that the measurable agent changes with contractility and, ideally, remains stable (unchanged) attributable to changes in other factors (e.g., load status, heart rate, or other non-cardiac effects). As part of clinical practice, measurements such as PRSW and ESPVR can be difficult to measure, so alternatively, other agents such as ejection fraction, wall motion / fractional shortening, and strain-based measures (Doppler, dense MRI) can be used to estimate systolic function.

[0035] Cardiac load (e.g., preload) is a measure of the degree of strain or “load” experienced by the myocardium before contraction. Because the myocardium can automatically change the amount of force generated depending on the length of the muscle cells at the start of contraction, there can be significant overlap between the concepts of contractile force, contractility, and preload. To distinguish these concepts, a force-length curve (or Frank-Starling curve, at the chamber scale) can be used to represent the different levels of force generated depending on muscle length for a given contractile state (e.g., for a specific level of contractility). An “optimal” length that produces peak force can be assumed. Variations along the force-length curve can be considered contractile changes attributable to load or “load effect” rather than changes in contractility. Such contractile changes attributable to load effects can also be termed “length-dependent activation.”

[0036] For a given initial length, contractile changes are the peak force changes generated by cardiac cells and chambers. Such contractile changes are also known as "length-independent activation." Using the force-length relationship, contractile changes can be represented as changes in the entire force-length curve (scaled up or down), reflecting a new contractile state that depends on variations in muscle length within itself. Cardiotonic effects may be directly related to the exchange of calcium ions (Ca++) in muscle cells. This Ca++ exchange can be modified, for example, by altering the rate of Ca++ influx into cells, the rate of Ca++ release from the sarcoplasmic reticulum, and / or the sensitivity of troponin-C (which binds to C++ as part of the muscle contraction cycle). At the cardiac chamber level, contractility can also be modified by changing the total number of muscle cells contributing to the "sum" of cellular contractility. For example, contractile changes can be observed during disease states where a portion of the total muscle cells cannot efficiently circulate Ca++.

[0037] The systems, devices, and methods described herein enable support devices (e.g., mechanical circulatory support devices) that are wholly or partially located within an organ to assess one or more aspects of that organ's function. Specifically, the systems, devices, and methods described herein enable cardiac pump systems (e.g., percutaneous ventricular assist devices) to be used to assess cardiac function. For example, such devices can be used to estimate a measure of the natural cardiac contractility of a patient's myocardial layer.

[0038] Monitoring natural cardiac contractility may be helpful for patients supported by mechanical circulatory support (MCS) devices, as it can provide important information about the recovery of myocardial function. Cardiac recovery refers to the ability of the myocardium to restore its function and strength after a period of injury or disease, and contractility is an important component of cardiac recovery because it refers to the ability of the myocardium to contract and pump blood effectively.

[0039] In conditions such as heart failure, myocardial infarction, or cardiomyopathy, the heart muscle may weaken and lose its ability to contract effectively. However, with appropriate treatment and management, the heart may be able to restore its function and improve its contractility. The ability to continuously monitor and track trends in contractility allows physicians to make informed decisions regarding patient management.

[0040] Some embodiments of this disclosure relate to a device (e.g., a computing device) configured to determine natural cardiac contractility (also referred to herein simply as "contractility") using signals from a mechanical circulatory support (MCS) device. In some embodiments, the device may be configured to use only signals from the MCS device. In some embodiments, the device may be configured to continuously determine contractility at multiple time points such that a trend in contractility over a timescale of interest can be determined. In some embodiments, the amount and / or trend of contractility may be used to determine and / or recommend treatment strategies for patients supported by a mechanical circulatory device.

[0041] The inventors have recognized and understood that existing techniques for measuring contractility typically rely on the simultaneous measurement of left ventricular pressure and volume, and the complexity and invasiveness of this process often hinder its clinical application. Furthermore, existing techniques for measuring contractility do not tend to provide consistently accurate measurements, thus contractile trends are often unobservable. Therefore, some embodiments of this disclosure relate to an improved technique for estimating contractility, which utilizes signals associated with the pump of the MCS device during its operation.

[0042] The inventors have also recognized and understood that, because the pump of the MCS device can be positioned across the aortic valve in the left ventricle, the rate of change of pressure, measured using sensors on the pump, can be measured, reflecting natural cardiac contractile function. Furthermore, natural cardiac pumping can cause the pump to deviate from its set performance curve. The amount of deviation may be related to the strength of natural cardiac function. In some embodiments, the measurement of the rate of change of pressure and pump performance parameters, along with other information derived from pump signals (e.g., pressure, speed, motor current), are provided as input to a machine learning model trained to output an estimate of natural cardiac contractility.

[0043] Figure 1AAn illustrative embodiment of a blood pump assembly 100 according to the present disclosure is shown. The blood pump assembly 100 may include a pump 101, a pump housing 103, a proximal end 105, a distal end 107, a cannula 108, an impeller (not shown), a damage-resistant extension 102, a catheter 112, an inlet region 110, an outlet region 106, and a blood discharge orifice 117. In some embodiments, the catheter 112 may be connected to the inlet region 110 of the cannula 108. The inlet region 110 may be located near the proximal end 105 of the cannula, and the outlet region 106 may be located toward the distal end 107 of the cannula 108. The inlet region 110 may include a pump housing 103 having a peripheral wall 111 extending about an axis of rotation of the impeller blades, the peripheral wall 111 being located radially outward of the inner surface relative to the axis of rotation of the impeller. An impeller may be rotatably coupled to the pump 101 in the inlet region 110 adjacent to a blood discharge orifice 117 formed in the wall 111 of the pump housing 103. According to some embodiments, the pump housing 103 may be made of metal. An extension 102 (also referred to as a “tail wire”) may be connected to the distal end 107 of the cannula 108 and may assist in stabilizing and / or positioning the blood pump assembly 100 in the correct location within the heart. The tail wire 102 may be configured from a straight configuration to a partially curved configuration. The tail wire 102 may be at least partially made of a flexible material and may have dual rigidity. It should be understood that some embodiments of the pump assembly may not include the tail wire 102.

[0044] The cannula 108 may have a shape that matches (or is similar to) the anatomy of the patient's right ventricle. Figure 1A In the exemplary embodiment shown, the cannula has a proximal end 105 positioned near the patient's inferior vena cava and a distal end 107 positioned near the pulmonary artery. The cannula 108 may include a first segment S1 extending from the inflow region to point B between the inlet region 110 and the outlet region 106. The cannula 108 may also include a second segment S2 extending from point C between the inlet region 110 and the outlet region 106 to the outlet region 106. In some embodiments, points B and C may be positioned at the same location along the cannula 108. The first segment S1 of the cannula may be formed in an 'S' shape in a first plane. In some embodiments, segment S1 may have a curvature between 30 degrees and 180 degrees. The second segment S2 of the cannula may be formed in an 'S' shape in a second plane. In some embodiments, segment S2 may have a curvature between 30 degrees and 180 degrees (e.g., 40°, 50°, 60°, 70°, 80°, 90°, 100°, 110°, 120°, 130°, 140°, 150°, 160°, or 170°). The second plane may be different from the first plane. In some embodiments, the second plane may be parallel to or equivalent to the first plane.

