Medical devices, systems, clinical decision support tools, and related methods

Abstract

The present disclosure provides embodiments relating to medical devices, systems, clinical decision support tools, and related methods. These embodiments may include the development and implementation of clinical decision support tools using artificial intelligence. In some embodiments, artificial intelligence models may be trained using simulated data and one or more physiological modifiers. In accordance with one particular embodiment, a system is provided comprising a computing device configured to receive a first set of waveform data from a patient's aorta and a first set of waveform data from a patient's left ventricle. The computing system additionally analyzes the first sets of waveform data using an artificial intelligence (Al) model trained using simulated data and physiological modifiers. The system may provide a recommendation regarding a transaortic valve replacement procedure (TAVR) in the form of a first clinical decision support tool (CDST).

Description

Attorney Docket XMT00017W01MEDICAL DEVICES, SYSTEMS, CLINICAL DECISION SUPPORT TOOLS, AND RELATED METHODSCROSS REFERENCE TO RELATED APPLICATIONS

[0001] The present application claims benefit to U.S. Provisional Patent Application No. 63 / 763,905, filed on Feb. 26, 2025, entitled MEDICAL DEVICES, SYSTEMS, CLINICAL DECISION SUPPORT TOOLS, AND RELATED METHODS, the disclosure of which is incorporated by reference herein in its entirety.BACKGROUND

[0002] The present disclosure relates generally to medical devices incorporating sensors, as well as systems and methods which may incorporate such devices, along with methods of using such systems and devices to provide practitioners with clinical support decision tools.

[0003] In one, non-limiting example, such medical devices may include intraluminal devices, such as guidewires or catheters, having one or more sensors for measuring one or more physiological parameters and / or for imaging.

[0004] Guidewire devices are often used to lead, or to guide, catheters or other interventional devices to a targeted anatomical location within a patient’s body. For example, guidewires may be passed into and through a patient’s vasculature in order to reach the target location, which may be, for example, at or near the patient’s heart or brain. Radiographic imaging is conventionally utilized to assist in navigating a guidewire to the targeted location. Guidewires are available with various outer diameter sizes. Widely utilized sizes include 0.010, 0.014, 0.016, 0.018, 0.024, and 0.035 inches in diameter, for example, though they may also be smaller or larger in diameter.

[0005] In some instances, a guidewire may be used to gather physiological information from within a patient. For example, so-called “pressure wires” conventionally incorporate a single pressure sensor to detect the blood pressure within a blood vessel of a patient.

[0006] In many instances, a guidewire is placed within the body during the interventional procedure so that it can be used to guide one or more catheters or otherAttorney Docket XMT00017USP1interventional devices to the targeted anatomical location. For example, a catheter can be guided to a targeted location and, once in place, be used to image the targeting location, to aspirate clots or other occlusions, or to deliver drugs, stents, embolic devices, radiopaque dyes, replacement valves, or other devices or substances for treating the patient.

[0007] These types of interventional devices can also include sensors located at their distal ends in order to provide added functionality to the device. For example, intravascular ultrasound (IVUS) is an imaging technique that utilizes a catheter with an ultrasound imaging sensor attached to the distal end. Ultrasound may be utilized to image within targeted vasculature (typically the coronary arteries).

[0008] There are several challenges associated with incorporating sensors into intraluminal devices. For example, such interventional devices have very limited space to work in, given the stringent dimensional constraints involved. Moreover, integrating the sensors with the interventional devices in a way that maintains effective functionality can be challenging.

[0009] As such, there is an ongoing need for improved medical devices that effectively integrate sensors and can help provide data in a more efficient manner and / or provide data previously unobtainable in a practical manner. Additionally, given that the use of these new devices gives access to new data in real-time, the present disclosure provides new methods and approaches that may be employed to leverage the use of such new devices and systems.SUMMARY

[0010] The present disclosure provides various embodiments relating to medical devices, systems, clinical decision support tools, and related methods. These embodiments may include the development and implementation of clinical decision support tools using artificial intelligence. In some embodiments, artificial intelligence models may be trained using simulated data and one or more physiological modifiers.

[0011] In accordance with one particular embodiment, a system is provided comprising: at least one computing device configured to: receive a first set of waveform data from a patient’s aorta and a first set of waveform data from a patient’s left ventricle; analyze the first set of waveform data from the patient’s aorta and the first set of waveform data from the patient’s left ventricle using an artificial intelligence (Al) model trained using simulated data andAttorney Docket XMT00017USP1physiological modifiers; and provide a recommendation regarding a transaortic valve replacement procedure (TAVR) in the form of a first clinical decision support tool (CDST).

[0012] In one embodiment, the at least one computing device is further configured to: receive a second set of waveform data from the patient’s aorta and a second set of waveform data from the patient’s left ventricle; analyze the second set of waveform data from the patient’s aorta and the second set of waveform data from the patient’s left ventricle using an artificial intelligence (Al) model that was trained using simulated data and physiological modifiers; and provide a recommendation regarding repositioning a valve in a patient’s heart in the form of a second CDST.

[0013] In one embodiment, the second CDST is based on a recommendation relating to a paravalvular leak.

[0014] In one embodiment, the physiological modifier includes at least one of vascular resistance, compliance, cardiac preload, contractility, and heart rate.

[0015] In one embodiment, the Al model is also trained using historical data.

[0016] In one embodiment, the first CDST is based on a recommendation relating to stenosis in a patient’s aortic valve.

[0017] In one embodiment, the first CDST is based on a recommendation relating to regurgitation through a patient’s aortic valve.

[0018] In one embodiment, the system further comprises a guidewire having a first sensor providing the first set of waveform data from a patient’s left ventricle and a second sensor providing the first set of waveform data from a patient’s aorta.

[0019] In one embodiment, the system further comprises a device coupled with the guidewire configured to receive the first set of waveform data from the patient’s left ventricle from the first sensor and the first set of waveform data from the patient’s aorta from the second sensor.

[0020] In one embodiment, the device is configured for wireless communication with the computing device.

