Pre-therapy sensing for neuromodulation therapy

Abstract

An example control device for performing a denervation procedure on a patient includes a memory and processing circuitry coupled to the memory. The processing circuitry is configured to receive, from one or more sensors and prior to delivery of neuromodulation therapy via a neuromodulation catheter, pre-therapy sensor data. The pre-therapy sensor data includes one or more parameter measurements for a period of time prior to the delivery of the neuromodulation therapy. The processing circuitry is configured to determine, based on the pre-therapy sensor data, whether the delivery of neuromodulation therapy would cause a deviation, in which the deviation includes one or more patient parameters or device parameters exceeding one or more patient parameter thresholds or device parameter thresholds, and transmit an output based on the determination.

Description

PRE-THERAPY SENSING FOR NEUROMODULATION THERAPY

[0001] This application claims the benefit of U.S. Provisional Patent Application Serial No. 63 / 664,102, filed June 25, 2024, the entire content of which is incorporated herein by reference.TECHNICAL FIELD

[0002] This disclosure generally relates to neuromodulation therapy.BACKGROUND

[0003] A denervation procedure may include ablating target nerves by delivering neuromodulation therapy, such as ablative energy, to a target treatment site via a neuromodulation catheter. For example, renal denervation may include ablating renal nerves. Responsiveness of a patient to the denervation procedure may depend on various factors related to the patient and the procedure, such as an anatomy of the patient and a position of the catheter in the patient.SUMMARY

[0004] In general, the disclosure describes systems and methods for delivering neuromodulation therapy that use one or more parameter measurements of the patient detected prior to delivering the neuromodulation therapy to inform or modify subsequent delivery of the neuromodulation therapy. A neuromodulation system includes a neuromodulation catheter and a control device coupled to the neuromodulation catheter. The neuromodulation catheter includes one or more therapy delivery elements, such as electrodes, for interfacing with a vessel of a patient and delivering the neuromodulation therapy to the vessel. The control device delivers a control signal to the neuromodulation catheter that causes the neuromodulation catheter to deliver the neuromodulation therapy according to various parameters.

[0005] In addition to causing delivery of the neuromodulation therapy, the control device receives pre-therapy sensor data detected prior to delivery of neuromodulation therapy and, in some instances, in-therapy sensor data detected during delivery of the neuromodulation therapy. The pre-therapy sensor data includes one or more parametermeasurements, such as impedance or temperature, for a period of time prior to the delivery of the neuromodulation therapy. The control device uses this pre-therapy sensor data to determine whether the delivery of neuromodulation therapy would cause one or more patient parameters or device parameters to exceed one or more patient parameter thresholds or device parameter thresholds (referred to herein as “deviations”). For example, the control device may utilize a model trained on pre-therapy sensor data and other patient data for a population of patients having undergone neuromodulation therapy, which may recognize characteristics of the pre-therapy sensor data of the patient that indicate the presence of conditions, such as poor electrode contact or a particular composition of adjacent tissues, that tend to cause the patient parameters or device parameters to exceed the patient or device parameter thresholds after neuromodulation therapy has begun.

[0006] The control device may transmit an output based on the determination of whether determine, based on the pre-therapy sensor data, whether the delivery of neuromodulation therapy would cause a deviation. For example, during delivery of neuromodulation therapy, the control device may monitor whether the in-therapy sensor data exceeds one or more safety or efficiency thresholds. Rather than wait for the detected in-therapy sensor data to exceed one of these interlocks, the control device may take an action, such as generating an indication of the patient or device threshold that would be exceeded or modifying subsequent delivery of the neuromodulation therapy, that can avoid or reduce the likelihood of the in-therapy sensor data exceeding the threshold. In this way, neuromodulation systems and methods described herein may more enable neuromodulation therapy in a reduced amount of time.

[0007] In one example, this disclosure is directed to a control device for performing a denervation procedure on a patient that includes a memory and processing circuitry coupled to the memory. The processing circuitry is configured to receive, from one or more sensors and prior to delivery of neuromodulation therapy via a neuromodulation catheter, pre-therapy sensor data. The pre-therapy sensor data includes one or more parameter measurements for a period of time prior to the delivery of the neuromodulation therapy. The processing circuitry is configured to determine, based on the pre-therapy sensor data, whether the delivery of neuromodulation therapy would cause a deviation, inwhich the deviation includes one or more patient parameters or device parameters exceeding one or more patient parameter thresholds or device parameter thresholds, and transmit an output based on the determination. In another example, the disclosure is directed to a neuromodulation system that includes a neuromodulation catheter and the control device described above.

[0008] In another example, this disclosure is directed to a method for performing a denervation procedure that includes receiving, by processing circuitry from one or more sensors and prior to delivery of neuromodulation therapy via a neuromodulation catheter, pre-therapy sensor data. The pre-therapy sensor data includes one or more parameter measurements for a period of time prior to the delivery of the neuromodulation therapy. The method further includes determining, by the processing circuitry and based on the pre- therapy sensor data, whether the delivery of neuromodulation therapy would cause a deviation, in which the deviation includes one or more patient parameters or device parameters exceeding one or more patient parameter thresholds or device parameter thresholds, and transmit an output based on the determination.

[0009] Further disclosed herein is a control device for performing a denervation procedure on a patient that includes a memory and processing circuitry coupled to the memory, wherein the processing circuitry is configured to receive, from one or more sensors and prior to delivery of neuromodulation therapy via a neuromodulation catheter, pre-therapy sensor data, wherein the pre-therapy sensor data includes one or more parameter measurements for a period of time prior to the delivery of the neuromodulation therapy, wherein the processing circuitry is configured to determine, based on the pre- therapy sensor data, whether the delivery of neuromodulation therapy would cause a deviation, in which the deviation includes one or more patient parameters or device parameters exceeding one or more patient parameter thresholds or device parameter thresholds, and transmit an output based on the determination.

[0010] Further details of one or more examples of this disclosure are set forth in the accompanying drawings and in the description below. Other features, objects, and advantages will be apparent from the description and drawings, and from the claims.

[0011] The above summary is not intended to describe each illustrated example or every implementation of the present disclosure.BRIEF DESCRIPTION OF DRAWINGS

[0012] FIG. 1 is a conceptual diagram illustrating an example neuromodulation system for responding to one or more predicted deviations during delivery of neuromodulation therapy using pre-therapy sensor data, in accordance with some examples of the current disclosure.

[0013] FIG. 2A is a block diagram illustrating an example control device configured to respond to one or more predicted deviations during delivery of neuromodulation therapy using pre-therapy sensor data, in accordance with some examples of the current disclosure.

[0014] FIG. 2B is a flow diagram illustrating an example technique for responding to one or more predicted deviations during delivery of neuromodulation therapy using pre- therapy sensor data, in accordance with some examples of the current disclosure.

[0015] FIG. 3 A is a block diagram illustrating an example computing device configured to generate a model for predicting one or more deviation during delivery of neuromodulation therapy using pre-therapy sensor data, in accordance with some examples of the current disclosure.

[0016] FIG. 3B is a flow diagram illustrating an example technique for generating a trained model for predicted deviations during delivery of neuromodulation therapy using pre-therapy sensor data, in accordance with some examples of the current disclosure.

[0017] FIG. 3C is a block diagram of a contrastive learning method for generating a trained model, in accordance with some examples of the current disclosure.

[0018] FIG. 4A is a graph of power delivery during a denervation procedure, in accordance with some examples of the current disclosure.

[0019] FIG. 4B is a graph of impedance over time prior to delivery of neuromodulation therapy, in accordance with some examples of the current disclosure.

[0020] FIG. 5 is a scatter plot of predicted deviations by the trained encoder, in accordance with some examples of the current disclosure.DETAILED DESCRIPTION

[0021] Denervation therapy, such as renal denervation (RDN) therapy, may be used to render a nerve inert, inactive, or otherwise completely or partially reduced in function,such as by ablation or lesioning of the nerve. Denervating an overactive nerve may provide a therapeutic benefit to a patient. For example, renal denervation may mitigate symptoms associated with renal sympathetic nerve overactivity, such as hypertension. Denervation therapy may include delivering electrical energy, thermal energy, and / or a chemical agent to a target nerve. In the case of renal denervation therapy, the denervation energy or agents can be delivered, for example, via a therapy delivery device (e.g., a neuromodulation catheter) disposed in a blood vessel (e.g., the renal artery) proximate to the renal nerve.

[0022] A significant proportion of individual neuromodulation therapy deliveries in denervation procedures, such as RDN procedures, may partially or wholly terminate due to an unexpected deviation in local conditions detected during the denervation procedure. For example, a control device of a neuromodulation system applying radiofrequency (RF) energy may receive in-therapy sensor data, such as temperature and impedance measurements, detect a treatment out of range (TOOR) condition existing on one or more electrodes, and react by prematurely terminating energy delivery on one or all of the electrodes. A clinician operating the neuromodulation catheter may continue to deliver energy on less than all the electrodes or may reposition the neuromodulation in an attempt to avoid the TOOR condition. Such delays may increase an operating time of the denervation therapy.