[0045] Although shown as having an 'S' shape, it will be understood that other embodiments of the blood pump assembly may be formed with other shapes (e.g., a 'U' shape), or may have no shape at all when outside the body. In such embodiments, the cannula may be formed of a flexible material such that the cannula can bend during insertion and achieve the desired shape once inside the patient's heart.

[0046] In some embodiments, the blood pump assembly 100 may be percutaneously inserted through the internal jugular vein, across the right atrium, and into the right ventricle. When properly positioned, the blood pump assembly 100 may deliver blood from an inlet region 110 (located medial to the patient's right atrium) through a cannula 108 to a blood discharge orifice 117 of a pump housing 103 located in the pulmonary artery. Alternatively, in some embodiments, the blood pump assembly 100 may be percutaneously inserted through the femoral artery and into the left ventricle to deliver blood from the left ventricle into the aorta.

[0047] Figure 1B The blood pump assembly 100 is shown as part of a cardiac support system 120. The cardiac support system 120 may also include a controller 130 (e.g., the Automated Impella Controller from ABIOMED, ​​Inc., Danvers, Mass.). ® The system comprises, as shown herein, a display 140, a purge subsystem 150, a connector cable 160, an embolization plug 170, and a repositioning unit 180. As shown, a controller 130 may include a display 140. The controller 130 may be configured to monitor and control the operation of the blood pump assembly 100. During operation, the purge subsystem 150 may be configured to deliver purge fluid through a catheter 112 to the blood pump assembly 100 to prevent blood from entering the motor of the heart pump (not shown). In some embodiments, the purge fluid is a glucose solution (e.g., a 5% aqueous glucose solution containing 25 or 50 IU / mL heparin, although the solution does not need to contain heparin in all embodiments). The connector cable 160 provides an electrical connection between the blood pump assembly 100 and the controller 130. The embolization plug 170 may connect the catheter 112, the purge subsystem 150, and the connector cable 160. In some embodiments, the embolization unit 170 includes a storage device (e.g., a memory) configured to store, for example, operating parameters to facilitate the transfer of the patient to another controller when needed. The repositioning unit 180 can be used to reposition the blood pump assembly 100 in the patient's heart (e.g., by maintaining the position of the pump assembly relative to the patient).

[0048] like Figure 1BAs shown, in some embodiments, the cardiac support system 120 may include a purge subsystem 150, which has a container 151, a supply line 152, a purge cartridge 153, a purge disc 154, a purge tube 155, a check valve 156, an accumulator 157, a perfusion filter 158, and a side arm 159. The container 151 may be, for example, a bag or a bottle. As will be understood, in other embodiments, the cardiac support system 120 may not include a purge subsystem. In some embodiments, purge fluid may be stored in the container 151. The supply line 152 may provide a fluid connection between the container 151 and the purge cartridge 153. The purge cartridge 153 may control the manner in which the purge fluid in the container 151 is delivered to the blood pump assembly 100. For example, the purge cartridge 153 may include one or more valves for controlling the pressure and / or flow rate of the purge fluid. The purge disc 154 may include one or more pressure and / or flow sensors for measuring the pressure and / or flow rate of the purge fluid. As shown, the controller 130 may include a purge chamber 153 and a purge disc 154. A purge tube 155 may provide a fluid connection between the purge disc 154 and a check valve 156. An accumulator 157 may provide additional filling volume during changes in the purge fluid. In some embodiments, the accumulator 157 may include a flexible rubber diaphragm that provides additional filling volume via an expansion chamber. An infusion filter 158 may help prevent bacterial contamination and air ingress into the catheter 112. A side arm 159 may provide a fluid connection between the infusion filter 158 and the plug 170. Although shown with separate purge tubes and connector cables, it will be understood that in some embodiments, the cardiac support system 120 may include a single connector with both fluid lines and electrical wires that can be connected to the controller 130.

[0049] During operation, controller 130 may be configured to receive measurements from one or more pressure sensors (not shown) included as part of blood pump assembly 100 and purge disc 154. Controller 130 may also be configured to control the operation of the motors (not shown) of blood pump assembly 100 and purge cartridge 153. In some embodiments, controller 130 may be configured to control and measure the pressure and / or flow rate of the purge fluid via purge cartridge 153 and purge disc 154. During operation, after exiting purge subsystem 150 via side arm 159, purge fluid may be guided through purge lumen (not shown) within conduit 112 and embolization 170. Conduit 112, connector cable 160, and sensor cables (not shown) within embolization 170 may provide electrical connections between components of blood pump assembly 100 (e.g., one or more pressure sensors) and controller 130. The motor cable (not shown), connector cable 160, and embolization plug 170 within catheter 112 provide an electrical connection between the motor of blood pump assembly 100 and controller 130. During operation, controller 130 can be configured to receive measurements from one or more pressure sensors of blood pump assembly 100 via sensor cables (e.g., optical fibers) and control the electrical power delivered to the motor of blood pump assembly 100 via the motor cables. By controlling the power delivered to the motor of blood pump assembly 100, controller 130 can be operated to control the speed of the motor.

[0050] Various modifications can be made to one or more of the cardiac support system 120 and its components. For example, one or more additional sensors can be added to the blood pump assembly 100. In another example, a signal generator can be added to the blood pump assembly 100 to generate a signal indicating the rotational speed of the motor of the blood pump assembly 100. As another example, one or more components of the cardiac support system 120 can be separate. For example, the display 140 can be incorporated into another device that communicates with the controller 130 (e.g., wirelessly or via one or more cables).