[0021] In another particular embodiment, a system is provided which comprises: a guidewire having a pair of pressure sensors in a distal section of the guidewire; a control unit selectively coupled with a proximal end of the guidewire; an external computing device in wireless communication with the control unit, the external computing unit coupled with aAttorney Docket XMT00017USP1network. The external computing unit, with or without assistance from the network, is configured to: receive, from the guidewire, a first set of waveform data from a patient’s aorta and a first set of waveform data from a patient’s left ventricle; analyze the first sets of waveform data using an artificial intelligence (Al) model trained using simulated data and physiological modifiers; and provide a recommendation regarding a transaortic valve replacement procedure (TAVR) in the form of a clinical decision support tool (CDST).

[0022] In accordance with another embodiment, a method is provided which comprises: obtaining a first set of waveform data from a patient’s aorta and a first set of waveform data from a patient’s left ventricle; and analyzing the first set of waveform data from the patient’s aorta and the first set of waveform data from the patient’s left ventricle using an artificial intelligence (Al) model trained using simulated data and physiological modifiers.

[0023] In one embodiment, the physiological modifiers include at least one of vascular resistance, compliance, cardiac preload, contractility, and heart rate.

[0024] In one embodiment, analyzing the first set of waveform data from the patient’s aorta and the first set of waveform data from the patient’s left ventricle includes determining whether there is stenosis in a patient’s aortic valve.

[0025] In one embodiment, analyzing the first set of waveform data from the patient’s aorta and the first set of waveform data from the patient’s left ventricle includes determining a severity of stenosis in the patient’s aortic valve.

[0026] In one embodiment, analyzing the first set of waveform data from the patient’s aorta and the first set of waveform data from the patient’s left ventricle includes determining whether there is regurgitation through a patient’s aortic valve.

[0027] In one embodiment, the method further comprises deploying an artificial valve within the patient’s heart.

[0028] In one embodiment, the method further comprises: subsequent deploying the artificial valve, obtaining a second set of waveform data from a patient’s aorta and a second set of waveform data from a patient’s left ventricle; and analyzing the second set of waveform data from the patient’s aorta and the second set of waveform data from the patient’s left ventricle using an artificial intelligence (Al) model trained using simulated data and physiological modifiers.Attorney Docket XMT00017USP1

[0029] In one embodiment, analyzing the second set of waveform data from the patient’s aorta and the second set of waveform data from the patient’s left ventricle includes determining the existence of a paravalvular leak.

[0030] In one embodiment, analyzing the second set of waveform data from the patient’s aorta and the second set of waveform data from the patient’s left ventricle includes determining a severity of the paravalvular leak.

[0031] Other embodiments are described hereinbelow, and features, elements, or components of one embodiment may be combined with features, elements, or components of other embodiments without limitation.BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The foregoing and other advantages of the invention will become apparent upon reading the following detailed description and upon reference to the drawings in which:

[0033] FIG. 1 illustrates a guidewire system according to an embodiment of the present disclosure;

[0034] FIG. 2 is an enlarged, detail view of a distal end of a guidewire in accordance with an embodiment of the present disclosure;

[0035] FIG. 3 is a partial cross-sectional view of a sensor mounting in a portion of a guidewire in accordance with an embodiment of the present disclosure;

[0036] FIG. 4 is an enlarged, detail view of a proximal end of a guidewire in accordance with an embodiment of the present disclosure;

[0037] FIG. 5 is a plan view of a controller removably coupled with a proximal portion of the guidewire in accordance with an embodiment of the present disclosure;

[0038] FIG. 6 is an enlarged, detail view of a distal end of a guidewire in accordance with another embodiment of the present disclosure;

[0039] FIG. 7 depicts a distal portion of a guidewire disposed within a patient’s heart in accordance with an embodiment of the present disclosure;

[0040] FIG. 8 is a flow diagram illustrating a method in accordance with an embodiment of the present disclosure;

[0041] FIG. 9 shows representative systolic and diastolic waveforms with certain portions highlighted;Attorney Docket XMT00017USP1

[0042] FIG. 10 is a flow diagram showing a process of preparing a trained model, and using the model real-time to provide a clinical decision support tool.DETAILED DESCRIPTION

[0043] Various embodiments described herein are directed toward the incorporation of electronic devices (e.g., sensors and transducers) into medical devices, systems incorporating such medical devices, and related methods.

[0044] In some embodiments, devices associated with cardiovascular, neurovascular, and endovascular procedures are provided having sensors integrated therewith. For example, guidewires or catheters may include sensors, transducers or other electronic or optical components integrated into the structure for detecting, imaging or measuring physiological data (e.g., pressure, flow rate, etc.), providing imaging data (e.g., ultrasound images), and providing that data to a healthcare provider in real time during an associated procedure.

[0045] In some embodiments, other sensors or electronic elements are associated with the device. For example, sensors configured to detect the presence of biological components may be incorporated into or otherwise associated with the device. In some embodiments, a transceiver unit having an antenna structure may be associated with the device for enabling wireless transmission of data.

[0046] Referring to FIG. 1, a guidewire system 100 is illustrated according to an embodiment of the present disclosure. As shown, the guidewire system 100 includes a guidewire 102, a proximal device which, in some embodiments, may include a control unit 104 for providing power to, and communication with, sensors or other electronic or optical components associated with the guidewire 102. The system 100 further includes a plurality of sensors 106 (see, e.g., FIGS. 2 and 3) associated with a distal end of the guidewire 102. The control unit 104 may include, for example, a power source (e.g., a battery), a data signal processor, a memory device, and a transmitter / receiver (referred to herein as a transceiver). In some embodiments, such components may be disposed, entirely or partially, within a body or housing of the control unit 104.