[0023] The present disclosure describes various aspects of techniques related to delivering neuromodulation therapy that uses one or more parameter measurements of the patient detected prior to delivering the neuromodulation therapy to inform or modify subsequent delivery of the neuromodulation therapy. In addition to receiving the intherapy sensor data, the control device receives pre-therapy sensor data detected prior to delivery of neuromodulation therapy. The pre-therapy sensor data includes one or more parameter measurements, such as impedance or temperature measurements, for a period of time prior to the delivery of the neuromodulation therapy. The control device uses this pre-therapy sensor data to determine a predicted deviation of the neuromodulation therapy.

[0024] To determine the predicted deviation, the control device may utilize a model trained on pre-therapy sensor data and other patient data for a population of patients having undergone neuromodulation therapy. The trained model may receive as input thepre-therapy sensor data, as well as other patient-specific data, which may include characteristics that indicate physiological conditions of the patient that tend to cause deviations after neuromodulation therapy has begun. For example, the pre-therapy sensor data may include impedance data having a signature waveform that reflects both cyclical deviations, such as heart rate and breathing, and non-cyclical deviations that may indicate an anomalous condition associated with a potential deviation during delivery of the neuromodulation therapy.

[0025] In response to the predicted deviation being outside a threshold, such as violating a TOOR condition described above, the control device may transmit an output based on the predicted deviation. For example, during delivery of neuromodulation therapy, the control device may monitor whether the in-therapy sensor data exceeds one or more safety or efficiency thresholds. Rather than wait for the detected in-therapy sensor data to exceed one of these interlocks, the control device may take an action that can avoid or reduce the likelihood of the in-therapy sensor data exceeding the threshold. Outputs may include informational outputs, such as an indication of the predicted deviation that can be viewed and responded to by a clinician, or automated outputs, such as a modification of subsequent delivery of the neuromodulation therapy that attempts to ameliorate the predicted deviation.

[0026] In this way, the example techniques improve the technology of neuromodulation therapy delivery by integrating in a practical application the detection of parameter measurements prior to delivering the neuromodulation therapy. For instance, as described in more detail below, determining prediction deviations, such as those determined by the trained model computations, may not be possible by the clinician for accurate prediction. With the example techniques described in this disclosure, a neuromodulation system configured to determine a predicted deviation of the neuromodulation therapy based on the pre-therapy sensor data and transmit an output based on the predicted deviation may enable delivery of neuromodulation therapy in a reduced amount of time as compared to other neuromodulation systems that do not perform the example techniques or clinician generated prediction.

[0027] FIG. 1 is a diagram illustrating an example neuromodulation system including a neuromodulation catheter 102 configured to deliver neuromodulation therapy, acomputing system 114, and a control device 112. Computing system 114 may be configured to send a control signal to control device 112 or otherwise control the general operation of one or both of catheter 102 or control device 112. In some examples, the control signal sent by computing system 114 to control device 112 is configured to cause control device 112 to, based on the control signal received from computing system 114, generate an electrical signal. Control device 112 may generate the electrical signal sent to catheter 102 to cause catheter 102 to deliver neuromodulation therapy. For example, the neuromodulation therapy may include renal neuromodulation, hepatic neuromodulation, or any other neuromodulation therapy.

[0028] In general, the devices, systems, and techniques described herein may be used to perform neuromodulation and stimulation from within any suitable anatomical lumen that has nerves adjacent to the anatomical lumen. Example anatomical lumens include the celiac trunk and its branches (including the common hepatic artery and its branches (including the gastroduodenal artery and its branches, the right gastric artery and its branches, and the proper hepatic artery and its branches), the left gastric artery and its branches, and the splenic artery and its branches), the superior mesenteric artery and its branches, the gonadal artery and its branches, the inferior mesenteric artery and its branches, and the like. Further, although the disclosure primarily describes neuromodulation from within one or more arteries, the devices, systems, and techniques of the disclosure also may be applied to neuromodulation from within one or more veins, such as a renal vein and its branches, a hepatic vein and its branches, an intercostal vein and its branches, or the like.

[0029] As shown in FIG. 1, system 100 includes a neuromodulation catheter 102, which includes a handle 104 and an elongated member 108 attached to handle 104. Elongated member 108 may have any suitable outer diameter, and the diameter can be constant along the length of elongated member 108 or may vary along the length of elongated member 108. In some examples, elongated member 108 can be 2, 3, 4, 5, 6, or 7 French or another suitable size.

[0030] Elongated member 108 includes a distal portion 108 A and a proximal portion 108b. Distal portion 108A includes one or more therapy delivery devices 110. In the example shown in FIG. 1, therapy delivery device 110 includes a plurality of therapydelivery elements, such as electrodes 106A, 106B, 106C, 106D, disposed along elongated member 108. Therapy delivery device 110 may be electrically connected to control device 112 via a plurality of electrical conductors disposed within elongated member 108 and / or handle 104. Control device 112 may, via the plurality of electrodes 106 of therapy delivery device 110, deliver therapeutic energy (e.g., RF energy) to and / or sense one or more parameters of tissue of a vessel wall of a blood vessel of the patient. Although FIG. 1 illustrates catheter 102 as having four electrodes 106, other example catheters may include greater or fewer electrodes 106.

[0031] Distal portion 108 A of elongated member 108 is configured to be advanced within an anatomical lumen of a human patient to locate therapy delivery device 110 at a target tissue site within or otherwise proximate to the anatomical lumen. For example, elongated member 108 may be configured to position therapy delivery device 110 within a blood vessel, a ureter, a duct, an airway, or another naturally occurring lumen within the human body. The examples described herein focus on the anatomical lumen being a blood vessel, such as a renal vessel, but it will be understood that similar techniques may be used with other anatomical lumens. In certain examples, intravascular delivery of therapy delivery device 110 includes percutaneously inserting a guidewire (not shown in FIG. 1) into a vessel of a patient and moving elongated member 108 and / or electrodes 106 along the guidewire until therapy delivery device 110 reach a target tissue site (e.g., a renal artery). For example, distal portion 108A of elongated member 108 may define a passageway for engaging the guidewire for delivery of therapy delivery device 110 using over-the-wire (OTW) or rapid exchange (RX) techniques. In other examples, neuromodulation catheter 102 can be a steerable or non-steerable device configured for use without a guidewire. In still other examples, neuromodulation catheter 102 can be configured for delivery via a guide catheter or sheath (not shown in FIG. 1), or other guide device.

[0032] Once at the target tissue site, therapy delivery device 110 can be configured to deliver therapy, such as RF energy, to provide or facilitate neuromodulation therapy at the target tissue site. For ease of description, the following discussion will be primarily focused on delivering RF energy, in which example therapy delivery device 110 includes a plurality of electrodes 106. Electrodes 106 may deliver RF energy to the tissue of thepatient. It will be understood, however, that in other examples, therapy delivery device 110 may include elements or structures configured to deliver other types of therapy in addition to or instead of RF energy such as, but not limited to, microwave energy or variations of the energy application such as gating or pulsing.

[0033] In the example shown in FIG. 1, therapy delivery device 110 is configured to assume a delivery configuration in which therapy delivery device 110 defines a relatively smaller radial extent (a relatively low profile), and a radially expanded configuration in which therapy delivery device 110 defines a relatively larger radial extent. Distal portion 108 A may be delivered through vasculature of the patient to the target tissue site while therapy delivery device 110 is in the delivery configuration. In some examples, in the radially expanded configuration, therapy delivery device 110 defines a helical, a spiral, a loop, a basket, or a stent-like configuration. In the radially expanded configuration, therapy delivery device 110 is configured to position one or more electrodes of the plurality of electrodes carried by therapy delivery device 110 near a vessel wall, e.g., in apposition to the vessel wall. In other examples, therapy delivery device 110 includes a balloon configured to expand from a relatively low profile delivery configuration to an expanded configuration in order to position one or more electrodes of the plurality of electrodes carried by therapy delivery device 110 near a vessel wall.

[0034] The plurality of electrodes 106 may be positioned along therapy delivery device 110 such that, when therapy delivery device 110 is in the radially expanded configuration, the electrodes 106 are spaced around an inner perimeter (e.g., circumference) of the vessel wall. In some examples, electrodes 106 may be positioned along therapy delivery device 110 such that, when therapy delivery device 110 is in the radially expanded configuration, electrodes 106 are substantially evenly spaced around an inner perimeter (e.g., circumference) of the vessel wall.

[0035] In some examples, catheter 102 is configured to deliver stimulation to a target site before, during, or after delivering neuromodulation therapy. For example, catheter may include a plurality of stimulation electrodes along elongate member 108 and / or may be configured to use electrodes 106 for stimulation. Stimulation prior to neuromodulation may be indicative of a potential effectiveness of delivery of the neuromodulation therapy, as will be described further below.