[0051] As described herein, signals from an MCS device (e.g., blood pump assembly 100) can be used to estimate natural cardiac contractility. Specifically, according to some embodiments of this disclosure, the controller of the MCS device can determine pressure signals (e.g., left ventricular pressure) used at least in part to estimate or mimic one or more measures of cardiac contractility. Figure 2AVarious features 200 that can be used to estimate natural cardiac contractility 220 according to some embodiments of this disclosure are illustrated schematically. As described herein, because contractility may not be directly measurable, feature 200 may include one or more contractility surrogate features (examples of which are described herein), such as ventricular wall movement 202, ejection fraction 204, end-systolic pressure-volume relationship (ESPVR), and preload supplemental stroke work (PRSW). When compared with other contractility surrogate features (e.g., ejection fraction), some contractility surrogate features (e.g., ESPVR and PRSW) have been shown to be more representative of natural cardiac contractility, for example, as measured in animal studies. Figure 2A As shown, feature 200 may also include information determined based on signals (e.g., pressure signals) associated with the operation of the heart pump. For example, one or more features may be derived from the LV pressure signal normalized to pressure, including, but not limited to, the maximum slope of the signal during systole (dP / dt max 206) or the slope of the signal during systole. In some embodiments, feature 200 may include normalized features. For example, feature 200 may include dP / dt normalized to mean pressure, normalized to end-diastolic pressure before the start of systole, or normalized to the pressure integral during the systolic cycle up to the point of maximum dP / dt. Feature 200 may be used alone or in any suitable combination to estimate natural cardiac contractility, examples of which are described herein.

[0052] The inventors have recognized and understood that some features (e.g., feature 200) do not provide the best estimate of natural cardiac contractility when considered in isolation. Accordingly, some embodiments of this disclosure relate to using machine learning methods to process MCS device signals and / or features derived from MCS device signals to empirically achieve an estimated measure of improved cardiac contractility. Figure 2B This illustration demonstrates how machine learning models according to some embodiments of the present disclosure can be used to estimate cardiac contractility. Figure 2B As shown, one or more pump features 230 (examples of which include, but are not limited to, features 222 determined from pump signals (e.g., motor current, pressure, pump flow rate), features 224 determined from pressure waveform signals (e.g., dP / dt max), and correlated features (e.g., contraction area) 226) can be provided as input to a machine learning (ML) model 240, which is trained to output an estimated measure 250 of cardiac contractility based on signals associated with the pump of the MCS device. Figure 2B In this example, the model-estimated measure of cardiac contractility 250 is the pump preload supplement stroke work index (PPSWi).

[0053] Figure 3A is a flowchart of process 300 for determining an estimated measure of cardiac contractility (e.g., pump PRSWi) using a machine learning model according to some embodiments. Although pump PRSWi is described herein as an estimated measure of cardiac contractility, it should be understood that the machine learning model may alternatively be trained to output estimates of other agents of cardiac contractility, such as ESPVR. In action 302, one or more MCS device signals (e.g., pressure signals, motor current, blood flow rate) are received. Process 300 may then proceed to action 304, in which one or more intercardiac features are estimated at least in part based on the received MCS device signals. Non-limiting examples of intercardiac features that may be estimated in action 304 include the average pump flow rate between two consecutive heartbeats, the average aortic pressure between two consecutive heartbeats, the end-diastolic left ventricular (LV) pressure (LVED), dP / dt max defined as the maximum rate of rise of LV pressure during LV systole, and the average motor current of the pump between two consecutive heartbeats. Process 300 may then proceed to action 306, where one or more inter-cardiac features estimated in action 304 are provided as input to a trained machine learning model. In example process 300, the trained machine learning model may be a feedforward dense neural network. Process 300 may then proceed to action 308, where an estimated measure of cardiac contractility is output from the trained machine learning model. In example process 300, the output of the trained machine learning model is pump PRSWi.

[0054] It should be understood that, according to some embodiments of this disclosure, any suitable model architecture can be used for trained machine learning models. Figure 3B The image shows an instance architecture of an ML model 350 that can be used according to some embodiments of this disclosure. Figure 3B In the examples shown, ML model 350 is implemented using a deep neural network (DNN) architecture; however, it should be understood that other ML network architectures can be used alternatively. Furthermore, ML model 350... Figure 3B The diagram is described as containing three DNNs (DNN 320, 322, and 324). It should be understood that the use of three DNNs is merely exemplary and, according to embodiments of this disclosure, any number of DNNs, including but not limited to a single DNN or more than three DNNs, may be used alternatively.

[0055] like Figure 3BAs shown, the ML model 350 receives AoP-based features 310, pump function features 312, and correlation features 314 as input. AoP-based features 310 may include features associated with pressure signals measured or derived from pressure sensors located on the pump of the MCS device. For example, AoP-based features 310 may include systolic pressure measurements, diastolic pressure measurements, time-based pressure measurements (e.g., dP / dt max), or any other suitable pressure measurements associated with pressure signals (e.g., LV pressure signals). Pump function features 312 may include features associated with the operation of the pump itself, examples of which include, but are not limited to, motor current, pump speed, and pump flow rate. Correlation features 314 may include features that depend at least in part on AoP-based features 310 and pump function features 312. For example, Figure 4 This graph displays the LV pressure (y-axis) versus motor current (x-axis) at position 400. Values ​​of the characteristics derived from this relationship (examples of which include...) Figure 4 The contraction area shown in the figure can be included in the correlation feature 314. Each of the AoP-based features 310, pump function features 312, and correlation features 314 can be provided as input to the ML model 350, which can be configured to output the value of PRSWi 326 as an estimated measure of cardiac contractility when trained (e.g., using preclinical data).

[0056] Despite the ML model 350 in Figure 3B The model is shown to receive three different types of input (i.e., AoP-based feature 310, pump function feature 312, and correlation feature 314), but it should be understood that some embodiments may include fewer (e.g., one or two) or more (e.g., more than three) types of features as input to the ML model 350, and embodiments of this disclosure are not limited in this respect. Furthermore, within each of the different types of features provided as input, any suitable number of features may be used. For example, in one embodiment, seven AoP-based features 310, three pump function features 312, and nine correlation features 314 may be used. In another instance, five AoP-based features 310, six pump function features 312, and seven correlation features 314 may be used.

[0057] In some embodiments, features provided as input to ML model 350 may be classified and provided to separate network components (e.g., separate DNNs) before being combined to determine an estimated measure of cardiac contractility (e.g., pump PRSWi). This approach facilitates the use of simpler and / or more compact ML models that can be implemented on processors with limited processing resources (e.g., on the controller of a cardiac pump system). In some embodiments, it may not be necessary to classify features before processing them using the ML model. For example, a set of features (e.g., intercardiac features) determined from MCS device signals may be provided as input to a trained neural network without classification, and the network may learn over time how to weight the different features to determine the best estimated measure of cardiac contractility.

[0058] In some embodiments, a leave-one-out training procedure can be used to train the ML model 350. For example, multiple ML models can be defined, each with the same architecture and hyperparameters. Each of the multiple ML models can be trained on separate feature datasets, and the final ML model can be defined and trained based on one or more of the multiple ML models.