[0047] The system 100 may further include an external computing device 110 (also referred to as a “hub”). The external computing device 110 may include, e.g., a stationary or handheld computer, a stationary or handheld display, a tablet computer, a smart phone, or otherAttorney Docket XMT00017USP1input and / or output device. In one embodiment, as depicted in FIG. 1, the external computing device 110 may be in wireless communication with the control unit 104. Any of a variety of wireless protocols may be utilized (e.g., Bluetooth, Zigbee, Wi-Fi, etc.).

[0048] The system may further include a monitor 112, or a “boom” used by medical personnel during a procedure to review data, images, and other information relating to the procedure (e.g., waveforms associated with pressures in the aorta and in the ventricle). The monitor 112 may be in wired or wireless communication with the external computing device 110 so as to display information obtained by the sensors 106 or other electronic or optical components associated with the guidewire 102. For example, if the guidewire were being used in association with a transcatheter aortic valve implantation (or replacement) - TAVI or TAVR - procedure, the sensors 106 may detect or determine a first pressure in the left ventricle and a second pressure in the aorta. Each of the sensors 106 may provide a signal representative of the obtained pressures to the control unit 108, which in turn relays them (either as they are or as a modified signal) to the external computing device 110. The external computing device 110 then relays the pressure data to the monitor 112 and presents it in a recognizable form (e.g., as number and / or as a wave form) so that an interventional cardiologist may review the sensed pressures and determine if a valve replacement is necessary - or when measuring after the initial placement of a new valve, determine if subsequent actions need to be taken (such as reseating the valve to eliminate or reduce regurgitation).

[0049] The external computing device 110 is also in communication with one or more networks 114. In one embodiment, the network 114 may include the hospital’s (or other healthcare facility’s) computing system for access to, for example, electronic healthcare records (EHRs) which may be relevant to the current procedure. Access to such information may be beneficial, for example, to consider a specific patient’s health history as it pertains to the instant procedure. In other embodiments, the network 114 may include a plurality of computing devices similar to the external computing device 110, which may distributed throughout a hospital, throughout numerous hospitals or healthcare facilities, or throughout the world. In some embodiments, a network 114 may additionally include personal or mobile computing devices, such as phones, tablets, or notebook computers.

[0050] Further, the external computing device 110 may be in communication with a global cloud 116 or database having information relating to the instant procedure. The hospitalAttorney Docket XMT00017USP1computing system 114 and the cloud 116 may each be in communication with the external computing device 110 through wireless or through wired connections.

[0051] In one example, the cloud 116 may contain information relating to similar procedures including information relating to individuals in a similar demographic as the patient undergoing a particular procedure, their response to different interventions, their pressure or flow rates during a similar procedure, and other relevant data. The global cloud 116 may include computing ability to implement machine learning (or artificial intelligence) to apply the information within the global cloud to a specific procedure in light of the data being collected during the procedure. For example, a comparison of a pressure curve / waveform associated with the aortic pressure with the pressure curve / waveform of the left ventricular pressure may yield an index that is useful in determining whether an interventional act is required. In some embodiments, the index may be based strictly on a direct comparison of such pressures.However, in some embodiments, the index may be based on a dynamic analysis of the pressure curves, the past health history of the patient (e.g., as obtained from the EHR), data associated with the diagnosis and procedure outcomes of other individuals that may satisfy certain health and / or demographic criteria (e.g., age, race, weight, other diagnosed conditions, etc.). Thus, the index can be a dynamic tool to more accurately determine actions to be taken (or not taken) during a specific procedure as the procedure is being conducted as will be discussed in further detail below. Such analysis may be performed, for example, in the cloud 116, on the network 114, on the external computing device 110, or using some combination thereof.

[0052] It is noted that the system 100 may be defined to include certain basic elements (e.g., the guidewire 102 including its sensors 106, the control unit 104, and the external computing device 110), or it may be defined to include additional elements including the network 114 and / or the global cloud 116). The system 100 may additionally include other components including those conventionally found in a catheterization lab, such as the monitor 112, a patient bed 118, and / or an imaging device 120 for providing CT, X-ray, fluoroscopy, or other imaging information during the procedure. It is noted that the monitor 112 may be coupled with the imaging device 120 and may be configured to show imaging information as well as physiological information e.g., waveforms) and that such information may be displayed simultaneously or individually as selectively determined by a practitioner.Attorney Docket XMT00017USP1

[0053] Referring briefly to FIG. 2, a distal section of the guidewire 102 is depicted. The distal section includes a curved e.g., spiraled) or coiled section 130, sometimes referred to as a “pigtail”, that is configured to engage with a portion of the patient’s anatomy, effectively anchoring the guidewire in a desired position within a patient’s heart in an atraumatic fashion. The portion proximal of the coiled section 130 further includes two or more sensors 106 that are longitudinally spaced along a length of the guidewire 102. In one embodiment, the sensors 106 may be configured as pressure sensors (e.g., piezoelectric or capacitive-type pressure sensors). In one embodiment, the distal most sensor 106 A may be positioned at a location a distance Di taken from a tangent line of the proximal most portion of the coiled section 130 and which extends perpendicular to the length of the guidewire 102 as indicated in FIG. 2. In one embodiment, the distance Di may be between approximately 1 cm and approximately 3.5 cm. In one particular embodiment, the distance Di may be between approximately 2 cm and approximately 2.5 cm. Such measurements referred to above being measured from the tangent line to the center of the distal -most sensor 106 A as depicted in FIG. 2.

[0054] The sensors are spaced apart a distance “D2” so that one may be positioned in a patient’s left ventricle while the other is positioned within the patient’s aorta. In one embodiment, distance D2 may be approximately 9 centimeters (cm) apart. In one embodiment, distance D2 may be approximately 10 cm apart. In one embodiment, distance D2 may be approximately 11 cm apart. In another embodiment, distance D2 may be between approximately 8 cm and approximately 11 cm apart. In another embodiment, distance D2 may be between approximately 7 cm and approximately 12 cm apart. Such measurements referred to above being measured from the center of one sensor 106A to the center of the adjacent sensor 106B as depicted in FIG. 2.