[0036] In some examples, the neural response to the stimulation, or the differences in neural responses to the stimulation after neuromodulation may be indicative of effectiveness of denervation. Thus, a clinician may alternate delivery of stimulation with one or more sessions of delivery of denervation therapy, to gauge the progress and effectiveness of the denervation therapy. For example, at least one physiological parameter (for example, a blood pressure, a flow rate, or constriction in vasculature) may be measured and compared before and after delivering a stimulation signal, and a change in a magnitude of the at least one physiological parameter may indicate attenuation of neural traffic associated with the nerve. For example, low change or an absence of change in the physiological parameter may indicate that the nerve is ablated and that no further ablation is necessary. Likewise, a change in the physiological parameter greater than a predetermined threshold may indicate that further ablation is required. Catheter 102 may thus deliver denervation therapy via the plurality of electrodes 106, and computing system 114 or control device 112 may monitor denervation by delivering stimulation via the plurality of electrodes 106 (or other sensing element).

[0037] Each electrode 106 may be coupled to a respective thermocouple. For example, control device 112 may sense a respective temperature of each electrode 106 via the respective thermocouple. In some examples, control device 112 may terminate denervation therapy in response to a temperature sensed by a thermocouple, for example, to avoid excessive heating of catheter 102 or neighboring tissue.

[0038] Catheter 102 may be communicatively coupled to control device 112 (e.g., a therapy delivery device). Control device 112 includes signal generation circuitry configured to generate and deliver RF energy to the patient via the one or more electrodes disposed on therapy delivery device 110. Control device 112 further includes sensing circuitry configured to sense one or more parameters (e.g., temperature, impedance, or the like) from the patient at or near a target tissue site via the one or more electrodes and / or sensors and adjust the RF energy (e.g., between monopolar energy and bipolar energy) based on the sensed parameters. For example, control device 112 can be configured to .

[0039] During a denervation procedure, a clinician may position distal portion 108 A at a target tissue site, such as by expanding distal portion 108 A into an expanded configuration. In the expanded configuration, therapy delivery device 110 may place atleast one electrode 106 at a first location relative to the vessel wall, for example, corresponding to a first rotational location. A clinician may control a therapy delivery device 110 to deliver, provide, or facilitate neuromodulation therapy at the target tissue site, for example, through the vessel wall at the target tissue site to target tissue adjacent to the blood vessel. The neuromodulation therapy may include, but is not limited to, radiofrequency (RF) energy, or the like. The clinician may rotate handle 104, or otherwise proximal portion 108B, to apply a torque to distal portion 108 A and cause therapy delivery device 110 to rotate about central longitudinal axis L from the first rotational location to a second rotational location. The clinician may control system 100 to deliver neuromodulation therapy or stimulation at the second rotational location, or after further successive rotational locations of therapy delivery device 110.

[0040] FIG. 2A is a block diagram illustrating an example control device 112 of FIG. 1. In this example, control device 112 includes signal generation circuitry 206, sensing circuitry 208, control circuitry 210, user interface 212, communications circuitry 214, memory 216, and a power source 218 that provides operational power to the other components. The various circuitry may be, or include, programmable or fixed function circuitry configured to perform the functions attributed to the respective circuitry.

[0041] Control device 112 is electrically connected to electrodes 106A-106D disposed on therapy delivery device 110 of catheter 102 (not pictured) via electrical conductors 204A-204D (also referred to as “electrical conductors 204”). In the example shown in FIG. 2A, each electrode 106A-106D is electrically coupled to a separate electrical conductor 204A-204D such that each electrode is independently and separately activatable. Electrical conductors 204 may have any suitable configuration, e.g., electrical wires, feedthrough assemblies, or the like extending inside handle 104 and elongated member 108 to therapy delivery device 110.

[0042] Signal generation circuitry 206 is configured to generate and deliver energy, e.g., in the form of an electrical signal, to the target tissue via neuromodulation catheter 102. Although RF energy and signals are primarily referred to herein, in other examples, the energy can be other types of energy and signals, such as, but not limited to, microwave energy. Signal generation circuitry 206 is configured to deliver the generated RF signals to the target tissue site through one or more selected electrodes 106 and respectiveelectrical conductors 204. Signal generation circuitry 206 may include, as examples, current or voltage sources, capacitors, charge pumps, or other signal generation circuitry. In some examples, signal generation circuitry 206 may receive sensed electrical signals from one or more of electrodes 106. Signal generation circuitry 206 may transmit the sensed electrical signals to sensing circuitry 208 and / or control circuitry 210.

[0043] Electrodes 106 are electrically connected to signal generation circuitry 206 of control device 112 through electrical conductors 204. In some examples, as illustrated in FIG. 2, each of electrodes 106 is separately electrically connected to signal generation circuitry 206 via a corresponding electrical conductor of electrical conductors 204. In other examples, electrodes 106 may be electrically connected, via electrical conductors 204, to a switching circuitry configured to selectively couple signal generation circuitry 206 and / or sensing circuitry 208 to selected combinations of electrodes 106.

[0044] Sensing circuitry 208 is configured to sense one or more parameters and / or receive one or more sensed parameters, such as temperature and / or impedance. In some examples, the sensed one or more parameters may be received from one or more sensors on neuromodulation catheter 102. For example, sensing circuitry 208 can be configured to sense the one or more parameters at or near the target tissue site via one or more sensors on catheter 102. In some examples, the one or more sensed parameters may be received from one or more sensors external to neuromodulation catheter 102. For example, sensing circuitry 208 can be configured to sense the one or more parameters at or near the target tissue site using another medical device, or may sense the one or more parameters at a site that is not at or near the target tissue site. Sensing circuitry 208 has any suitable configuration and may, for example, include filters, amplifiers, analog-to-digital converters, or other circuitry configured to sense electrical signals via electrodes 106 or to convert the sensed electrical signals to one or more parameters (e.g., temperature / change in temperature of tissue near selected electrodes of electrodes 106, impedance / change in impedance sensed by the selected electrodes 106). In some examples, sensing circuitry 208 includes one or more thermocouples (e.g., connected to each of electrodes 106 via a pair of corresponding wires, which can, but need not, include conductors 204). In these examples, sensing circuitry 208 provides signals based on which control circuitry 210 candetermine the temperature and / or change in temperature of tissue near the selected electrodes of electrodes 106 via the one or more thermocouples.

[0045] In some examples, sensing circuitry 208 is configured to sense one or more patient or device parameters during delivery of neuromodulation therapy as in-therapy sensor data. Patient or device parameters may include any parameters of the patient or catheter 102 that result from delivery of the neuromodulation therapy. The in-therapy sensor data may provide an indication as to an effectiveness or safety of the delivery of neuromodulation therapy. The one or more patient or device parameter measurements may include, but are not limited to, temperature measurements, such as may be detected by the thermocouples; impedance measurements, such as may be detected by electrodes 106; physiological measurements (e.g., blood pressure), such as may be detected by an external sensor; and any other measurements that may indicate a responsiveness of the patient to the neuromodulation therapy.

[0046] Sensing circuitry 208 is configured to sense one or more parameters prior to delivery of neuromodulation therapy as pre-therapy sensor data. The pre-therapy sensor data may provide an indication as to a potential effectiveness or safety of the delivery of neuromodulation therapy. For example, various localized conditions at the target site, such as contact between electrodes 106 and a wall of the vessel, may result in reduced effectiveness of delivery of the neuromodulation therapy, as indicated by the in-therapy sensor data collected during delivery. Absent the pre-therapy sensor data, these localized conditions may not be detected until the in-therapy sensor data exceeds a threshold, such as a TOOR condition. However, various parameter measurements taken prior to delivering the denervation therapy may provide an indication as to the localized conditions. For example, various characteristics of the parameter measurements may be correlated with deviations in expected patient or device parameters once delivery of the neuromodulation therapy has begun. These characteristics may be too subtle for a clinician or other operator to identify and associate with a potential deviation. The one or more parameters may include any parameter that may indicate a condition that may result in a deviation in delivery of neuromodulation therapy including, but not limited to, temperature measurements, such as may be detected by the thermocouples; impedance measurements, such as may be detected by electrodes 106; anatomical measurements (e.g.,vessel diameter), such as may be detected by a balloon or other medical device positioned in the vessel; pressure measurements, such as may be detected by the balloon or other medical device; ultrasound measurements, such as may be detected by an ultrasound transducer or other medical device; and the like.

[0047] In some examples, neuromodulation catheter 102 includes one or more sensors configured to capture the pre-therapy sensor data. For example, electrodes 106, thermocouples, or any other sensors attached to catheter 102 may detect the one or more parameter measurements and deliver the parameter measurements to sensing circuitry 208. The sensors that generate the in-therapy sensor data may also generate the pre-therapy sensor data.