[0059] In some embodiments, estimated measures of natural cardiac contractility, such as those determined by the output of an ML model, can be visualized, trend-analyzed, and / or longitudinally tracked. For example, estimated contractility values ​​can be displayed on one or more displays associated with the MCS device. Figure 5 Example plot 500 shows an estimated measure of cardiac contractility (pump PRSWi) 514 determined using a trained machine learning model, as described herein. Trend analysis of the pump PRSWi value 514 is performed over a relative period of several hours and compared with a measured (e.g., real-world) PRSWi value 510 (e.g., when a stress test is performed) and a measured dP / dt max value 512 determined based on the LV pressure signal of the cardiac pump system. Figure 5 As can be observed, the pump PRSWi value 514 more closely tracks the measured PRSWi value 510 compared to the standalone dP / dt max metric 512, demonstrating the effectiveness of using ML-based techniques to reliably estimate cardiac contractility. For example, when an increase in contractility is induced around the 10:50 mark in graph 500, there is a good correspondence between the pump PRSWi metric 514 and other PRSW metrics. Furthermore, graph 500 shows that after the initial induction of contractility (e.g., after the 11:00 mark in graph 500), the pump PRSWi metric 514 more closely tracks the real-world PRSWi metric 510 compared to the dP / dt max metric 512. Such comparisons demonstrate the ability of trained ML models to accurately predict natural cardiac contractility based on signals associated with the pump in the MCS device.

[0060] Figure 6A and 6B Further explanation is provided regarding the comparison of the pump PRSWi value, determined as the output of the trained ML model using the techniques described in this paper, with surrogate measures conventionally used to estimate contractility. Ideally, an accurate estimated measure of cardiac contractility would be sensitive to changes in contractility but insensitive to changes in load. Figure 6A The curve in 600 and Figure 6B The curve 610 in the figure shows that the pump measurement performance (pump PRSWi) determined using the technique described herein is based on the sensitivity to both load and contractility between the reference (PRSWi) and input (dP / dt max) values ​​of the model, thus indicating that the pump signal-based technique described herein for determining contractility performs similarly to existing techniques for determining contractility.

[0061] Figure 7 This document describes a process 700 for estimating a measure of cardiac contractility using signals from a mechanical circulation support (MCS) device and a trained machine learning model, according to some embodiments of this disclosure. Process 700 may begin at action 710, where a set of features is determined during operation of the MCS device. For example, as described herein, one or more signals associated with operation of the MCS device (e.g., pressure signals, flow rate signals, motor current signals, etc.) may be processed to extract one or more features (e.g., one or more inter-cardiac features) that may be included in the set of features. Process 700 may then proceed to action 712, where the set of features is provided as input to a machine learning (ML) model trained to output a measure of cardiac contractility. For example, as described herein, the ML model may be trained (e.g., using preclinical data) to estimate PRSWi based on features extracted from pump signal data.

[0062] Process 700 may then proceed to action 714, wherein the action is performed at least in part based on a measure of cardiac contractility output from the ML model. In some embodiments, the action may be to display the value and / or trend of the cardiac contractility measurement on a user interface associated with the MCS device. In some embodiments, the cardiac contractility measurement may be measured over time, compared to a baseline measurement at a previous time point, or as an absolute or relative (e.g., percentage) change in contractility relative to the baseline measurement. In such embodiments, an indication of the change in contractility relative to the baseline measurement may be displayed. In some embodiments, the action may be to provide clinical decision support (CDS) to a user associated with the MCS device (e.g., a physician or other healthcare professional). For example, when the cardiac contractility measure is above (or below) a threshold, an alert or other indication may be provided to the user, indicating that the patient's cardiac function is improving (or declining) and that the user can take appropriate action based on the alert / indication to improve patient management. In some embodiments, the action may be to use the cardiac contractility measurement to modify one or more other measures that can be used to facilitate patient management. In some embodiments, the action may be to adjust the operating conditions of the MCS device (with or without user assistance). For example, the pump rate of the MCS device may be adjusted at least in part based on a measure of cardiac contractility. In this way, the measure of cardiac contractility can be used as feedback to adjust the support provided by the MCS device according to the patient's needs to further facilitate the patient's return to natural cardiac function.

[0063] In some embodiments, once the MCS device is deployed, cardiac contractility measures (e.g., PRSWi) can be made available for continuous use and trend analysis via user initiation or automated system initiation. Contractility measures can be used in a wider range of clinical decision support (CDS) algorithms for patient management on MCS devices, such as device escalation, device weaning, device support titration, secondary device titration, ventilator titration / extubation, or pharmacological titration of vasoactive or cardiotonic drugs.

[0064] As a measure of natural cardiac recovery, contractility can be visualized, trend-analyzed, and longitudinally tracked on a daily / weekly basis. This measure can have a significant short-term and / or long-term prognostic impact on patient outcomes.

[0065] Figure 8 Examples are shown of a portion of a user interface 800 that can display an estimated measure of cardiac contractility according to some embodiments. For example... Figure 8As shown, an estimated measure of cardiac contractility can be displayed as a numerical value 810 and / or a trend over time 820. In some embodiments, the user interface 800 may include a time window selector 830 configured to allow a user to interact with the user interface to display trends (e.g., trend 820) over different time scales.

[0066] As described herein, in some embodiments, estimated cardiac contractility measurements can be used to modify one or more other metrics determined based on signals associated with the MCS device to provide improved patient management. For some patients with implanted MCS devices, as the patient's natural cardiac function recovers, it is desirable to wean the patient off the device (and possibly eventually explant the device). In some cases, the recovery of natural cardiac function can be characterized as the recovery of the heart's contractile function. Accordingly, in some embodiments, estimated measures of contractility can be at least partially used to provide clinicians with guidance on how and / or when to wean the patient off the MCS device. Figure 9 This illustration describes an example of a portion of a user interface 900 for an offline assisted clinical decision support tool according to some embodiments of the present disclosure. Figure 9 In the example user interface 900 shown, the offline assistive tool includes a stability score 910, which indicates the relative stability of the patient, as determined by multiple variables or measures associated with the MCS device. In some embodiments, at least one of the measures used to determine the stability score 910 may be an estimated cardiac contractility measure (e.g., pump PRSWi). Based on the patient's stability score 910, a healthcare provider may decide to modify the amount of support provided to the patient by the MCS device, with the aim of weaning the patient off the device or providing additional support where warranted.

[0067] The concept of contractile reserve can be defined as the “extra capacity” or “reserve” that the heart can use to increase contraction in response to changes in physiological state. For example, following an acute event (e.g., acute myocardial infarction (AMI)), decreased contractility may be present. In response, the heart may partially increase the contractility of the remaining available muscle cells to maintain afterload / output / end-organ delivery. Similarly, in events such as hemorrhage / blood loss, a decrease in mean blood pressure may trigger the heart to partially respond to the event by increasing contractility to maintain delivery. Further complicating matters, these changes are often superimposed on other response mechanisms, including changes in load, heart rate (chronotropic), and / or vascular properties (compliance and resistance), all of which tend to alter the hemodynamic state in parallel with changes in the true inotropic effect.