[0055] In one embodiment, the distance Di may be approximately 2.5 cm while the distance D2may be approximately 10 cm. In another embodiment, the distance Di may be approximately 2 cm while the distance D2 may be approximately 10 cm. In another embodiment, distance Di may be approximately 2.5 cm, and distance D2may be approximately 11 cm.

[0056] Referring now to FIG. 3, a cross-sectional view of a sensor 106 mounted in the guidewire 102 is provided. The guidewire 102 includes a core wire 140 which may be formed of a metallic material such as, for example, stainless steel, titanium, nickel-titanium, or anotherAttorney Docket XMT00017USP1suitable metal or metal alloy. The core wire 104 has a portion removed to form a pocket 142 or a void. The pocket 142 may be formed by machining, laser ablation, or other appropriate manufacturing techniques. Next to the pocket 142, a shelf or a stepped region 144 is formed within the core wire 140. A first portion of the sensor 106 is attached to the stepped region 144 (e.g., such as by adhesive) such that another portion of the sensor 106 is cantilevered into the pocket 142, leaving a gap or a space between the underside of the cantilevered sensor portion and the bottom of the pocket 142 formed in the core wire 140.

[0057] A housing 146 is positioned over the core wire 104, the sensor 106 and the pocket 142. The housing 146 may be formed of a metallic material such as, for example, stainless steel, titanium, nickel-titanium, or another suitable metal or metal alloy. In one embodiment, the housing exhibits a longitudinal length “L” of approximately 0.5 cm. In other embodiments, the length L may be between approximately 1.0 cm and 0.25 cm. The housing 146 may help maintain the position of the sensor 106 or otherwise secure the sensor 106 to the core wire 104. The housing 146 additionally provides support to the core wire 104 in the region where the pocket 142 and stepped region have been formed such that the core wire 140 may withstand bending forces applied in that region of the core wire 140 when the guidewire 102 is being navigated through a tortuous path of a patient’s anatomy. An opening 147 is formed in the housing 146 to provide fluid communication between the pocket 142 and the external environment in which the guidewire 102 is placed. Thus, for example, if the guidewire is positioned such that the sensor 106 is located within a patient’s aorta, the pocket is in fluid communication with the blood that is flowing within the aorta, enabling the sensor 106 to determine the blood pressure at that location.

[0058] Another material layer 148 may be positioned about the core wire 140 in locations adjacent to the housings 146 to provide a common diameter and provide a smooth outer surface for the guidewire 102 along its longitudinal extent and minimize or eliminate any abrupt transitions that might otherwise occur (e.g., a stepped transition that might occur along the length of the core wire 140 and the housings 146). The material layer 148 may include, for example, a polymer material such as polyimide. In some embodiments, the material layer 148 may extend over the housings 146, while still leaving an opening (e.g., associated with opening 147) for fluid communication into the pocket 142.Attorney Docket XMT00017USP1

[0059] The configuration of the cantilevered sensor 106 within the pocket enables the sensor 106 to avoid or minimize inaccurate pressure readings that might otherwise be induced by the bending of the core wire 140. In other words, the bending of the core wire 140 at the location of the sensor 106 does not subject the sensor to a false reading because the cantilevered portion (which may either completely or substantially include the pressure detecting portion of the sensor, such as an associated membrane) is “free” from the bending (which might occur at the attached portion of the sensor 106) and does not register such bending forces as it would if the entire sensor were adhered or otherwise attached to the core wire 140.

[0060] The resulting guidewire 102 may have a size such that the outer diameter (e.g., after application of other outer members and / or coatings) is about 0.008 inches to about 0.040 inches, though larger or smaller sizes may also be utilized depending on particular application needs. For example, particular embodiments may have outer diameter sizes corresponding to standard guidewire sizes such as approximately 0.010 inches, 0.014 inches, 0.016 inches, 0.018 inches, 0.024 inches, 0.035 inches, 0.038 inches, or other such sizes common to guidewire devices.

[0061] While the guidewire 102 has been primarily described as including pressure sensors, it is noted that other sensors may be used in addition to such pressure sensors or in the alternative of such pressure sensors. For example, the sensors may additionally, or alternatively, be configured to determine flow rate or to sense the presence of biological components or measure physiological parameters in the targeted anatomical location (e.g., in the blood). Example biological components that may be detected / measured include sugar levels, pH levels, CO2 levels (CO2 partial pressure, bicarbonate levels), oxygen levels (oxygen partial pressure, oxygen saturation), temperature, and other such substrates and physiological parameters. The one or more sensors may be configured to sense the presence, absence, or levels of biological components such as, for example, immune system-related molecules (e.g., macrophages, lymphocytes, T cells, natural killer cells, monocytes, other white blood cells, etc.), inflammatory markers (e.g., C-reactive protein, procalcitonin, amyloid A, cytokines, alpha- 1 -acid glycoprotein, ceruloplasmin, hepcidin, haptoglobin, etc.), platelets, hemoglobin, ammonia, creatinine, bilirubin, homocysteine, albumin, lactate, pyruvate, ketone bodies, ion and / or nutrient levels (e.g., glucose, urea, chloride, sodium, potassium, calcium, iron / ferritin, copper, zinc, magnesium, vitamins, etc.), hormones (e.g., estradiol, follicle-Attorney Docket XMT00017USP1stimulating hormone, aldosterone, progesterone, luteinizing hormone, testosterone, thyroxine, thyrotropin, parathyroid hormone, insulin, glucagon, cortisol, prolactin, etc.), enzymes (e.g., amylase, lactate dehydrogenase, lipase, creatine kinase), lipids (e.g., triglycerides, HDL cholesterol, LDL cholesterol), tumor markers (e.g., alpha fetoprotein, beta human chorionic gonadotrophin, carcinoembryonic antigen, prostate specific antigen, calcitonin), and / or toxins (e.g., lead, ethanol).