[0048] In some examples, neuromodulation system 100 includes external sensors 220 communicatively coupled to sensing circuitry 208 and configured to detect parameter measurements. External sensors 220 may measure physiological parameters that may not be measured by neuromodulation catheter 102 (e.g., due to space constraints) and / or that may be measured on parts of a patient that are away from the target tissue site. For example, sensors 220 may be configured to detect parameters at a measurement site away from the target tissue site that may nevertheless be indicative of the effectiveness of the delivery of the neuromodulation therapy. As will be discussed further below, such parameters may be identified through analysis of population parameter data to generate a trained model that may use the parameters as inputs, and that may otherwise appear unrelated to a clinician.

[0049] The pre-therapy sensor data includes one or more parameter measurements that are captured over a period of time prior to the delivery of the neuromodulation therapy. The period of time may include any period of time for which a localized condition that may affect the delivery of neuromodulation therapy may be detected. For example, for parameter measurements detected by neuromodulation catheter 102, the period of time may include less than or equal to a period of time for which electrodes 106 are deployed at a treatment site. On the other hand, for parameter measurements obtained by external sensors 220, such as an additional medical device deployed inside or outside a patient, the period of time may be longer, such as minutes, hours, or days. In some examples, the period of time may be sufficiently long to capture at least one period of a physiologicalcycle or other regular cycles of the patient. For example, the period of time may be sufficiently long to capture a breath and / or heart beat of a patient, such that a non-cyclical deviation may be differentiated from cyclical deviations. In some examples, the period of time is at least five seconds, such as at least ten seconds.

[0050] In some examples, the one or more parameter measurements of the pre-therapy sensor data include impedance measurements. Impedance measurements may be particularly useful for capturing information about localized conditions at the target treatment site, as impedance may be sensitive to changes in tissue or fluid contacting electrodes 106. A waveform of an impedance measurements may have a number of characteristics that can be differentiated including, but not limited to, a frequency of impedance, an amplitude of impedance, a variability of impedance, a phase of impedance, and the like. In some examples, sensing circuitry 208 is configured to process pre-therapy data to extract various characteristics of the pre-therapy data. For example, sensing circuitry may be configured to determine various characteristics of impedance including, but not limited to, amplitude, phase angle, real part (resistance), imaginary part (reactance), frequency dependence, complex impedance, time constant, or impedance spectrum.

[0051] In some examples, the one or more parameter measurements of the pre-therapy sensor data include contact pressure measurements. For example, neuromodulation catheter 102 may include one or more pressure sensors configured to detect a contact pressure between catheter 102 and a wall of the vessel, a contact pressure between one or more electrodes 106 and the wall of the vessel, and / or any other contact pressure that may indicate an amount of contact between electrodes 106 and the vessel prior to delivery of the neuromodulation therapy. Such contact pressure prior to the delivery of the neuromodulation therapy may provide an indication of a quality of contact after delivery of neuromodulation therapy has begun. Characteristics of the contact pressure may include, but are not limited to, a magnitude of pressure, a variability of pressure, a change in pressure, and the like.

[0052] In some examples, the one or more parameter measurements of the pre-therapy sensor data include temperature measurements. For example, temperature measured prior to the delivery of the neuromodulation therapy may provide an indication of a compositionof tissue surrounding the vessel, including an expected change in the patient or device parameters that result from that particular composition.

[0053] Memory 216 may store computer-readable instructions that, when executed by control circuitry 210, cause control circuitry 210 and control device 112 to perform various functions described herein. Memory 216 may be a storage device or other non- transitory medium. Memory 216 may include any volatile, non-volatile, magnetic, optical, or electrical media, such as random-access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), electrically-erasable programmable ROM (EEPROM), flash memory, or any other digital media.

[0054] In the example of FIG. 2A, memory 216 stores one or more trained models 222. As will be described further below, trained model 222 may be used by control circuitry 210 to determine a predicted deviation of the delivery of neuromodulation therapy based on pre-therapy sensor data. Each trained model 222 may be generated from population parameter data for a population of patients that have undergone neuromodulation therapy. The population parameter data includes, for each patient of the population, one or more parameter measurements for a period of time prior to the delivery of the neuromodulation therapy, as well as an indication of whether a deviation of the neuromodulation therapy occurred, such as a TOOR condition or other threshold being exceeded. The population parameter data on which the trained model is trained may also include additional information regarding the patient, such as demographic data, that may affect delivery of neuromodulation therapy. A variety of trained models may be used including, but not limited to, decision trees, such as classification and regression trees; ensemble methods, such as random forest or gradient boosting machines (GBM); linear models, such as linear regression or logistic regression; support vector machines (SVM), such as for classification or regression; neural networks, such as feedforward neural networks (FNN), convolutional neural networks (CNN), recurrent neural networks (RNN), or generative adversarial networks (GANs); Bayesian models; clustering algorithms, such as K-means or hierarchical clustering; or the like.

[0055] In the example of FIG. 2 A, memory 216 stores one or more threshold conditions 224. Threshold conditions 224 may include any conditions limiting deviation in delivery of neuromodulation therapy. Threshold conditions 224 include one or moresafety or effectiveness thresholds that, if exceeded, may cause control circuitry 210 to modify delivery of neuromodulation therapy, such as by terminating neuromodulation therapy to one or all of electrodes 106, changing an amount of neuromodulation therapy to one or all of electrodes 106, or taking some other action. Threshold conditions 224 may include patient threshold conditions that correspond to thresholds for patient parameters and device threshold conditions that correspond to thresholds for device parameters. Patient parameter thresholds may include, but are not limited to, temperature or impedance thresholds of tissues at the target treatment site. Device parameter thresholds may include, but are not limited to, temperature, impedance, or operability thresholds of catheter 102.

[0056] In some examples, a threshold includes a patient parameter or device parameter threshold for one or more patient or device parameter measurements, such as the one or more patient or device parameter measurements of the in-therapy sensor data. In some examples, the patient or device parameter measurements may include at least one of a temperature or an impedance, such that the patient or device threshold includes at least one of a temperature out of range or an impedance out of range. For example, the threshold conditions may include an absolute temperature and / or impedance, a change in temperature and / or impedance, a variability of temperature and / or impedance, or other threshold condition related to temperature and / or impedance that may indicate a deviation from a desired delivery of neuromodulation therapy.

[0057] UI 212 may be configured to output information to and / or receive input from, e.g., a clinician or another user. The clinician may use UI 212 to input information (e.g., threshold condition values) into control device 112. In some examples, the clinician uses UI 212 to instruct control circuitry 210 and / or signal generation circuitry 206 to begin delivery of RF energy to the target tissue site, terminate delivery of RF energy to the target tissue site (e.g., by powering off electrodes 106), and / or adjust the delivered RF energy to the target tissue site (e.g., by switching between monopolar and bipolar energy, by adjust amplitude of the delivered RF energy). The clinician may interact with UI 212 via one or more inputs including tactile, auditory, and / or visual input.

[0058] UI 212 may also output information to the clinician or another user. The outputted information may include, but is not limited, positions of one or more electrodes of electrodes 106 within the patient, a predicted deviation in the delivery of theneuromodulation therapy at one or more electrodes 106, the in-therapy sensor data (e.g., temperature and / or impedance values, such as for each of electrodes 106), or parameters of the RF energy delivered to the patient (e.g., amplitude and / or frequency of the RF energy).

[0059] In some examples, UI 212 notifies the clinician of the predicted deviation of the delivery of neuromodulation therapy. For example, the notification displayed by UI 212 may include a value of the predicted deviation, a classification of the predicted deviation, a classification of a threshold condition 224 that may be exceeded, or any other information related to the predicted deviation. In some examples, UI 212 notifies the clinician of deviation of threshold conditions 224. For example, the notification displayed by UI 212 may include a type of the exceeded threshold condition 224, a threshold value corresponding to threshold condition 224, and / or a determined value (e.g., determined and / or sensed by control circuitry 210 and / or sensing circuitry 208).

[0060] Communications circuitry 214 supports communication between control device 112 and one or more other computing devices, computing systems, and / or cloud computing environments. Control circuitry 210 of control device 112 may retrieve from the one or more other computing devices, computing systems, and / or cloud computing environments values for one or more threshold conditions, additional patient data (e.g., patient parameter data 314 of FIG. 3 A) that may be input into trained model 222, or the like. Control circuitry 210 may transmit, via communications circuitry 214, the pretherapy sensor data, the in-therapy sensor data, and / or the predicted deviation to the one or more other devices, systems, and / or cloud computing environments.

[0061] Control circuitry 210 (alternatively referred to as “processing circuitry 210”) may include any one or more of a microprocessor, a controller, a digital signal processor (DSP), an application-specific integrated circuitry (ASIC), a field-programmable gate array (FPGA), discrete logic circuitry, or any other circuitry configured to provide the functions attributed to control circuitry 210 herein and may be embodied as firmware, hardware, software, or any combination thereof.