[0068] To isolate and measure systolic reserve, a cardiac stress test is typically performed using exercise or pharmacological agents to trigger inotropic changes. In a typical stress test, the heart is stimulated (physically or pharmacologically) to induce changes in cellular Ca++ disposal, which leads to inotropic changes. Repeated stimulation of the heart to push it to a “maximum” inotropic state can represent the highest possible contractility (zero remaining reserve). Total systolic reserve can then be measured by subtracting peak-state contractility from baseline-state contractility, where the difference is the “systolic reserve.” Systolic reserve can be standardized to baseline or other factors. Alternatively, the ratio of peak-state to baseline-state contractility measures can be used to quantify systolic reserve on a standardized basis. Some typical clinical measures of systolic reserve include changes in ejection fraction or changes in the “wall motion score index” (WMSI). Other measures of systolic reserve may include imaging-based changes in left ventricular (LV) size or overall LV strain (e.g., circumferential or longitudinal).

[0069] Measurements of systolic reserve have demonstrated a strong correlation with clinical outcomes, particularly long-term recovery following acute events. The concept of “natural cardiac recovery” is intrinsically linked to the restoration of systolic reserve. Accordingly, quantifying systolic reserve in patients where restoring natural cardiac function is a treatment goal of hospitalization may be important. In some cases, knowledge of systolic reserve can serve as a marker for deciding on weaning / external support devices (e.g., mechanical circulatory support), reducing pharmacological support, and / or discharging the patient. Similarly, measurements of systolic reserve can be a useful indicator of a patient's quality of life in managing future stressors (e.g., climbing stairs), stressors that may not be easily assessed while the patient is resting in a hospital bed.

[0070] Contractile reserve has been shown to be a valuable prognostic indicator of short-term and long-term outcomes for several different cardiac diseases. Accordingly, the inventors have recognized and understand that measuring, estimating, and / or predicting contractile reserve may be a valuable tool for patient management in patients receiving mechanical circulatory support. The inventors have further recognized and understand that existing solutions for determining contractile reserve often fail to accurately measure contractility. Therefore, estimates of contractile reserve based on such inaccurate contractility measurements may have compounded errors, which may limit the clinical applicability of such measures. Additionally, prior art often requires intervention (e.g., stress testing) to induce changes in the patient's cardiac state and to measure differences in contractility across patients with different cardiac states. Limitations of such techniques may include their inability to be consistently performed and, in some cases, their inability to be performed at all (e.g., when the patient cannot tolerate the stress event and / or when time / resource constraints exist). Furthermore, prior art may not utilize signals that are most closely associated with changes in actual cellular contractility, such as LV pressure or LV volume.

[0071] The inventors have recognized and understood that the techniques described herein for estimating natural cardiac contractility using signals from a mechanical circulatory device can also be useful for estimating contractile reserve. For example, cardiac contractility can be measured at multiple time points (or windows) using the techniques described herein. Then, multiple measurements are used to estimate the patient's contractile reserve. For example, the difference in contractility measured at two different time points can be used to estimate the patient's contractile reserve. In some embodiments, cardiac reserve can be measured by using the operation of the heart pump itself to apply different strains to the patient's heart, rather than by subjecting the patient to conventional cardiac stress tests (e.g., induced by exercise or pharmaceutical preparations). For example, changes in pump function may induce changes in contractility due to a "mini-stress test." In some embodiments, the pump may be configured to operate continuously and the prediction or trend analysis of cardiac reserve over time can be performed using one or more techniques described herein.

[0072] In some embodiments, trend contractility can be determined during a typical clinical stress test. For example, the device can queue or store a first contractility value determined prior to the stress test (e.g., using the techniques described herein), perform the stress test (e.g., using a drug such as dobutamine or using exercise), and store a second contractility value measured at peak strain during the stress test. The first and second contractility values ​​can be used to determine a patient's contractility reserve. For example, the contractility reserve can be determined based on the ratio of the first to the second contractility values, the difference between the first and second contractility values, the percentage difference between the first and second contractility values, or a regression fit / estimation of the peak strain (in the case of collecting multiple strain points).

[0073] In some embodiments, small changes in systolic function can be induced in a patient such that, according to known transformations, the degree of change correlates with the patient's true (or near-true) systolic reserve. Changes in the pump rate of a mechanical circulatory support device over short time periods (e.g., ~1-minute time periods) may induce changes in coronary infusion pressure, resulting in changes in coronary flow and temporary changes in systolic function, which may trigger a reserve effect (to some extent) to restore homeostasis, where the degree of change correlates with peak reserve. Coronary infusion pressure can represent the difference between aortic pressure and left ventricular pressure. Different pump flow rates can alter coronary flow / infusion, which can simulate or otherwise mimic a dipyridamole stress test. Run times at high and low flow rates can be longer to allow the heart to respond to changes in coronary flow, and changes in systolic and diastolic function can be mapped to systolic reserve. In some embodiments, when the heart transitions from a first “non-stressed” state to a “stressed” state, systolic reserve can be estimated using pressure measurements from the heart pump (e.g., reflecting coronary infusion pressure).

[0074] In some embodiments, systolic reserve can be determined at least in part based on a combination of electrocardiogram (ECG) information and information from left ventricular (LV) pressure waveforms. For example, ECG-based lead signals can be used to measure excitation-contraction coupling (EC coupling) components (e.g., the delay between the onset of electrical contraction and the mechanical upstroke), which may reflect excess capacity of calcium channels. One or more extracted features of the ECG waveform (e.g., QRS interval, PR interval, T wave amplitude, morphology of the electrical depolarization and repolarization phases, etc.) can be combined with information associated with the profile of LV pressure measured from a pump in a mechanical circulatory support device to predict a patient's systolic reserve.

[0075] In some embodiments, systolic reserve can be determined at least in part based on a combination of known left ventricular end-systolic volume (LV ESV) and information from left ventricular pressure waveforms (e.g., LV SP / LV ESV). As a supplement to or alternative to using ECG information, information from other sources (e.g., first and second heart sounds) containing information about the perturbation / degree of the initial contraction can also be used. When combined with one or more characteristics of the LV pressure signal, the relationship between heart sounds and pressure generation can indicate the expected “hidden reserve.” For example, the LV pressure waveform can reflect the degree of contraction (force) and heart sounds can reflect the degree of movement (tension / shortening / valve action).

