Use of surrogate model for stenting

Abstract

Example systems and techniques are disclosed that may determine a recommended stent and a recommended stent location. An example system includes one or more memories configured to store surrogate machine learning model, the surrogate machine learning model being trained on imaging data of a plurality of patients and simulation output data based on at least a portion of the imaging data of the plurality of patients. The system includes processing circuitry communicatively coupled to the one or more memories. The processing circuitry is configured to receive imaging data of at least a portion of a vasculature of a patient. The processing circuitry is configured to execute the surrogate machine learning model to determine a recommended stent and a recommended stent location within the vasculature of the patient. The processing circuitry is configured to output an indication of the recommended stent and the recommended stent location for display.

Description

A0013365W001 / 1241-330W001USE OF SURROGATE MODEL FOR STENTING

[0001] This application claims the benefit of U.S. Provisional Patent Application No. 63 / 727,768, filed on December 4, 2024, the content of which is hereby incorporated by reference.TECHNICAL FIELD

[0002] This disclosure relates to the use of images captured during a medical procedure.BACKGROUND

[0003] During a medical procedure, a clinician may use an imaging system to be able to visualize internal anatomy of a patient. Such an imaging system may display anatomy, medical instruments, or the like, and may be used to diagnose a patient condition or assist in guiding a clinician in moving a device, such as a medical instrument to an intended location inside the patient. Imaging systems may use sensors to capture video images which may be displayed during the medical procedure. Imaging systems include angiography systems, N-plane angiography systems, ultrasound imaging systems, computed tomography (CT) scan systems, magnetic resonance imaging (MRI) systems, isocentric C-arm fluoroscopic systems, positron emission tomography (PET) systems, intravascular ultrasound (IVUS), optical coherence tomography (OCT), cardiac computed tomography angiography (CCTA), as well as other imaging systems.

[0004] A medical procedure may include the implantation of a stent in vasculature of a patient. In some cases, an implanted stent may be malapposed, such that at least one stent strut is not in full contact with a vessel luminal wall or where a distance between a strut and a luminal border of the vessel is greater than the thickness of the strut plus polymer. Stent malapposition may increase a risk of restenosis and thrombosis in the patient.SUMMARY

[0005] In general, this disclosure is directed to various techniques and medical systems for using a surrogate machine learning model during a medical procedure to provide recommendation(s) to a clinician. The surrogate machine learning model may be used to reduce the likelihood of stent malapposition, and thereby improve patient outcomes. Such recommendations may include an identification of a stent to use in a patient and / or a location or position in which to implant the stent. In some examples, the recommendations may further include an identification of a balloon, a guidewire, and / or a catheter to use during the procedure.A0013365W001 / 1241-330W001In some examples, the recommendations may include predicted fluid flow characteristics and / or a likelihood of stent malapposition.

[0006] The surrogate machine learning model may be a surrogate for one or more simulations that may make near real-time recommendations impossible or highly unlikely. Some simulations of fluid flow characteristics, malapposition, turbulence, wall shear stress, and / or the like, run on actual current patient imaging data for a plurality of different stents and / or stent locations may take days to run. Because of the time involved in running such simulations, use of the simulations to provide near real-time recommendations during a current medical procedure on a current patient may not be feasible.

[0007] According to the techniques of this disclosure, a surrogate machine learning model may be trained based on imaging data of a plurality of patients and simulation output of simulations run on such imaging data. During a medical procedure, imaging data of a current patient undergoing the medical procedure may be input into the trained machine learning model. The trained machine learning model may determine fluid flow characteristics, malapposition, turbulence, wall shear stress, and / or the like for the current patient for each of a plurality of stents and / or stent locations. Because the surrogate machine learning model is trained based on imaging data and simulation output data, the surrogate machine learning model may determine similar data to that of the simulation(s), but for imaging data for a current patient during a current medical procedure. Thus, the surrogate machine learning model may act as a surrogate for one or more simulation(s), providing near-real time recommendations, which otherwise may not be possible.

[0008] In one example, a medical system includes: one or more memories configured to store a surrogate machine learning model, the surrogate machine learning model being trained on imaging data of a plurality of patients and simulation output data based on at least a portion of the imaging data of the plurality of patients; and processing circuitry communicatively coupled to the one or more memories, the processing circuitry being configured to: receive imaging data of at least a portion of a vasculature of a patient; execute the surrogate machine learning model to determine, based on the imaging data of the at least the portion of the vasculature of the patient, a recommended stent and a recommended stent location within the vasculature of the patient; and output an indication of the recommended stent and the recommended stent location for display.