[0062] Referring now to FIGs. 4 and 5, a proximal end of the guidewire 102 and a control unit 104 for attachment to the proximal end of the guidewire 102, respectively, are shown. As shown in FIG. 4, a proximal end of the guidewire 102 may include a plurality of electrodes or electrical contacts 160A-160E. The contacts 160A-160E are configured to make electrical connection with corresponding contacts or electrodes (not shown) in the control unit 104. In the embodiment shown, each of the contacts 160A-160E may be electrically coupled with wires or conductors that extend a length of the guidewire 102 and that are, in turn, electrically coupled with the sensors 106 in the distal section of the guidewire 102.

[0063] For example, two separate trifilar windings may be used to connect the sensors 106 to the contacts 160A-160E. In one embodiment, a first electrode 160A may be coupled with a power connection of a first sensor 106A using a first strand of a first trifilar winding, a second electrode 160B may be coupled with a data connection of a first sensor 106A using a second strand of a first trifilar winding, and a common connection may be made between the first sensor 106 A and a third electrode 160C using a third strand of the first trifilar winding. Additionally, a fourth electrode 160D may be coupled with a power connection of a second sensor 106B using a first strand of a second trifilar winding, a fifth electrode 160E may be coupled with a data connection of a second sensor 106B using a second strand of a second trifilar winding, and a common connection may be made between the second sensor 106B and the third electrode 160C using a third strand of the second trifilar winding. In another embodiment, instead of using trifilar windings to provide an electrical connection between the sensors 106 to the contacts, a flexible circuit may be used such as is described in U.S. Provisional Application No.63 / 712,175, entitled MEDICAL DEVICES AND RELATED METHODS, filed on Oct. 25, 2024, the disclosure of which is incorporated by reference herein in its entirety.

[0064] The guidewire 102 may additionally include a keyed or locking feature 162 in its proximal section. The locking feature 162 may include a shoulder or reduced diameterAttorney Docket XMT00017USP1section configured for engagement with a locking structure or mechanism 164 located within a housing 166 of the control unit 104 (see FIG. 5). The locking feature 162 and locking mechanism 164 work together to retain the guidewire 102 in a desired longitudinal position relative to the control unit 104 during use of the guidewire.

[0065] As indicated in FIG. 5, the guidewire 102 may be coupled with the control unit 104 by sliding the guidewire 102 in a direction parallel to its length (as indicated at 168) through an opening 170 formed in a surface of the housing 166 (FIG. 5 depicts the guidewire inserted into the housing as indicated by dashed lines). The control unit 104 (in addition to components and features previously mentioned) may include input features 172 (e.g., buttons, sliders, touchpads, directional pads, switches, etc.) to provide control of or communication with the sensors 106, transmission of data, and / or control of external components such as the external computing device 110 or monitor 112 (see FIG. 1). Additionally, the control unit 104 may include output features 174 (e.g., lights, screens, audio speakers, etc.) to provide feedback regarding operational status of the guidewire 102 or other components of the system 100.

[0066] Referring now to FIG. 6, the distal end of a guidewire 102 is depicted in accordance with another embodiment of the present disclosure. The guidewire is generally constructed similarly to that which has been described previously herein. The guidewire 102 may include a pair of spaced apart sensors 106 A and 106B used, for example, to simultaneously determine pressure in a left ventricle and an aorta, respectively, of a patient. The guidewire may additionally include an imaging sensor 190 longitudinally disposed between the two sensors 106A and 106B. In one embodiment, the imaging sensor may include an ultrasound transducer (or an array of ultrasound transducers) that enables imaging of for example, the aortic valve. In one example, the imaging sensor may be used to determine, for example, the size of a replacement valve that may need to be used. In another example, the imaging sensor may be used in an effort to determine whether there is stenosis in the valve, whether there is regurgitation, or whether some other condition exists within a patient’s anatomy. In other examples, the imaging sensor may be used to determine whether a new valve has been suitably positioned or whether regurgitation exists after initial placement. In one embodiment, such an imaging sensor 190 may include one or more transducers such as described in PCT Patent Application No. PCTUS2023 / 022337, entitled CMUT MEDICAL DEVICES, FABRICATION METHODS, SYSTEMS AND RELATED METHODS, and filed on May 16, 2023, theAttorney Docket XMT00017USP1disclosure of which is incorporated by reference herein in its entirety. In another embodiment, the imagine sensor 190 may be constructed in a manner such as described in PCT Patent Application Publication No. WO2023 / 196559 entitled MEDICAL DEVICES, SENSORS FOR MEDICAL DEVICES AND RELATED METHODS, and filed on April 7, 2023, the disclosure of which is incorporated by reference herein in its entirety. In one embodiment, the imaging sensor 190 may be positioned approximately an equal distance from each of the pair of sensors 106 A and 106B. In other embodiments, the imaging sensor 190 may be offset towards one sensor (106A) or the other sensor (106B).

[0067] In other embodiments, different numbers of and types of sensors may be used. For example, in some embodiments, ten or more pressures sensors may be used to determine a pressure gradient within a portion of a patient’s vasculature. In some embodiments, one or more pressure sensors may be combined with one or more flow sensors to determine multiple physiological variables. In some of these embodiments, a different type of atraumatic tip may be provided on the guidewire (as compared to the coil 130 shown in FIG. 2) as will be appreciated and recognized by one having ordinary skill in the art.