[0076] In some embodiments, systolic reserve can be determined at least in part based on morphological signals in available waveforms (including, but not limited to, LV diastolic filling curves, current systolic values, heart rate, cardiac output, and / or cardiac power output (CPO)). In such embodiments, the focus may be on diastolic / relaxation during early diastole and the range of true passive cycles during diastole (rapid diastole / short relaxation), which may be associated with a greater capacity for systolic reserve. When the heart operates at volume or near-volume (lower reserve) levels, there is likely to be little / no relaxation time and little true passive cycle.

[0077] In some embodiments, a system of lumped parameter models of the systolic circulation and the heart (e.g., all available signals) can be used, at least in part, to determine systolic reserve. In this embodiment, the model can then be used to simulate a stress test, and the systolic reserve can be estimated from the simulation. The systolic reserve can be determined based on actual contractility (e.g., PRSW) or via another simulation agent. For example, estimates of cardiac contractility using the techniques described herein (e.g., pump PRSWi) can be determined at multiple time points, and a patient's systolic reserve can be determined based on multiple determinations of cardiac contractility.

[0078] In some embodiments, systolic reserve can be determined at least in part by combining information from imaging (e.g., ultrasound) and the relationship between ventricular wall strain / movement, as determined from mechanical circulatory support devices, and the resulting LV pressure. In such embodiments, knowledge of the LV pressure waveform morphology associated with strain rate, ventricular wall motion, or end-systolic volume can be used to determine a patient's systolic reserve.

[0079] It should be understood that one or more of the foregoing concepts can be combined in any suitable manner to determine a patient's systolic reserve. For example, in some embodiments, systolic reserve can be determined at least in part based on heart sound information, imaging information, and information from LV pressure waveforms.

[0080] Figure 10 This description describes a process 1000 for determining contractile reserve according to some embodiments of the present disclosure. Process 1000 may begin with action 1010, where a first cardiac contractility is determined based on signals recorded at a first time. For example, a set of features may be determined from signals recorded from a mechanical circulatory support device at a first time (e.g., a single time point or time window), and said set of features may be provided as input to a trained machine learning model, where the output of the trained machine learning model is the first cardiac contractility. Process 1000 may then proceed to action 1012, where a second cardiac contractility is determined based on signals recorded at a second time. For example, a set of features may be determined from signals recorded from a mechanical circulatory support device at a second time (e.g., a single time point or time window), where the second time is after the first time. The set of features may be provided as input to a trained machine learning model, where the output of the trained machine learning model is the second cardiac contractility. Process 1000 may then proceed to action 1014, where contractile reserve is determined based on the first and second cardiac contractility. For example, contractile reserve can be determined based on the ratio of the first to the second cardiac contractile values, the difference between the first and the second cardiac contractile values, the percentage difference between the first and the second cardiac contractile values, or a regression fit / estimation of peak strain (in the case of collecting multiple strain points).

[0081] In some embodiments, the contractile reserve determined in action 1014 may be used to perform further actions, including but not limited to providing clinical decision support to a healthcare provider, determining a patient’s off-board status or stability score, or other appropriate actions described herein in conjunction with measures for determining cardiac contractility.

[0082] Therefore, having described several aspects and embodiments of the technology set forth in this disclosure, it should be understood that various changes, modifications, and improvements will readily occur to those skilled in the art. Such changes, modifications, and improvements are intended to be within the spirit and scope of the technology described herein. For example, those skilled in the art will readily conceive of various other components and / or structures for performing the functions described herein and / or obtaining the results and / or one or more advantages described herein, and each of such changes and / or modifications is considered to be within the scope of the embodiments described herein. Those skilled in the art will recognize or be able to confirm many equivalents of the specific embodiments described herein using only conventional experimentation. Therefore, it should be understood that the foregoing embodiments are presented by way of example only and that embodiments of this disclosure can be practiced in ways other than those specifically described. Furthermore, any combination of two or more features, systems, articles, materials, kits, and / or methods described herein is included within the scope of this disclosure, provided that such features, systems, articles, materials, kits, and / or methods do not conflict with each other.

[0083] The above embodiments can be implemented in any of a variety of ways. One or more aspects and embodiments of this disclosure relating to the execution of processes or methods can be performed or controlled using program instructions executable by means of an apparatus (e.g., a computer, processor, or other means). In this regard, various inventive concepts can be embodied in computer-readable storage media (or multiple computer-readable storage media) encoded with one or more processes (e.g., computer memory, one or more floppy disks, laser disks, optical disks, magnetic tapes, flash memory, field-programmable gate arrays, or other circuit configurations in semiconductor devices, or other tangible computer storage media), which, when said one or more programs are executed on one or more computers or other processors, perform the methods implementing one or more of the various embodiments described above. The computer-readable media or several computer-readable media may be transportable, such that the program or several programs stored thereon can be loaded onto one or more different computers or other processors to implement each of the aspects described above. In some embodiments, the computer-readable media may be a non-transitory media.

[0084] The above embodiments of this technology can be implemented in any of a variety of ways. For example, the embodiments can be implemented using hardware, software, or a combination thereof. When implemented in software, the software code can be executed on any suitable processor or set of processors, whether located in a single computer or distributed across multiple computers. It should be understood that any component or set of components performing the functions described above can generally be considered as a controller controlling the functions described above. The controller can be implemented in a variety of ways (e.g., using dedicated hardware or using general-purpose hardware (e.g., one or more processors) programmed using microcode or software to perform the functions listed above), and when the controller corresponds to multiple components of a system, it can be implemented in a combination of ways.

[0085] Furthermore, it should be understood that a computer can be embodied in several ways, as non-limiting examples, such as a rack-mount computer, desktop computer, laptop computer, or tablet computer. Additionally, a computer can be embedded in a device that is not typically considered a computer but has appropriate processing capabilities, including a personal digital assistant (PDA), a smartphone, or any other suitable portable or stationary electronic device.

[0086] In addition, a computer may have one or more input and output devices. These devices are particularly useful for presenting a user interface. Examples of output devices that can be used to provide a user interface include printers or display screens for visual presentation and speakers or other sound-producing devices for auditory presentation. Examples of input devices that can be used for a user interface include keyboards and pointing devices, such as mice, touchpads, and digitizers. As another example, a computer may receive input information via voice recognition or in other auditory formats.

[0087] Such computers can be interconnected in any suitable form through one or more networks, including local area networks (LANs) or wide area networks (WANs), such as enterprise networks and intelligent networks (INs) or the Internet. These networks can be based on any suitable technology and operate according to any suitable protocol, and can include wireless networks, wired networks, or fiber optic networks.