[0009] In another example, a method includes: receiving, by processing circuitry, imaging data of at least a portion of a vasculature of a patient; executing, by the processing circuitry, a surrogate machine learning model to determine, based on the imaging data of the at least the portion of the vasculature of the patient, a recommended stent and a recommended stent location within the vasculature of the patient, wherein the surrogate machine learning model is trained on imaging data of a plurality of patients and simulation output data based on at least a portion of the imagingA0013365W001 / 1241-330W001 data of the plurality of patients; and outputting, by the processing circuitry, an indication of the recommended stent and the recommended stent location for display.

[0010] In another example, non-transitory computer readable media stores instructions, which, when executed, cause processing circuitry to: receive imaging data of at least a portion of a vasculature of a patient; execute a surrogate machine learning model to determine, based on the imaging data of the at least the portion of the vasculature of the patient, a recommended stent and a recommended stent location within the vasculature of the patient, wherein the surrogate machine learning model is trained on imaging data of a plurality of patients and simulation output data based on at least a portion of the imaging data of the plurality of patients; and output an indication of the recommended stent and the recommended stent location for display.

[0011] These and other aspects of the present disclosure will be apparent from the detailed description below. In no event, however, should the above summaries be construed as limitations on the claimed subject matter, which subject matter is defined solely by the attached claims.

[0012] This summary is intended to provide an overview of the subject matter described in this disclosure. It is not intended to provide an exclusive or exhaustive explanation of the apparatus and methods described in detail within the accompanying drawings and description below. Further details of one or more examples are set forth in the accompanying drawings and the description below.BRIEF DESCRIPTION OF DRAWINGS

[0013] FIG. l is a schematic perspective view of one example of a system for using imaging data according to one or more aspects of this disclosure.

[0014] FIG. 2 is a block diagram of one example of a computing device in accordance with one or more aspects of this disclosure.

[0015] FIG. 3 is a flow diagram illustrating an example technique of using a surrogate model for stenting according to one or more aspect of this disclosure.

[0016] FIG. 4 is a flow diagram illustrating an example frontend of a technique for the use of a surrogate model for stenting according to one or more aspects of this disclosure.

[0017] FIG. 5 is a flow diagram illustrating an example technique for collection and / or generation of training data according to one or more aspects of this disclosure.

[0018] FIG. 6 is a conceptual diagram illustrating an example of training model(s) according to one or more aspects of this disclosure.

[0019] FIG. 7 is a flow diagram illustrating example use of surrogate model for stenting techniques of this disclosure.A0013365W001 / 1241-330W001

[0020] FIG. 8 is a conceptual diagram illustrating an example machine learning model according to one or more aspects of this disclosure.

[0021] FIG. 9 is a conceptual diagram illustrating an example training process for a machine learning model according to one or more aspects of this disclosure.DETAILED DESCRIPTION

[0022] Imaging systems may be used to assist a clinician in a medical procedure, such as a diagnostic medical procedure, a therapeutic medical procedure, such as a percutaneous coronary intervention (PCI) procedure, a stent implantation procedure, or the like, or any combination thereof. For example, imaging systems may be used to determine the presence of lesions within a vasculature of a patient that may be limiting or obstructing blood flow within the vasculature of the patient. Imaging systems may also be used when performing a stent implantation to address the presence of such lesions. While described primarily herein with respect to the vasculature of a patient, imaging systems described herein may be used for other medical purposes and are not limited to cardiovascular purposes. Imaging systems may generate image and / or video data via sensors. This image / video data may be displayed during a medical procedure and / or be recorded for later use. The image / video data may include representations of portions of vasculature or heart of a patient, including one or more lesions which may be restricting blood flow through the portion of the vasculature or the heart of the patient, a geometry and location within a blood vessel or the heart of such lesions, and / or any medical instrument which may be within a field of view of one or more sensors of the imaging system. In some examples, contrasting fluid may be injected into the vasculature of the patient and the imaging data may include fluoroscopy imaging.