[0068] Referring now to FIG. 7, the distal section of the guidewire 102 is shown to be positioned within a patient’s anatomy such that a first sensor 106 A is located within the patient’s left ventricle 150 and a second sensor 106B is positioned within the patient’s aorta 152. As noted above, this enables pressure readings (or pressure wave forms) to be recorded simultaneously at both locations to determine, for example, the level of regurgitation that a patient is experiencing across the aortic valve. The pressure readings may be displayed on a monitor 112 (FIG. 1) for a practitioner to review and analyze. The guidewire 102, thus, may be used to help in the diagnosis of a patient’s condition, but also used after the diagnosis to guide a catheter delivering a replacement valve to the target location if needed or desired. Subsequent the delivery and placement of a new valve 154 (shown in dashed lines), the guidewire 102 may be again used to measure pressures in the left ventricle and the aorta to determine if the valve implantation / replacement was successful, or if repositioning of the valve may be desirable. Thus, the guidewire 102 may be used throughout the valve replacement procedure without removal of the guidewire 102 until the practitioner is satisfied with the placement and securement of the new valve.Attorney Docket XMT00017USP1

[0069] Referring now to FIG. 8 in association with FIGS. 1, 2 and 7, a flow diagram associated with such a procedure 200 is shown. As indicated at 202, a guidewire (or other elongated devices) is positioned within a patient’s heart. At 204, pressures (or waveforms showing pressure in real time) within the aorta and within the left ventricle are recorded using pressure sensors located on the guidewire. At 206, the pressure / waveform data obtained from the pressure sensors may be utilized in association with an artificial intelligence model to conduct analysis on the waveform data. At 208, a clinical decision support tool (CDST) is provided to assist a physician determine the existence of, and severity of, aortic regurgitation (AR), aortic stenosis (AS), or both. The CDST may provide information as a binary recommendation (recommending the placement of a new valve, or recommending against placement of a new valve), or it may provide a graded assessment (e.g., mild, moderate, severe), or provide the information in some other form (e.g., as a percentage of leak, Regurgitant Orifice Area (ROA)) for either or both AR and AS. In some embodiments, act 208 may include providing information to the physician in the form of one or more known indices. For example, any of the following indices could be utilized:

[0070] The Aortic Regurgitation Index (ARI) which is defined as ((DBP-LVEDP) / SBP) X 100; where DBP is the diastolic blood pressure, LVEDP is the left ventricle end-diastolic pressure, and SBP is the systolic blood pressure.

[0071] The Time Integrated Aortic Regurgitation Index (TIARI) which is defined as ((DAD / DD) / (LVSA / SD)) X 100; wherein DAD is the diastolic area difference, DD is the diastolic duration, LVSA is the left ventricle systolic area, and SD is the systolic duration (see FIG. 9 regarding these values).

[0072] The Diastolic Pressure Time index Adjusted for Systolic Blood Pressure (DPTIadj) which is defined as (DPTI / SBP) X 100; wherein DPTI is the diastolic pressure time index (the area under the diastolic portion of the pressure waveform.

[0073] The Dicrotic Notch Index (DNI) which is defined as ((SBP-DNP) / PP); wherein DNP is the dicrotic notch pressure, and PP is the pulse pressure (the difference between the systolic blood pressure and the diastolic blood pressure).

[0074] The Dicrotic Aortic Regurgitation Index (DARI) which is defined as ((SBP -LVEDP) / (DNP - DBP)).Attorney Docket XMT00017USP1

[0075] In addition to the indices noted above, other regurgitation indices, as well as stenosis indices may be calculated or considered in providing a CDST at act 208.

[0076] At 210, a valve is positioned, such as in a transaortic valve replacement (TAVR) procedure (sometimes referred to as transaortic valve implantation (TAVI)) using a catheter that is guided to the valve location by way of the guidewire (e.g., the guidewire is disposed within a lumen of the catheter to guide it to the desired location). The TAVR / TAVI procedure may use any known replacement valve.

[0077] After placement of the valve, updated waveform data is obtained from within the aorta and within the left ventricle, using the guidewire 102, as indicated at 212.

[0078] At 214, the pressure / waveform data obtained from the pressure sensors is utilized in association with an artificial intelligence model to conduct analysis on the waveform data. At 216, a clinical decision support tool (CDST) is provided to assist a physician determine the existence of, and severity of, any paravalvular leak PVL).

[0079] The CDST may provide information as a binary recommendation (recommending the repositioning of the valve, or recommending against repositioning of the valve), or it may provide a graded assessment (e.g., mild, moderate, severe), or provide the information in some other form (e.g., as a percentage of leak, Regurgitant Orifice Area (ROA)). In some embodiments, act 216 may include providing information to the physician in the form of one or more known indices, including any of the indices described above, or in any of the other forms discussed herein. Another index that could be used in consideration of PVL analysis, includes the following:

[0080] The ARI Ratio, which is defined as ARIpost / ARIpre; wherein ARIpost is the ARI measured post placement of a valve, and ARIpre is the ARI measure prior to placement of the valve.

[0081] At 218, a determination is made whether to make any adjustments to the valve. This determination may be made by comparing pressures taken pre- and post-placement of the valve, including comparison of actual pressures, comparison of pressure ratios (e.g., a ratio of aortic pressure to ventricular pressure - or the inverse), a comparison to data in the cloud -which may include the use of machine learning or artificial intelligence to compare like data and previously documented outcomes. If adjustments are desired, adjustments are performed as indicated at 220 and the procedure can return to the act of obtaining updated waveform data asAttorney Docket XMT00017USP1indicated at 212. This cycle may be repeated as necessary until the CDST regarding PVL indicates that the valve is sufficiently seated or that acceptable levels of PVL have been obtained, at which time the procedure may be completed as indicated at 222.

[0082] Regarding acts 206, 208, 214, and 216, various methods of machine learning (ML) or artificial intelligence (Al) may be utilized.

[0083] In some embodiments, the Al model can be developed by receiving 1) simulated data; 2) real-world, historic physiologic data; or both. Such data can include waveforms, measurements, or derived parameters. In one embodiment, simulated data may include data from a platform such as HARVI, which is a simulation-based environment for learning about cardiovascular physiology, hemodynamics and therapeutics (seehttps : / / 'harvi. online / site / welcome / )

[0084] Use of a platform such as HARVI enables the simulation of controlled physiological conditions to produce dual aortic and ventricular waveforms, which can be further manipulated to represent multiple disease states or procedural scenarios. Such a system integrates time-series hemodynamic waveforms generated under controlled conditions that replicate physiologic states, and its outputs can serve as a robust ground truth or training dataset for one or more machine learning models. This simulated data, in conjunction with real patient data, may be used to train, validate, and deploy one or more machine learning models that predict metrics relevant to clinical outcomes.