[0062] Control circuitry 210 is configured to control signal generation circuitry 206 to deliver neuromodulation to a tissue of a patient based on one or more parameters sensed by sensing circuitry 208 prior to delivering the denervation therapy. FIG. 2B is a flow diagram illustrating an example technique performed by control circuitry 210, whichincludes responding to one or more predicted deviations during delivery of neuromodulation therapy using pre-therapy sensor data, in accordance with some examples of the current disclosure. The method of FIG. 4 will be described with respect to control circuitry of FIG. 2 A.

[0063] Prior to delivering neuromodulation therapy via neuromodulation catheter 102, control circuitry 210 is configured to receive, from one or more sensors, pre-therapy sensor data (230). As described above, the pre-therapy sensor data includes one or more parameter measurements for a period of time prior to the delivery of the neuromodulation therapy. For example, sensors on neuromodulation catheter 102 may detect the parameter measurements, such as impedance measurements, for the period of time prior to delivery of neuromodulation therapy. Additionally, control circuitry 210 may receive other data, such as demographic data, input data, medical history data, or other data that influence a safety or effectiveness of the delivery of the neuromodulation therapy.

[0064] Control circuitry 210 is configured to determine, based on the pre-therapy sensor data, whether the delivery of neuromodulation therapy would cause a deviation (232). The deviation includes one or more patient parameters or device parameters exceeding one or more patient parameter thresholds or device parameter thresholds. The pre-therapy sensor data may be indicative of a potential deviation in delivery of the neuromodulation therapy. A deviation in delivery of neuromodulation therapy may include any variation in the delivery of the neuromodulation therapy that may be associated with or result in an undesirable outcome, such as a patient or device parameter measurement exceeding a threshold. Deviations in the neuromodulation therapy may include, but are not limited to, high impedance, low impedance, high temperature, low temperature, temperature instability, low change in temperature, high change in impedance, impedance instability, high initial impedance, low initial impedance, prolonged treatment time, high initial temperature, low initial temperature, low power, high temperature in intended deenergized electrode. These deviations may indicate undesired local conditions that affect how the neuromodulation therapy is delivered to or received by the tissue of the patient, which can include local conditions associated with catheter 102, such as movement of electrodes 106, poor contact of electrodes 106 with a vessel, poor connection of electrodes 106, or malfunction of catheter 102, or localconditions associated with the patient, such as a composition of adjacent tissues that may receive neuromodulation therapy differently than the target tissues.

[0065] In some examples, control circuitry 210 may use trained model 222 to determine the predicted deviation. When executing trained model 222, control circuitry 210 may be configured to receive as input the pre-therapy sensor data, as well as additional data that may indicate whether a deviation is likely to occur, such as patient parameter data 314 of FIG. 3 A. Trained model 222 may correlate various characteristics of the pre-therapy sensor data with various deviations associated with whether a deviation, such as a TOOR condition, is likely to occur. For example, amplitude, frequency, periodicity, and other characteristics of the pre-therapy sensor data can be differentiated and associated with deviations. In some examples, control circuitry 210 may be configured to identify various characteristics and / or patterns of the pre-therapy sensor data and input those characteristics into trained model 222. As described above, trained model 222 is generated from population parameter data 314 for a population of patient that have undergone neuromodulation therapy. Population parameter data 314 includes, for each patient of the population, one or more parameter measurements for a period of time prior to the delivery of the neuromodulation therapy, as well as an indication of whether a deviation of the neuromodulation therapy occurred.

[0066] Control circuitry 210 is configured to transmit an output based on the determination of whether the delivery of neuromodulation therapy would cause a deviation. Control circuitry 210 may determine that the delivery of neuromodulation therapy would not cause the deviation, in response to the determination that the delivery of neuromodulation therapy would not cause the deviation, control circuitry 210 may transmit an output indicating to proceed with the denervation procedure (238).

[0067] Control circuitry 210 may determine that the delivery of neuromodulation therapy would cause the deviation. In response to the determination that the delivery of neuromodulation therapy would cause a deviation, control circuitry 210 is configured to transmit the output based on the determination (236). In some examples, the output may include a notification to a clinician. The notification may include a classification of the patient or device threshold that would be exceeded. For example, control circuitry 210 may output an indication that a particular TOOR condition, such as related to atemperature out of range or an impedance out of range, would be exceeded if delivery of neuromodulation therapy were to commence. Additionally or alternatively, the notification may include a recommended action. For example, in response to the determination that the delivery of neuromodulation therapy would cause the deviation, control circuitry 210 may transmit an output including an indication to adjust the denervation procedure by at least one of repositioning the neuromodulation catheter, deselecting a therapy delivery element, or modifying the delivery of the neuromodulation therapy. As another example, in response to the determination that the delivery of neuromodulation therapy would cause the deviation, control circuitry may transmit an output indicating to withhold with the denervation procedure.

[0068] In some examples, the output may modify the delivery of the neuromodulation. In response to the determination that the delivery of neuromodulation therapy would cause the deviation, control circuitry 210 may transmit an output that includes a modification to the delivery of the neuromodulation therapy. For example, control circuitry 210 may generate a control signal to signal generation circuitry 206 that modifies delivery of the neuromodulation therapy so that one or more patient or device parameters are less likely to exceed the patient or device threshold during delivery of the neuromodulation therapy. Such modifications may include, but are not limited to, reducing power, increasing power, energizing fewer electrodes, or other actions that may modify the patient or device parameters detected as in-therapy sensor data during delivery of neuromodulation therapy.

[0069] Computing system 114 may be configured to generate trained model 222. FIG. 3 A is a block diagram illustrating an example computing system 114 of FIG. 1.Computing system 114 may include a workstation, a desktop computer, a laptop computer, a smart phone, a tablet, a dedicated computing device, or any other computing device capable of performing the techniques of this disclosure. In the example of FIG. 3 A, computing system 114 includes a memory 302, processing circuitry 304, a display 306, a network interface 308, an input device(s) 310, and an output device(s) 312, each of which may represent any of multiple instances of such a device within the computing system, for ease of description.

[0070] Display 306 may be configured to display information to a user, such as a patient or a clinician. Display 306 may be touch sensitive or voice activated (e.g., via oneor more sensors which may include one or more microphones), enabling display 306 to serve as both an input and output device. Alternatively, a keyboard (not shown), mouse (not shown), or other data input devices (e.g., input device(s) 310) may be employed. Input device(s) 310 may include any device that enables a user to interact with computing system 114, such as, for example, a mouse, keyboard, foot pedal, touch screen, augmented-reality input device receiving inputs such as hand gestures or body movements, or voice interface. Output device(s) 312 may include any connectivity port or bus, such as, for example, parallel ports, serial ports, universal serial busses (USB), or any other similar connectivity port known to those skilled in the art. Network interface 308 may be adapted to connect to a network, such as a local area network (LAN) that includes a wired network or a wireless network, a wide area network (WAN), a wireless mobile network, a Bluetooth network, or the internet. Computing system 114 may receive updates to its software, for example, applications 318, via network interface 308.

[0071] Memory 302 of computing system 114 includes any non-transitory computer- readable storage media for storing data or software that is executable by processing circuitry 304 and that controls the operation of computing system 114. In one or more examples, memory 302 may include one or more solid-state storage devices, such as flash memory chips, or one or more mass storage devices connected to processing circuitry 304 through a mass storage controller (not shown) and a communications bus (not shown). Memory 302 may be configured to store patient parameter data 314, population parameter data 316, and applications 318.

[0072] Patient parameter data 314 may include any data related to a patient, including data indicating a physiological condition of the patient that may, in combination with pretherapy sensor data, indicate whether a deviation in delivery of neuromodulation therapy is likely to occur. As will be described further below regarding model generation, patient parameters may be identified from fixed effects of population parameter data 316, and patient parameter data 314 may include values that correspond to these patient parameters. Patient parameters may include any of patient demographic characteristics, patient imaging characteristics, patient physiological characteristics, patient procedural and medication history, and other measurable or estimable patient characteristics that may influence delivery of neuromodulation therapy. Patient demographic characteristics mayinclude, but are not limited to, age, sex, or race. Patient imaging characteristics may include, but are not limited to, characteristics determined through imaging, characteristics indicated by anatomy, or other characteristics that may be visually indicated using images of an anatomy of a patient. Patient physiological characteristics may include, but are limited to, body mass index (BMI), arterial stiffness, or other measurable physiological characteristics. Patient procedural and medication history may include, but is not limited to, presence of co-morbidities, presence of certain prescribed medication classes, or other conditions in medical history that may influence and / or be correlated with delivery of neuromodulation therapy.

[0073] Patient parameter data 314 may include data from a variety of sources, which may include input data, sensor data, and medical records data. Input data may include any values of patient parameters input by a user. For example, a patient or clinician may enter values of one or more patient parameters directly into input device 310, such as during a patient interview or patient diagnostic testing. Sensor data may include any physiological data of the patient obtained in real-time or near real-time, such as by using an external sensor. Medical records data may include any physiological data or medical history data of the patient obtained previously.