[0023] Stent malapposition, where at least one stent strut is not in full contact with a vessel luminal wall or where a distance between a strut and a luminal border of the vessel is greater than the thickness of the strut plus polymer, is a common occurrence in coronary procedures that increases the risk of restenosis and thrombosis. See K. Mahadevan, C. Cosgrove, J. Stange, “Factors Influencing Stent Failure in Chronic Total Occlusion Coronary Intervention,” Interv Cardiol, 2021 Oct. 12. Poor procedural technique, stent diameter, and stent design selection play a role in stent malapposition, particularly when combined with variable arterial composition of patients. Attempts at further characterizing lesions through imaging may result in additional radiation exposure to the patient and additional procedural time, which extends the use of medical resources and personnel. Additionally, characterizing lesion morphology and flow may result in the use of additional medical devices.

[0024] The techniques of this disclosure may reduce radiation exposure to the patient, procedural time, and reduce the need for additional medical devices during a medical procedure.A0013365W001 / 1241-330W001As such, the techniques of this disclosure provide a practical application of technology which may result in improved patient outcomes, as well as preserve medical resources.

[0025] As discussed above, clinician stent selection, positioning, and deployment contributes to high stent malapposition rates (62% for drug-eluting stents, see., e.g., Zandvoort et. al). Stent malapposition may result in adverse procedural outcomes, including stent restenosis. Computational fluid dynamics (CFD), finite element analysis (FEA), fluid-structure interaction (FSI) simulations may be used to model the deployment of stents to select optimal treatment lesions (where multiple parts of the coronary tree are affected), deployment sites, and devices (e.g., stents). Unfortunately, such simulations are time-intensive incurring great cost and may take several days to mesh, run, and analyze, even for a skilled analyst. However, using surrogate modelling, a representative anatomical dataset may be simulated ahead of time (e.g., prior to a medical procedure) using a mix or combination of real and synthetic patient data. Such a dataset may be used to train a surrogate model using one of many methodologies, such as a deep neural network. The surrogate model may be configured to provide a near instantaneous analysis of the coronary anatomy and predict how multiple devices (e.g., stents) may perform. The analysis of processing circuitry executing the surrogate model may be provided to a clinician for interpretation and / or may select for, or propose, to the clinician, the predicted best-performing device(s) for use during the medical procedure, which may lessen clinical burden on the clinician.

[0026] FIG. l is a schematic perspective view of one example of a system for using imaging data according to one or more aspects of this disclosure. System 100 includes a display device 110, a table 120, an imager 140, and a computing device 150. System 100 may be an example of a system for use in an emergency room or a Cath lab. In some examples, system 100 may include other devices, not shown for simplicity purposes. In some examples, system 100 may also include server 160, which may be co-located with the other devices of system 100 or may be located elsewhere. In some examples, system 100 may include additional imager(s) 170. In the examples where additional imager(s) 170 are included in system 100, additional imager(s) 170 may include one or more imagers different than imager 140. System 100 may be used during a medical procedure, such as a medical procedure to implant a stent into a vessel of a patient.

[0027] Computing device 150 may include, for example, an off-the-shelf device such as a laptop computer, desktop computer, tablet computer, smart phone, or other similar device or may include a specific purpose device. Computing device 150 may perform various control functions with respect to imager 140. In some examples, computing device 150 may include a guidance workstation. Computing device 150 may control the operation of imager 140 and receive the output of imager 140 and may receive imaging data from imager 140. Computing device 150 may executeA0013365W001 / 1241-330W001 one or more machine learning models to determine procedural recommendations to provide to a clinician.

[0028] Display device 110 may be configured to output instructions, images, and messages relating to the medical procedure(s), such as any procedural recommendations determined by computing device 150. For example, display device 110 may display imaging data obtained through imager 140, additional imager(s) 170, and / or one or more procedural recommendations. Table 120 may be, for example, an operating table or other table suitable for use during a medical procedure.

[0029] In the example of FIG. 1, imager 140, such as an angiography imager, fluoroscopy imager, a CT imager, or other imaging device, may be used to image relevant portions of the patient’s anatomy during a medical procedure to visualize the anatomy, a medical instrument, and / or a device to be implanted, such as a stent, inside the patient’s body through the generation of imaging data. While primarily described herein as an angiography imager, imager 140 may be any type of imaging device, such as an angiography device, an N-plane angiography device, a fluoroscopy device, an isocentric C-arm fluoroscopic device, a CT device, a CCTA device, an IVUS device, an OCT device, an MRI device, a PET device, an ultrasound device, or the like.