[0088] Furthermore, as described, some aspects can be embodied in one or more methods. Actions performed as part of said method can be ordered in any suitable manner. Accordingly, embodiments in which actions are performed in a different order than that described can be constructed, which may include performing some actions simultaneously, even if shown as sequential actions in the illustrative embodiments.

[0089] As defined and used herein, all definitions should be understood as controlling dictionary definitions, definitions in referenced literature, and / or the general meaning of the defined terms.

[0090] As used in this article, the indefinite articles “a” and “an” should be understood to mean “at least one” unless there is an explicit indication to the contrary.

[0091] As used herein, the phrase “and / or” should be understood as “any one or both” of the elements so combined, that is, elements that exist together in some cases and separately in others. Multiple elements listed with “and / or” should be interpreted in the same way, that is, “one or more” of the elements so combined. In addition to the elements specifically identified by the “and / or” clause, other elements may optionally exist, whether related to or unrelated to the specifically identified elements. Thus, as a non-limiting example, when used in conjunction with open-ended language such as “including,” a reference to “A and / or B” may in one embodiment refer to only A (optionally including elements other than B); in another embodiment, refer to only B (optionally including elements other than A); in yet another embodiment, refer to both A and B (optionally including other elements); and so on.

[0092] As used herein, the phrase "at least one" in relation to a list of one or more elements should be understood to mean any one or more of the elements selected from the list of elements, but not necessarily including at least one of each element specifically listed in the list of elements, and does not exclude at least one of any combination of elements in the list of elements. This definition also allows for the optional presence of elements referred to by the phrase "at least one" other than those specifically identified in the list of elements, whether related to or not the specifically identified elements. Therefore, as a non-limiting example, "at least one of A and B" (or equivalently, "at least one of A or B", or equivalently "at least one of A and / or B") in one embodiment may refer to at least one that optionally includes more than one A and does not include B (and optionally includes elements other than B); in another embodiment, it optionally includes more than one B and does not include A (and optionally includes elements other than A); in yet another embodiment, it optionally includes at least one of more than one A and optionally includes more than one B (and optionally includes other elements); etc.

[0093] Furthermore, the phrases and terms used herein are for descriptive purposes and should not be considered restrictive. The use of “comprising,” “including,” “having,” “containing,” “involving,” and variations thereof in this document means to cover the items listed thereafter and their equivalents, as well as additional items.

Claims

1. A computer-implemented method, comprising: Receive a set of signals from the mechanical circulation support device; A set of features is determined using a computer processor, at least in part, based on the set of signals; The set of features is provided as input to a machine learning model, which is trained to output a measure of cardiac contractility. and The action is performed at least in part based on the measure of cardiac contractility output by the machine learning model.

2. The computer implementation method according to claim 1, wherein the set of signals includes at least one pressure signal and / or at least one pump function signal.

3. The computer implementation method according to claim 2, wherein determining the set of features includes determining at least one feature using a combination of a first pressure signal of the at least one pressure signal and a first pump function signal of the at least one pump function signal.

4. The computer implementation method of claim 3, wherein the at least one feature of the combination of the first pressure signal of the at least one pressure signal and the first pump function signal of the at least one pump function signal includes the maximum rate of rise of left ventricular pressure during systole.

5. The computer implementation method according to claim 1, wherein determining a set of features includes: One or more intercardiac features are determined at least in part based on the set of signals.

6. The computer-implemented method according to claim 5, wherein determining the one or more intercardiac features includes: Determine one or more of the following: average pump flow rate, average aortic pressure, end-diastolic left ventricular pressure, maximum rate of rise of left ventricular pressure during systole, and average motor current.

7. The computer implementation method according to claim 1, wherein the measure of cardiac contractility is an estimate of the preload supplemental stroke work index.

8. The computer implementation method according to claim 1, wherein the machine learning model comprises a feedforward dense neural network.

9. The computer implementation method of claim 1, wherein performing the action based at least in part on the measure of cardiac contractility output by the machine learning model comprises: The measurement of cardiac contractility is displayed on the user interface associated with the mechanical circulatory support device.

10. The computer implementation method of claim 9, wherein the indication of the measure of cardiac contractility is a trend of cardiac contractility over a specific time range.

11. The computer-implemented method of claim 1, wherein performing the action based at least in part on the measure of cardiac contractility output by the machine learning model comprises: Determine a stability score for patients in whom the mechanical circulatory support device is implanted, the stability score being based at least in part on the measure of cardiac contractility; and The stability score is displayed on the user interface associated with the mechanical circulation support device.

12. The computer-implemented method of claim 1, wherein performing the action based at least in part on the measure of cardiac contractility output by the machine learning model comprises: Treatment recommendations are provided for patients in whom the mechanical circulatory support device is implanted, wherein the treatment recommendations are determined at least in part based on the measure of cardiac contractility.

13. The computer-implemented method of claim 1, wherein performing the action based at least in part on the measure of cardiac contractility output by the machine learning model comprises: Adjust the operating conditions of the mechanical circulation support device.

14. The computer-implemented method according to claim 13, wherein adjusting the operating conditions of the mechanical circulation support device includes: Adjust the pump speed of the mechanical circulation support device.

15. The computer-implemented method of claim 1, wherein performing the action based at least in part on the measure of cardiac contractility output by the machine learning model comprises: Contractile reserve is determined at least in part based on the aforementioned measure of cardiac contractility.

16. The computer-implemented method according to claim 15, wherein... The measure of cardiac contractility includes a first cardiac contractility output from the machine learning model at a first time based on signals from the mechanical circulatory support device, and a second cardiac contractility output from the machine learning model at a second time based on signals from the mechanical circulatory support device after the first time. Determining the contractile reserve includes determining the contractile reserve based on the first cardiac contractility and the second cardiac contractility.

17. The computer-implemented method of claim 16, wherein determining the contractile reserve based on the first cardiac contractility and the second cardiac contractility comprises: The contractile reserve is determined based on the difference between the first cardiac contractility and the second cardiac contractility.

18. A controller for a mechanical circulation support device, the controller comprising: At least one hardware processor configured to: A set of features is determined at least in part based on a set of signals received from the mechanical circulation support device; The set of features is provided as input to a machine learning model, which is trained to output a measure of cardiac contractility. and The action is performed at least in part based on the measure of cardiac contractility output by the machine learning model.

19. The controller of claim 18, wherein the set of signals includes at least one pressure signal and / or at least one pump function signal.

20. The controller of claim 19, wherein determining the set of features comprises: At least one feature is determined by combining the first pressure signal of the at least one pressure signal with the first pump function signal of the at least one pump function signal.