[0074] As mentioned above, control device 112 may be configured to determine a predicted deviation in delivery of neuromodulation therapy based on a trained model that is derived from data for a population of patients that have undergone neuromodulation therapy. Population parameter data 316 may include any data that may indicate a physiological condition of a particular patient that may have influenced delivery of the neuromodulation therapy, such as a presence of one or more TOOR conditions. Population parameter data 316 may include medical records data for each patient in the population, which may include at least one parameter measurement detected over a period of time prior to the delivery of the neuromodulation therapy, and an indication of whether a deviation in the delivery of the neuromodulation therapy occurred. Additionally, population parameter data 316 may include additional parameters of the patient as fixed effects, which may be correlated with whether a deviation in the delivery of the neuromodulation therapy may occur. Fixed effects may include any variable of the patient of the population, including any of the parameters described above with respect to patientparameter data 314, including patient demographic characteristics, patient imaging characteristics, patient physiological characteristics, patient procedural and medication history, and other measurable or estimable patient characteristics that may influence whether a deviation in delivery of neuromodulation therapy occurs.

[0075] Applications 318 may be one or more software programs stored in memory 302 and executed by processing circuitry 304 of computing system 114. Applications 318 of memory 302 may include a model generation module 322, which may be executed by processing circuitry 304. For simplicity, functions performed, or configured to be performed, by applications 318 will be understood to be performed, or configured to be performed, by processing circuitry 304 according to a set of instructions.

[0076] Model generation module 322 may be configured to generate trained model 222 of FIG. 2A based on population parameter data 316. FIG. 3B is a flow diagram illustrating an example technique for generating a trained model for determining a predicted deviation of delivery of neuromodulation therapy based on pre-therapy sensor data. The example technique of FIG. 3B will be described with respect to model generation module 322.

[0077] Model generation module 322 may obtain medical records data for a population of patients that have undergone neuromodulation therapy (330). For example, model generation module 322 may access population parameter data 324 in memory 302 for the medical records data. The medical records data for each patient in the population may include pre-therapy sensor data that includes at least one parameter measurement sensed prior to delivery of the neuromodulation therapy, as well as at least some information related to whether or not a deviation in delivery of the neuromodulation therapy occurred. Population parameter data 316 may also include additional information regarding the patient, such as demographic data, that may affect delivery of neuromodulation therapy.

[0078] For each patient procedure, model generation module 322 may classify the delivery of neuromodulation therapy according to whether or not a deviation occurred (332). For example, model generation module 322 may determine whether a TOOR condition or other threshold condition was exceeded at any point during delivery of the neuromodulation therapy.

[0079] Model generation module 322 may determine a trained model based on the pretherapy sensor data and the classified deviations (334). A variety of trained models may be used including, but not limited to, decision trees, such as classification and regression trees; ensemble methods, such as random forest or gradient boosting machines (GBM); linear models, such as linear regression or logistic regression; support vector machines (SVM), such as for classification or regression; neural networks, such as feedforward neural networks (FNN), convolutional neural networks (CNN), recurrent neural networks (RNN), or generative adversarial networks (GANs); Bayesian models; clustering algorithms, such as K-means or hierarchical clustering; or the like.

[0080] Model generation module 322 may use machine learning to generate trained model 222. Machine learning may generally enable computing system 114 to analyze input data, including population parameter data 316, and identify an action to be performed responsive to the input data. Each trained model 222 may be trained using training data that reflects likely input data. Model generation module 322 may train model 222 to represent a relationship of the pre-therapy sensor data of patients in a population, including parameter measurements, and / or other fixed effects in the medical records data, to whether or not a deviation occurred and / or a magnitude of the deviation. For example, model generation module 322 may train the deep leaning model by adjusting weights of a hidden layer of a neural network model to balance the contribution of each input (e.g., the pre-therapy sensor data included in population parameter data 316).

[0081] Model generation module 322 may output the trained model to control device 112 for use in predicting deviations based on pre-therapy sensor data of a particular patient (336). In some examples, more than one trained model may be available, such that model generation module 322 may output multiple models, or may select a particular model.

[0082] FIG. 3C is a block diagram of a contrastive learning method for generating trained model 222. Contrastive learning is a type of self-supervised learning that trains models by making them distinguish between similar and dissimilar data points. The goal of contrastive learning is to learn useful representations by maximizing agreement between differently augmented views of a same data point while minimizing the agreement between views of different data points. Contrastive learning may be particularly effective in scenarios with limited labeled data, such as patient parameter data316, as contrastive learning frameworks can leverage large amounts of unlabeled data to pre-train the encoder, which can then be fine-tuned on smaller labeled datasets.

[0083] In the example of FIG. 3C, patient parameter data is input (“A”). A set of augmentations 340 are generated from the input by creating multiple transformed versions of the same input data, thereby generating diverse views of the same data point so that model 222 learns to recognize the essential features that remain invariant across these augmentations 340. For example, patient parameter data 316 may have various data that is removed to create the augmentations 340. Augmentations 340 may be input into encoder 342. Encoder 342 is a neural network, such as a transformer, that processes the augmented inputs and extracts feature representations. For example, encoder 342 may map the augmented inputs into a latent feature space, where similar inputs should be closer together. Extracted features 344 output by encoder 342 may represent the input data in a high-dimensional feature space. For example, extracted features 344 may capture relevant information from the input data that the model will use for contrastive learning. Projection head 346 may map extracted features into a space where the contrastive loss is applied.Projection head 346 may improve a quality of learned representations 348 by ensuring that the similarity is measured in an appropriate space. Projection head 348 may be a small neural network (e.g., a few fully connected layers) that further processes extracted features 344. Different augmented versions of the same input are considered “positive pairs,” while views of different inputs are considered “negative pairs.” The contrastive learning framework aims to maximize the agreement (similarity) between positive pairs and minimize the agreement between negative pairs. Parameters of encoder 342 and projection head 346 are optimized to minimize the contrastive loss. After training, encoder 342 learns to produce robust and meaningful representations of the input data, which can then be used for downstream tasks like classification, clustering, or retrieval. Encoder 342 may be output as trained model 222.

[0084] FIG. 4A is a graph of power delivery before, during, and after delivery of neuromodulation therapy. In the example of FIG. 4A, the procedure includes a pretherapy period A prior to delivery of neuromodulation therapy, a therapy period B during delivery of neuromodulation therapy, and a post-therapy period C after delivery of neuromodulation therapy. Therapy period B further includes an initial ramp subperiod Bl,a step up subperiod B2, and a sustained power subperiod B3. During therapy period B, control device 112 may receive in-therapy sensor data, such as temperature and impedance measurements, that indicate how neuromodulation therapy is being received by tissue of the patient. During pre-therapy period A, control device 112 may collect pre-therapy sensor data, such as impedance measurements, that indicate whether a deviation in the delivery of the neuromodulation therapy is likely to occur.

[0085] FIG. 4B is a graph of impedance over time, such as during pre-therapy period A. In the example of FIG. 4B, impedance is measured for a first electrode 400, second electrode 402, third electrode 404, and fourth electrode 406. As seen in FIG. 4B, the impedance signals include signal signatures that may result from various cycles, such as a cardiac cycle or respiratory cycle. The measured impedance signal may be influenced by the conductivity and make-up of the tissue through which a current passes. For example, as an artery contracts and relaxes, the artery conducts electrical impulses that propagate through tissue and the surrounding vasculature differently.

[0086] Experimental Methods

[0087] Different kinds of augmentations multiple combinations of augmentations were tested. The experiments covered using up to four augmented input samples to train a model, as the dataset was limited. The encoder network is set to a ID-ResNet of 12 layers, with a single hidden layer MLP as projection head. The encoder results in 128 features, after which a dropout of 25% occurs before passing through the MLP. For the first round of experimentation, only two views are used when training the model. The average performance of each augmentation was calculated to assess both the most optimal combination and the overall best performing augmentation. The next step in experimentation uses three augmentations, the top three best combinations of augmentations from the first step are used to add a new augmentation to as a third. These were again plotted out to assess the most optimal combination and the most influential augmentation. This same process is repeated for four combinations. For each of these experiments, the model was trained from scratch for each of the 10 different train-test combinations, on every combination the model was trained for 200 epochs with a temperature of 0.07, empirical testing showed this configuration to be enough. The experiments using three augmentations are trained for 1000 epochs with a temperature of0.25. The amount of epochs and temperature were increased as the model needed to assess more augmentations at the same time which is likely to take longer to train. For the experiment using 4 augmentations the amount of epochs is kept to 1000 and the temperature is still to 0.25. The results of these experiments provide insight into which augmentations are most beneficial, in turn providing insight into what can cause the occurrence of OR23.