[0030] Imager 140 may image a region of interest in the patient’s body. The particular region of interest may be dependent on anatomy, the medical procedure, patient symptoms, and / or the like. For example, when performing a stent implant procedure, the region of interest may include the vessel of the patient into which the stent is being implanted, or a portion thereof.

[0031] Additional imager(s) 170, may also be used to image relevant portions of the patient’s anatomy during a medical procedure to visualize the anatomy, a medical instrument, and / or a device to be implanted, such as a stent, inside the patient’s body through the generation of imaging data. Additional imager(s) 170 may include one or more of an angiography device, an N-plane angiography device, a fluoroscopy device, an isocentric C-arm fluoroscopic device, a CT device, a CCTA device, an IVUS device, an OCT device, an MRI device, a PET device, an ultrasound device, or the like.

[0032] Computing device 150 may be communicatively coupled to imager 140, display device 110, additional imager(s) 170, and / or server 160, for example, by wired, optical, or wireless communications. Server 160 may be a hospital server which may or may not be located in an emergency room or Cath lab of a hospital, a cloud-based server, or the like. Server 160 may be configured to store patient imaging data (such as angiography data), electronic healthcare or medical records, or the like. In some examples, server 160 may be configured to execute the machine learning model(s) and / or perform one or more of, or a portion of one or more of, the determinations associated therewith.A0013365W001 / 1241-330W001

[0033] Any of, or any combination of, computing device 150, imager 140, additional imager(s) 170, and / or server 160 may include one or more machine learning model(s). For example, computing device 150, and / or server 160 may receive imaging data of at least a portion of a vasculature of a patient, e.g., via imager 140 and / or additional imager(s) 170. Computing device 150 and / or server 160 may execute a surrogate machine learning model to determine, based on the imaging data of the at least the portion of the vasculature of the patient, a recommended stent and a recommended stent location with the vasculature of the patient. Computing device 150 and / or server 160 may output an indication of the recommended stent and the recommended stent location for display.

[0034] By outputting an indication of the recommended stent and the recommended stent location, system 100 may assist clinicians in more effectively and efficiently performing a stent implantation procedure while reducing or minimizing a likelihood of stent malapposition. As such, the techniques of this disclosure may improve patient outcomes and / or improve medical facility efficiency.

[0035] FIG. 2 is a block diagram of one example of a computing device in accordance with one or more aspects of this disclosure. Computing device 200 may be an example of computing device 150, imager 140, additional imager(s) 170, and / or server 160 of FIG. 1 and may include a workstation, a desktop computer, a laptop computer, a server, a smart phone, a tablet, a dedicated computing device, or any other computing device capable of performing the techniques of this disclosure.

[0036] In some examples, computing device 200 may be configured to perform processing, control and other functions associated with computing device 150, imager 140, additional imager(s) 170, and / or server 160. In some examples, computing device 200 represents multiple instances of computing devices, each of which may be associated with one or more of computing device 150, imager 140, additional imager(s) 170, and / or server 160. Computing device 200 may include, for example, a memory 202, processing circuitry 204, a display 206, a network interface 208, an input device(s) 210, or an output device(s) 212, each of which may represent any of multiple instances of such a device within the computing system, for ease of description.

[0037] While processing circuitry 204 appears in computing device 200 in FIG. 2, in some examples, features attributed to processing circuitry 204 may be performed by processing circuitry of any of computing device 150, imager 140, additional imager(s) 170, or server 160, or combinations thereof. In some examples, one or more processors associated with processing circuitry 204 in computing device 200 may be distributed and shared across any combination of computing device 150, imager 140, additional imager(s) 170, and / or server 1060. Additionally, in some examples, processing operations or other operations performed by processing circuitry 204A0013365W001 / 1241-330W001 may be performed by one or more processors residing remotely, such as one or more cloud servers or processors, each of which may be considered a part of computing device 200. Computing device 200 may be used to perform any of the techniques described in this disclosure, and may form all or part of devices or systems configured to perform such techniques, alone or in conjunction with other components, such as components of computing device 150, imager 140, additional imager(s) 170, server 160, or a system including any or all of such devices.

[0038] Memory 202 of computing device 200 includes any non-transitory computer-readable storage media for storing data or software that is executable by processing circuitry 204 and that controls the operation of computing device 150, imager 1040, additional imager(s) 170, or server 1060, as applicable. In one or more examples, memory 202 may include one or more solid-state storage devices such as flash memory chips. In one or more examples, memory 202 may include one or more mass storage devices connected to the processing circuitry 204 through a mass storage controller (not shown) and a communications bus (not shown).