[0085] The Al model may use one or more of a variety of learning techniques. For example, it may use any one, or some combination of, the following:

[0086] Supervised Learning, wherein the model is trained on labeled examples of waveform data and corresponding clinical labels (e.g., severity of PVL or recommended reexpansion). Potential approaches using this technique include: Neural Networks (fully connected, convolutional, or recurrent architectures); Tree-Based Methods (random forest, gradient-boosted trees, or decision trees); Regression Models (linear, logistic, polynomial, or other variants); Support Vector Machines (SVMs);

[0087] Semi-Supervised and Unsupervised Techniques, wherein, large volumes of unlabeled simulated or real-world waveforms may be available. Techniques such as clustering (e.g., k-means, DBSCAN) or autoencoders (for unsupervised feature extraction) can be used to discover novel physiologic patterns that inform or supplement the predictive model.Attorney Docket XMT00017USP1

[0088] Ensemble Methods, wherein a combination of different algorithms (e.g., bagging, boosting, or stacking) may be employed to improve robustness and accuracy. The ensemble can involve multiple supervised learning models trained on different subsets or representations of the simulated data, thereby capturing a range of physiologic conditions.

[0089] Feature Engineering: Before or during model training, the system may extract a wide variety of features from the waveform data, such as amplitude, frequency, phase shifts, integrals, derivatives, or other domain-specific metrics. The machine learning (ML) pipeline may use traditional statistical feature selection, or automated feature engineering (e.g., via deep learning embeddings) to select the most relevant signals. In some embodiments, such may include a digital representation of the data that is not interpreted visually.

[0090] In some embodiments, the Training Workflow of the Al model may include the following:

[0091] 1. Data Generation and Labeling: a simulation platform produces simulated waveforms under different physiologic states (e.g., mild to severe PVL conditions), creating labeled examples for training.

[0092] 2. Data Preprocessing: The waveforms may undergo normalization, filtering, or segmentation to isolate critical periods (e.g., diastolic phase).

[0093] 3. Model Training and Validation: Using a variety of algorithms, one or more machine learning models are trained to learn the relationship between waveform characteristics and regurgitation severity or related clinical outcomes. In some embodiments, cross-validation or hold-out validation sets are employed for performance assessment.

[0094] 4. Model Selection and Tuning: Based on performance metrics (e g., accuracy, sensitivity, specificity, mean squared error), the system may automatically or manually choose the best-performing model. Hyperparameter tuning methods (e.g., grid search, Bayesian optimization) may be used to further refine model performance.

[0095] 5. Clinical Integration: Once a model is selected, it is integrated with the system to provide real-time or near-real-time predictions on new waveforms acquired during TAVR procedures, aiding clinicians in decision-making (e.g., whether to perform a re-expansion or repositioning of a newly implanted valve).

[0096] The combination of simulation data (which could include data that is preapproved by the Food and Drug Administration (FDA)) with advanced machine learningAttorney Docket XMT00017USP1techniques addresses potential data scarcity issues and ensures that the training process reflects clinically relevant conditions. The controlled nature of the simulated waveforms enables the generation of data representing a broad range of physiologic and pathological states. By capturing these variations, the system’s Al engine can robustly learn signals indicative of, for example, regurgitation severity, stenosis, or other valve dysfunction, resulting in improved accuracy and reliability of TAVR / TAVI procedure assessments - whether considering pre-TAVR diagnostics or post-TAVR adjustments.

[0097] Although the specific Al algorithms may vary, the system may provide optional interpretability features. For instance, the system can display confidence scores or highlight waveform features contributing most to the regurgitation severity estimate. This fosters clinician trust and satisfies emerging regulatory guidelines for transparent Al in medical devices. Clinicians can use the system’s resultant CDST alongside their own expertise, ensuring human oversight remains central in critical TAVR / TAVI decisions.

[0098] In another embodiment of the present disclosure, multi-modal data fusion may be used, wherein the Al engine may also incorporate imaging data (e.g., echocardiography, angiography) alongside simulated waveforms to improve predictive power.

[0099] In another embodiment of the present disclosure, adaptive learning may be employed, wherein the model may incorporate continuous learning, and where new simulated or real-world TAVR / TAVI data are periodically introduced to refine model parameters in a secure and controlled manner.

[0100] Referring to FIG. 10, a flow diagram is shown as an example of training an Al model and then using the Al model in conjunction with conducting a procedure such as TAVR / TAVI. As previously noted initial training may include the use of simulated data 250 (e.g., simulated waveforms) and historical data 252 (e.g., actual waveforms from patients and events with known conditions or states). Additionally, data relating to “physiological modifiers” 254 may be included. These physiological modifiers 254 may include various parameters that change when a valve is implanted (i.e., pre-TAVR state is different that post-TAVR state) -parameters that are not explicitly addressed using some of the indices discussed hereinabove.

[0101] For example, published indices are generally based on the time course of decay of aortic pressure and the time course of rise of the left ventricle pressure during diastole.However, the rate and extent of aortic pressure decay during diastole are dependent on vascularAttorney Docket XMT00017USP1resistance, compliance, and heart rate. Additionally, the rate and extent of LV pressure rise during diastole are dependent on changes of preload, contractility, and heart rate. All of these parameters may change from pre-to-post implantation of a valve, making the interpretation of any known index. The inclusion of data relating to the change of one or more of these physiological modifiers (e.g. vascular resistance, compliance, preload, contractility, and heart rate) significantly improves the recommendation of the CDST as the system predicts the response of a patient’s heart function to specific treatments.

[0102] All of this information may be used to develop the trained Al model 256. The trained Al model 256, along with patient waveforms 258 (e.g., waveforms recorded during a live procedure) may then be fed into a computing system for the performance of real-time Al analysis 260 specific to the patient’s condition, thereby providing a recommendation to the physician (e.g., to implant a valve or not, to reposition a recently implanted valve ore not, etc.) through the CDST 262.