[0088] FIG. 5 is a scatter plot of predicted deviations by the trained encoder, such as encoder 342 of FIG. 3C. As illustrated in FIG. 5, the encoder predicted a substantial majority of both deviations and no deviations from the test dataset, as evidenced by the clustering and overlap of training and testing datapoints (e.g., blue and green) with no error and of training and testing datapoints (e.g., orange and red) with error.

[0089] Various examples have been described. These and other examples are within the scope of the following claims.

[0090] Example 1. A control device for performing a denervation procedure on a patient, comprising: a memory; and processing circuitry coupled to the memory, wherein the processing circuitry is configured to: receive, from one or more sensors and prior to delivery of neuromodulation therapy via a neuromodulation catheter, pre-therapy sensor data, wherein the pre-therapy sensor data includes one or more parameter measurements for a period of time prior to the delivery of the neuromodulation therapy; determine, based on the pre-therapy sensor data, whether the delivery of neuromodulation therapy would cause a deviation, wherein the deviation includes one or more patient parameters or device parameters exceeding one or more patient parameter thresholds or device parameter thresholds; and transmit an output based on the determination.

[0091] Example 2. The control device of Example 1, wherein the neuromodulation catheter includes the one or more sensors configured to detect the pre- therapy sensor data.

[0092] Example 3. The control device of Example 1 or Example 2, wherein the period of time comprises at least five seconds.

[0093] Example 4. The control device of any of Examples 1 to 3, wherein the memory stores a trained model, and wherein the processing circuitry is configured todetermine, using the trained model, whether the delivery of neuromodulation therapy would cause the deviation.

[0094] Example 5. The control device of Example 4, wherein the trained model is generated from population parameter data for a population of patients that have undergone neuromodulation therapy, and wherein the population parameter data includes, for each patient of the population: one or more parameter measurements for a period of time prior to the delivery of the neuromodulation therapy; and an indication of whether the delivery of neuromodulation therapy caused one or more patient parameters or device parameters to exceed one or more patient parameter thresholds or device parameter thresholds.

[0095] Example 6. The control device of any of Examples 1 to 5, wherein the one or more parameter measurements of the pre-therapy sensor data includes impedance measurements.

[0096] Example 7. The control device of any of Examples 1 to 6, wherein the processing circuitry is configured to receive, from the one or more sensors and during delivery of the neuromodulation therapy, in-therapy sensor data, wherein the in-therapy sensor data includes one or more patient parameter measurements.

[0097] Example 8. The control device of any of Examples 1 to 7, wherein to transmit the output, the processing circuitry is configured to transmit the output based on the determination and a classification of the one or more patient parameters or device parameters.

[0098] Example 9. The control device of Example 7 or 8, wherein the one or more patient parameters include at least one of a temperature or an impedance, and wherein the one or more patient parameter thresholds include at least one of a temperature threshold or an impedance threshold.

[0099] Example 10. The control device of any of Examples 1 to 9, wherein the processing circuitry is configured to determine that the delivery of neuromodulation therapy would not cause the deviation, and wherein to transmit the output, the processing circuitry is configured to, in response to the determination that the delivery of neuromodulation therapy would not cause the deviation, transmit the output indicating to proceed with the denervation procedure.

[0100] Example 11. The control device of any of Examples 1 to 9, wherein the processing circuitry is configured to determine that the delivery of neuromodulation therapy would cause the deviation, and wherein to transmit the output, the processing circuitry is configured to, in response to the determination that the delivery of neuromodulation therapy would cause the deviation, transmit the output indicating to withhold with the denervation procedure.

[0101] Example 12. The control device of any of Examples 1 to 9, wherein the processing circuitry is configured to determine that the delivery of neuromodulation therapy would cause the deviation, and wherein to transmit the output, the processing circuitry is configured to, in response to the determination that the delivery of neuromodulation therapy would cause the deviation, transmit the output including an indication to adjust the denervation procedure by at least one of repositioning the neuromodulation catheter, deselecting a therapy delivery element, or modifying the delivery of the neuromodulation therapy.

[0102] Example 13. The control device of any of Examples 1 to 9, wherein the processing circuitry is configured to determine that the delivery of neuromodulation therapy would cause the deviation, and wherein to transmit the output, the processing circuitry is configured to, in response to the determination that the delivery of neuromodulation therapy would cause the deviation, transmit the output that includes a modification to the delivery of the neuromodulation therapy.

[0103] Example 14. The control device of any of Examples 1 to 13, wherein the neuromodulation catheter is a radiofrequency (RF) catheter.

[0104] Example 15. The control device of Example 14, wherein the RF catheter comprises a plurality of electrodes, and wherein the pre-therapy sensor data comprises impedance measurements for each electrode of the plurality of electrodes.

[0105] Example 16. A neuromodulation system, comprising: a neuromodulation catheter comprises one or more therapy delivery elements; and a control device communicatively coupled to the neuromodulation catheter and configured to: receive, from one or more sensors and prior to delivery of neuromodulation therapy via a neuromodulation catheter, pre-therapy sensor data, wherein the pre-therapy sensor data includes one or more parameter measurements for a period of time prior to the delivery ofthe neuromodulation therapy; determine, based on the pre-therapy sensor data, whether the delivery of neuromodulation therapy would cause a deviation, wherein the deviation includes one or more patient parameters or device parameters exceeding one or more patient parameter thresholds or device parameter thresholds; and transmit an output based on the determination.

[0106] Example 17. The neuromodulation system of Example 16, wherein the neuromodulation catheter includes the one or more sensors configured to capture the pretherapy sensor data.

[0107] Example 18. The neuromodulation system of Example 16 or 17, wherein the period of time comprises at least five seconds.

[0108] Example 19. The neuromodulation system of any of Examples 16 to 18, wherein the memory stores a trained model, and wherein the processing circuitry is configured to determine, using the trained model, whether the delivery of neuromodulation therapy would cause the deviation.

[0109] Example 20. The neuromodulation system of Example 19, wherein the trained model is generated from population parameter data for a population of patients that have undergone neuromodulation therapy, and wherein the population parameter data includes, for each patient of the population: one or more parameter measurements for a period of time prior to the delivery of the neuromodulation therapy; and an indication of whether the delivery of neuromodulation therapy caused one or more patient parameters or device parameters to exceed one or more patient parameter thresholds or device parameter thresholds.

[0110] Example 21. The neuromodulation system of any of Examples 16 to 20, wherein the one or more parameter measurements of the pre-therapy sensor data includes impedance measurements.

[0111] Example 22. The neuromodulation system of any of Examples 16 to 21, wherein the processing circuitry is configured to receive, from the one or more sensors and during delivery of the neuromodulation therapy, in-therapy sensor data, wherein the intherapy sensor data includes one or more patient parameter measurements.

[0112] Example 23. The neuromodulation system of any of Examples 16 to 22, wherein to transmit the output, the processing circuitry is configured to transmit the outputbased on the determination and a classification of the one or more patient parameters or device parameters.

[0113] Example 24. The neuromodulation system of Example 22 or 23, wherein the one or more patient parameters include at least one of a temperature or an impedance, and wherein the one or more patient parameter thresholds include at least one of a temperature threshold or an impedance threshold.

[0114] Example 25. The neuromodulation system of any of Examples 16 to 24, wherein the processing circuitry is configured to determine that the delivery of neuromodulation therapy would not cause the deviation, and wherein to transmit the output, the processing circuitry is configured to, in response to the determination that the delivery of neuromodulation therapy would not cause the deviation, transmit the output indicating to proceed with the denervation procedure.

[0115] Example 26. The neuromodulation system of any of Examples 16 to 24, wherein the processing circuitry is configured to determine that the delivery of neuromodulation therapy would cause the deviation, and wherein to transmit the output, the processing circuitry is configured to, in response to the determination that the delivery of neuromodulation therapy would cause the deviation, transmit the output indicating to withhold with the denervation procedure.

[0116] Example 27. The neuromodulation system of any of Examples 16 to 24, wherein the processing circuitry is configured to determine that the delivery of neuromodulation therapy would cause the deviation, and wherein to transmit the output, the processing circuitry is configured to, in response to the determination that the delivery of neuromodulation therapy would cause the deviation, transmit the output including an indication to adjust the denervation procedure by at least one of repositioning the neuromodulation catheter, deselecting a therapy delivery element, or modifying the delivery of the neuromodulation therapy.

[0117] Example 28. The neuromodulation system of any of Examples 16 to 24, wherein the processing circuitry is configured to determine that the delivery of neuromodulation therapy would cause the deviation, and wherein to transmit the output, the processing circuitry is configured to, in response to the determination that the deliveryof neuromodulation therapy would cause the deviation, transmit the output that includes a modification to the delivery of the neuromodulation therapy.

[0118] Example 29. The neuromodulation system of any of Examples 16 to 28, wherein the neuromodulation catheter is a radiofrequency (RF) catheter.