21. The controller of claim 20, wherein the at least one feature of the combination of the first pressure signal using the at least one pressure signal and the first pump function signal using the at least one pump function signal includes: The maximum rate of increase in left ventricular pressure during systole.

22. The controller of claim 18, wherein the at least one hardware processor is further configured to determine the set of features by determining one or more intercardiac features based at least in part on the set of signals.

23. The controller of claim 22, wherein determining the one or more intercardiac characteristics comprises determining one or more of the following: average pump flow rate, average aortic pressure, end-diastolic left ventricular pressure, maximum rate of rise of left ventricular pressure during systole, and average motor current.

24. The controller of claim 18, wherein the measure of cardiac contractility is an estimate of the preload supplemental stroke work index.

25. The controller of claim 18, wherein the machine learning model comprises a feedforward dense neural network.

26. The controller of claim 18, wherein the at least one hardware processor is configured to perform actions based at least in part on the measure of cardiac contractility output by the machine learning model by displaying an indication of the measure of cardiac contractility on a user interface associated with the mechanical circulatory support device.

27. The controller of claim 26, wherein the indication of the measure of cardiac contractility is a trend of cardiac contractility over a specific time range.

28. The controller of claim 18, wherein the at least one hardware processor is configured to perform an action based at least in part on the measure of cardiac contractility output by the machine learning model through the following steps: Determine a stability score for patients in whom the mechanical circulatory support device is implanted, the stability score being at least in part based on the measure of cardiac contractility; and The stability score is displayed on the user interface associated with the mechanical circulation support device.

29. The controller of claim 18, wherein the at least one hardware processor is configured to perform actions by providing treatment recommendations to a patient in which the mechanical circulatory support device is implanted, based at least in part on the measure of cardiac contractility output by the machine learning model, wherein the treatment recommendations are determined at least in part on the measure of cardiac contractility.

30. The controller of claim 18, wherein the at least one hardware processor is configured to adjust the operating conditions of the mechanical circulatory support device based on the measure of cardiac contractility, performing actions at least in part based on the measure of cardiac contractility output by the machine learning model.

31. The controller of claim 30, wherein adjusting the operating conditions of the mechanical circulation support device comprises: Adjust the pump speed of the mechanical circulation support device.

32. The controller of claim 18, wherein the at least one hardware processor is configured to perform an action by determining a contractile reserve based at least in part on the measure of cardiac contractility, and at least in part on the measure of cardiac contractility output by the machine learning model.

33. The controller according to claim 32, wherein The measure of cardiac contractility includes a first cardiac contractility output from the machine learning model at a first time based on signals from the mechanical circulatory support device, and a second cardiac contractility output from the machine learning model at a second time based on signals from the mechanical circulatory support device after the first time. Determining the contractile reserve includes determining the contractile reserve based on the first cardiac contractility and the second cardiac contractility.

34. The controller of claim 33, wherein determining the contractile reserve based on the first cardiac contractility and the second cardiac contractility comprises: The contractile reserve is determined based on the difference between the first cardiac contractility and the second cardiac contractility.

35. A heart pump system comprising: A heart pump comprising at least one pressure sensor configured to sense pressure within a portion of a patient's heart; and The controller is configured to: A set of features is determined at least in part based on a set of signals received from the heart pump, the set of features including a first feature based on the sensed pressure; The set of features is provided as input to a machine learning model, which is trained to output a measure of cardiac contractility. and The action is performed at least in part based on the measure of cardiac contractility output by the machine learning model.

36. The heart pump system of claim 35, wherein the set of features further comprises a second feature based on a signal corresponding to the operating state of the heart pump.

37. The heart pump system of claim 35, wherein the first feature is further based on a signal corresponding to the operating state of the heart pump.

38. The cardiac pump system of claim 37, wherein the first feature includes the maximum rate of rise of left ventricular pressure during systole.

39. The heart pump system of claim 36, wherein the controller is further configured to determine the set of features by determining one or more inter-cardiac features based at least in part on the set of signals.

40. The cardiac pump system of claim 39, wherein determining the one or more intercardiac characteristics comprises determining one or more of the following: average pump flow rate, average aortic pressure, end-diastolic left ventricular pressure, maximum rate of rise of left ventricular pressure during systole, and average motor current.

41. The cardiac pump system of claim 35, wherein the measure of cardiac contractility is an estimate of the preload supplemental stroke work index.

42. The heart pump system of claim 35, wherein the machine learning model comprises a feedforward dense neural network.

43. The heart pump system of claim 35, further comprising: A display configured to show a user interface containing representations of one or more signals associated with the operation of the heart pump system. The controller is configured to perform actions based, at least in part, on the measure of cardiac contractility output by the machine learning model, by displaying an indication of the measure of cardiac contractility on the user interface.

44. The heart pump system of claim 43, wherein the indication of the measure of cardiac contractility is a trend of cardiac contractility over a specific time range.

45. The heart pump system of claim 35, further comprising: A display configured to show a user interface containing representations of one or more signals associated with the operation of the heart pump system. The controller is configured to perform actions based at least in part on the measure of cardiac contractility output through the following steps: Determine a stability score for patients in whom the heart pump is implanted, the stability score being at least in part based on the measure of cardiac contractility; and The stability score is displayed on the user interface.

46. ​​The heart pump system of claim 35, further comprising: A display configured to show a user interface containing representations of one or more signals associated with the operation of the heart pump system. The controller is configured to perform actions based at least in part on the measure of cardiac contractility output through the following steps: Treatment recommendations are provided on the user interface for patients in which the heart pump is implanted, wherein the treatment recommendations are determined at least in part based on the measure of cardiac contractility.

47. The heart pump system of claim 35, wherein the controller is configured to adjust the operating conditions of the mechanical circulatory support device by means of the measure of cardiac contractility, performing actions at least in part based on the measure of cardiac contractility output by the machine learning model.

48. The heart pump system of claim 47, wherein adjusting the operating state of the heart pump includes adjusting the pump speed of the heart pump.

49. The heart pump system of claim 35, wherein the controller is configured to perform actions by determining a contractile reserve based at least in part on the measure of cardiac contractility, and at least in part on the measure of cardiac contractility output by the machine learning model.

50. The heart pump system of claim 49, wherein... The measure of cardiac contractility includes a first cardiac contractility output from the machine learning model at a first time based on signals from the heart pump and a second cardiac contractility output from the machine learning model at a second time based on signals from the heart pump after the first time. Determining the contractile reserve includes determining the contractile reserve based on the first cardiac contractility and the second cardiac contractility.

51. The heart pump system of claim 50, wherein determining the contractile reserve based on the first cardiac contractility and the second cardiac contractility comprises: The contractile reserve is determined based on the difference between the first cardiac contractility and the second cardiac contractility.

Images