[0103] In various embodiments, a remote or cloud-based implementation may be utilized. The computation for training and inference may be performed in the cloud, allowing for continuous updates and easier deployment across different clinical sites. In other embodiments, a distributed network may be used in real-time to develop the CDST. For example, with reference to FIG. 1, the hub 110 may be connected to a network 114 of individual hubs configured to conduct EDGE computing. The real-time analysis required to arrive at a CDST recommendation could be done using such a network (e.g., known as EDGE Al). In some embodiments, the computing may be carried out directly at the hub / extemal computing device 110.

[0104] By leveraging the controlled simulation data and / or the historical data in combination with predictors associated with the physiological modifiers (and in using a flexible suite of machine learning techniques) the resulting system will deliver a robust, explainable, and clinically relevant metric for TAVR / TAVI assessments. This approach will ensure broad adaptability to new data and evolving clinical practices while maintaining a foundational rigor grounded in validated physiologic simulations.

[0105] While the disclosed embodiments may be susceptible to various modifications and alternative forms, specific embodiments have been shown by way of example in the drawings and have been described in detail herein. It is noted that features, elements, orAttorney Docket XMT00017USP1components of one embodiment may be combined with features, elements, or components of other embodiments without limitation. However, it should be understood that the invention is not intended to be limited to the particular forms disclosed. Rather, the invention includes all modifications, equivalents, and alternatives falling within the spirit and scope of the invention as defined by the following appended claims.

Claims

Attorney Docket XMT00017USP1CLAIMSWhat is claimed is:

1. A system comprising:at least one computing device configured to:receive a first set of waveform data from a patient’s aorta and a first set of waveform data from a patient’s left ventricle;analyze the first set of waveform data from the patient’s aorta and the first set of waveform data from the patient’s left ventricle using an artificial intelligence (Al) model trained using simulated data and physiological modifiers; andprovide a recommendation regarding a transaortic valve replacement procedure (TAVR) in the form of a first clinical decision support tool (CDST).

2. The system of claim 1, wherein the at least one computing device is further configured to:receive a second set of waveform data from the patient’s aorta and a second set of waveform data from the patient’s left ventricle;analyze the second set of waveform data from the patient’s aorta and the second set of waveform data from the patient’s left ventricle using an artificial intelligence (Al) model that was trained using simulated data and physiological modifiers; and provide a recommendation regarding repositioning a valve in a patient’s heart in the form of a second CDST.

3. The system of claim 2, wherein the second CDST is based on a recommendation relating to a paravalvular leak.

4. The system of claim 1, wherein the physiological modifier includes at least one of vascular resistance, compliance, cardiac preload, contractility, and heart rate.

5. The system of claim 1, wherein the Al model is also trained using historical data.Attorney Docket XMT00017USP16. The system of claim 1, wherein the first CDST is based on a recommendation relating to stenosis in a patient’s aortic valve.

7. The system of claim 1, wherein the first CDST is based on a recommendation relating to regurgitation through a patient’s aortic valve.

8. The system of claim 1, further comprising a guidewire having a first sensor providing the first set of waveform data from a patient’s left ventricle and a second sensor providing the first set of waveform data from a patient’s aorta.

9. The system of claim 7, further comprising a device coupled with the guidewire configured to receive the first set of waveform data from the patient’s left ventricle from the first sensor and the first set of waveform data from the patient’s aorta from the second sensor.

10. The system of claim 8, wherein the device is configured for wireless communication with the computing device.

11. A system comprising:a guidewire having a pair of pressure sensors in a distal section of the guidewire; a control unit selectively coupled with a proximal end of the guidewire;an external computing device in wireless communication with the control unit, the external computing unit coupled with a network, and wherein the external computing unit, with or without assistance from the network, is configured to:receive, from the guidewire, a first set of waveform data from a patient’s aorta and a first set of waveform data from a patient’s left ventricle; analyze the first sets of waveform data using an artificial intelligence (Al) model trained using simulated data and physiological modifiers; and provide a recommendation regarding a transaortic valve replacement procedure (TAVR) in the form of a clinical decision support tool (CDST).Attorney Docket XMT00017USP112. A method comprising:obtaining a first set of waveform data from a patient’s aorta and a first set of waveform data from a patient’s left ventricle;analyzing the first set of waveform data from the patient’s aorta and the first set of waveform data from the patient’s left ventricle using an artificial intelligence (Al) model trained using simulated data and physiological modifiers.

13. The method according to claim 13, wherein the physiological modifiers include at least one of vascular resistance, compliance, cardiac preload, contractility, and heart rate.

14. The method according to claim 13, wherein analyzing the first set of waveform data from the patient’s aorta and the first set of waveform data from the patient’s left ventricle includes determining whether there is stenosis in a patient’s aortic valve.

15. The method according to claim 15, wherein analyzing the first set of waveform data from the patient’s aorta and the first set of waveform data from the patient’s left ventricle includes determining a severity of stenosis in the patient’s aortic valve.

16. The method according to claim 13, wherein analyzing the first set of waveform data from the patient’s aorta and the first set of waveform data from the patient’s left ventricle includes determining whether there is regurgitation through a patient’s aortic valve.

17. The method according to claim 13, further comprising deploying an artificial valve within the patient’s heart.

18. The method according to claim 17, further comprising:subsequent deploying the artificial valve, obtaining a second set of waveform data from a patient’s aorta and a second set of waveform data from a patient’s left ventricle;analyzing the second set of waveform data from the patient’s aorta and the second set of waveform data from the patient’s left ventricle using an artificial intelligence (Al) model trained using simulated data and physiological modifiers.Attorney Docket XMT00017USP119. The method according to claim 17, wherein analyzing the second set of waveform data from the patient’s aorta and the second set of waveform data from the patient’s left ventricle includes determining the existence of a paravalvular leak.

20. The method according to claim 19, wherein analyzing the second set of waveform data from the patient’s aorta and the second set of waveform data from the patient’s left ventricle includes determining a severity of the paravalvular leak.

Images