[0119] Example 30. The neuromodulation system of Example 29, wherein theRF catheter comprises a plurality of electrodes, and wherein the pre-therapy sensor data comprises impedance measurements for each electrode of the plurality of electrodes.

[0120] Example 31. A method for performing a denervation procedure, comprising: receiving, by processing circuitry from one or more sensors and prior to delivery of neuromodulation therapy via a neuromodulation catheter, pre-therapy sensor data, wherein the pre-therapy sensor data includes one or more parameter measurements for a period of time prior to the delivery of the neuromodulation therapy; determining, by the processing circuitry and based on the pre-therapy sensor data, whether the delivery of neuromodulation therapy would cause a deviation, wherein the deviation includes one or more patient parameters or device parameters exceeding one or more patient parameter thresholds or device parameter thresholds; and transmitting, by the processing circuitry, an output based on the determination.

[0121] Example 32. The method of Example 31, wherein the neuromodulation catheter includes the one or more sensors configured to capture the pre-therapy sensor data.

[0122] Example 33. The method of Examples 31 or 32, wherein the period of time comprises at least five seconds.

[0123] Example 34. The method of any of Examples 31 to 33, further comprising determining, using a trained model, whether the delivery of neuromodulation therapy would cause the deviation.

[0124] Example 35. The method of Example 34, wherein the trained model is generated from population parameter data for a population of patients that have undergone neuromodulation therapy, and wherein the population parameter data includes, for each patient of the population: one or more parameter measurements for a period of time prior to the delivery of the neuromodulation therapy; and an indication of whether the deliveryof neuromodulation therapy caused one or more patient parameters or device parameters to exceed one or more patient parameter thresholds or device parameter thresholds.

[0125] Example 36. The method of any of Examples 31 to 35, wherein the one or more parameter measurements of the pre-therapy sensor data includes impedance measurements.

[0126] Example 37. The method of any of Examples 31 to 36, further comprising receiving, by the processing circuitry from the one or more sensors and during delivery of the neuromodulation therapy, in-therapy sensor data, wherein the in-therapy sensor data includes one or more patient parameter measurements.

[0127] Example 38. The method any of Examples 31 to 37, further comprising transmitting, by the processing circuitry, the output based on the determination and a classification of the one or more patient parameters or device parameters.

[0128] Example 39. The method of Example 37 or 38, wherein the one or more patient parameters include at least one of a temperature or an impedance, and wherein the one or more patient parameter thresholds include at least one of a temperature threshold or an impedance threshold.

[0129] Example 40. The method of any of Examples 31 to 39, further comprising: determining, by the processing circuitry, that the delivery of neuromodulation therapy would not cause the deviation, and in response to the determination that the delivery of neuromodulation therapy would not cause the deviation, transmitting, by the processing circuitry, the output indicating to proceed with the denervation procedure.

[0130] Example 41. The method of any of Examples 31 to 39, further comprising: determining, by the processing circuitry, that the delivery of neuromodulation therapy would cause the deviation; and in response to the determination that the delivery of neuromodulation therapy would cause the deviation, transmitting, by the processing circuitry, the output indicating to withhold with the denervation procedure.

[0131] Example 42. The method of any of Examples 31 to 39, further comprising determining, by the processing circuitry that the delivery of neuromodulation therapy would cause the deviation; and in response to the determination that the delivery of neuromodulation therapy would cause the deviation, transmitting, by the processing circuitry, the output including an indication to adjust the denervation procedure by at leastone of repositioning the neuromodulation catheter, deselecting a therapy delivery element, or modifying the delivery of the neuromodulation therapy.

[0132] Example 43. The method of any of Examples 31 to 39, further comprising: determining, by the processing circuitry, that the delivery of neuromodulation therapy would cause the deviation; and in response to the determination that the delivery of neuromodulation therapy would cause the deviation, transmitting, by the processing circuitry, the output that includes a modification to the delivery of the neuromodulation therapy.

[0133] Example 44. The method of any of Examples 31 to 43, wherein the neuromodulation catheter is a radiofrequency (RF) catheter.

[0134] Example 45. The method of Example 44, wherein the RF catheter comprises a plurality of electrodes, and wherein the pre-therapy sensor data comprises impedance measurements for each electrode of the plurality of electrodes.

Claims

CLAIMS1. A control device for performing a denervation procedure on a patient, comprising: a memory; and processing circuitry coupled to the memory, wherein the processing circuitry is configured to: receive, from one or more sensors and prior to delivery of neuromodulation therapy via a neuromodulation catheter, pre-therapy sensor data, wherein the pretherapy sensor data includes one or more parameter measurements for a period of time prior to the delivery of the neuromodulation therapy; determine, based on the pre-therapy sensor data, whether the delivery of neuromodulation therapy would cause a deviation, wherein the deviation includes one or more patient parameters or device parameters exceeding one or more patient parameter thresholds or device parameter thresholds; and transmit an output based on the determination.

2. The control device of claim 1, wherein the neuromodulation catheter includes the one or more sensors configured to detect the pre-therapy sensor data.

3. The control device of claim 1 or 2, wherein the period of time comprises at least five seconds.

4. The control device of any one of claims 1 to 3, wherein the memory stores a trained model, and wherein the processing circuitry is configured to determine, using the trained model, whether the delivery of neuromodulation therapy would cause the deviation.

5. The control device of claim 4, wherein the trained model is generated from population parameter data for a population of patients that have undergone neuromodulation therapy, and wherein the population parameter data includes, for each patient of the population:one or more parameter measurements for a period of time prior to the delivery of the neuromodulation therapy; and an indication of whether the delivery of neuromodulation therapy caused one or more patient parameters or device parameters to exceed one or more patient parameter thresholds or device parameter thresholds.

6. The control device of any one of claims 1 to 5, wherein the processing circuitry is configured to receive, from the one or more sensors and during delivery of the neuromodulation therapy, in-therapy sensor data, wherein the in-therapy sensor data includes one or more patient parameter measurements.

7. The control device of any one of claims 1 to 6, wherein to transmit the output, the processing circuitry is configured to transmit the output based on the determination and a classification of the one or more patient parameters or device parameters.

8. The control device of claim 6 or 7, wherein the one or more patient parameters include at least one of a temperature or an impedance, and wherein the one or more patient parameter thresholds include at least one of a temperature threshold or an impedance threshold.

9. The control device of any one of claims 1 to 8, wherein the processing circuitry is configured to determine that the delivery of neuromodulation therapy would not cause the deviation, and wherein to transmit the output, the processing circuitry is configured to, in response to the determination that the delivery of neuromodulation therapy would not cause the deviation, transmit the output indicating to proceed with the denervation procedure.

10. The control device of any one of claims 1 to 8,wherein the processing circuitry is configured to determine that the delivery of neuromodulation therapy would cause the deviation, and wherein to transmit the output, the processing circuitry is configured to, in response to the determination that the delivery of neuromodulation therapy would cause the deviation, transmit the output indicating to withhold with the denervation procedure.

11. The control device of any one of claims 1 to 8, wherein the processing circuitry is configured to determine that the delivery of neuromodulation therapy would cause the deviation, and wherein to transmit the output, the processing circuitry is configured to, in response to the determination that the delivery of neuromodulation therapy would cause the deviation, transmit the output including an indication to adjust the denervation procedure by at least one of repositioning the neuromodulation catheter, deselecting a therapy delivery element, or modifying the delivery of the neuromodulation therapy.

12. The control device of any one of claim 1 to 8, wherein the processing circuitry is configured to determine that the delivery of neuromodulation therapy would cause the deviation, and wherein to transmit the output, the processing circuitry is configured to, in response to the determination that the delivery of neuromodulation therapy would cause the deviation, transmit the output that includes a modification to the delivery of the neuromodulation therapy.

13. The control device of any one of claims 1 to 12, wherein the neuromodulation catheter is a radiofrequency (RF) catheter comprising a plurality of electrodes, and wherein the pre-therapy sensor data comprises impedance measurements for each electrode of the plurality of electrodes.

14. A neuromodulation system, comprising: a neuromodulation catheter comprises one or more therapy delivery elements; anda control device communicatively coupled to the neuromodulation catheter and configured to: receive, from one or more sensors and prior to delivery of neuromodulation therapy via a neuromodulation catheter, pre-therapy sensor data, wherein the pretherapy sensor data includes one or more parameter measurements for a period of time prior to the delivery of the neuromodulation therapy; determine, based on the pre-therapy sensor data, whether the delivery of neuromodulation therapy would cause a deviation, wherein the deviation includes one or more patient parameters or device parameters exceeding one or more patient parameter thresholds or device parameter thresholds; and transmit an output based on the determination.

15. The neuromodulation system of claim 14, wherein the neuromodulation catheter is a radiofrequency (RF) catheter comprising a plurality of electrodes, and wherein the pre-therapy sensor data comprises impedance measurements for each electrode of the plurality of electrodes.

Images