[0039] Although the description of computer-readable media herein refers to a solid-state storage, it should be appreciated by those skilled in the art that computer-readable storage media may be any available media that may be accessed by the processing circuitry 204. That is, computer readable storage media includes non-transitory, volatile and non-volatile, removable and non-removable media implemented in any method or technology for storage of information such as computer-readable instructions, data structures, program modules, or other data. For example, computer-readable storage media includes RAM, ROM, EPROM, EEPROM, flash memory or other solid state memory technology, CD-ROM, DVD, Blu-Ray or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that may be used to store the desired information and that may be accessed by computing device 200. In one or more examples, computer-readable storage media may be stored in the cloud or remote storage and accessed using any suitable technique or techniques through at least one of a wired or wireless connection.

[0040] Memory 202 may store machine learning model(s) 222 and simulation(s) 224. Machine learning model(s) 222 may include a surrogate machine learning model. The surrogate machine learning model may be trained based on imaging data 214, synthetic training data, and / or output of simulation(s) 224 run on imaging data 214. Simulation(s) 224 may include, for example, a CFD simulation application, an FEA simulation application, and / or an FSI simulation application. Simulation(s) 224 may be configured to generate pre-procedural (e.g., prior to stent implantation) simulated data based on imaging data 214. The pre-procedural simulated data may include fractional flow reserve (FFR) data and / or fluid distribution data (e.g., fluid flow map data) which may represent predictions or estimates of FFR and / or fluid flow associated with imaging data 214.A0013365W001 / 1241-330W001Simulation(s) 224 may also be configured to generate post-procedural (e.g., after stent implantation) simulated data based on imaging data 214, for each of a plurality of potential stents and a plurality of potential stent implantation locations. The post-procedural simulated data may include FFR data, fluid distribution data (e.g., fluid flow map data), stent malapposition data, turbulence (e.g., thrombogenic) data, and / or wall shear stress data, each of which may represent a prediction or estimate of such data if a respective stent is implanted at a respective location.

[0041] Memory 202 may store imaging data 214, patient imaging data 216, procedural recommendation(s) 226, and three-dimensional (3D) geometry 228 of patient anatomy. Imaging data 214 may be captured by imager 140 (FIG. 1) and / or additional imager(s) 170 during medical procedures of a plurality of patients. In some examples, imaging data 214 may therefore include a relatively large volume of imaging data which may be used as input to simulation(s) 224 and to train machine learning model(s) 222. Patient imaging data 216 may be captured by imager 140 (FIG. 1) and / or additional imager(s) 170 during a current medical procedure of a current patient.

[0042] Processing circuitry 204 may receive imaging data 214 and / or imaging data 216 from imager 140 and / or additional imager(s) 170, and store imaging data 214 and / or patient imaging data 216 in memory 202. During training of the surrogate machine learning model, processing circuitry 204 may convert imaging data 214 into 3D geometry 228. For example, processing circuitry 204 may build CAD geometry from imaging data 214. In some examples, processing circuitry 204 may execute edge detection, mesh, boundary and / or other techniques to convert imaging data 214 into 3D geometry 228.

[0043] Processing circuitry 204 may execute simulation(s) 224 on 3D geometry 228 to generate pre-procedural simulated data and / or post-procedural simulated data discussed above. Processing circuitry 204 may train the surrogate machine learning model using the pre-procedural simulated data, the post-procedural simulated data and imaging data 214.

[0044] During a current medical procedure, processing circuitry 204 may receive patient imaging data 216, which may include imaging data of a portion of a vasculature of a current patient. Processing circuitry 204 may execute the surrogate machine learning model of machine learning model(s) 222 on patient imaging data 216 to generate procedural recommendation(s) 226. Procedural recommendation(s) 226 may include a recommended stent to be used for implantation and a recommended implantation location for the recommended stent. In some examples, procedural recommendation(s) 226 may also include one or more of a recommended balloon to be used, a recommended guidewire to be used, a recommended catheter not be used, predicted fluid flow characteristics (e.g., fluid flow map data, FFR data, and / or the like), or a likelihood of stent malapposition using the recommended stent. The surrogate machine learning model may be configured to rank, order, and / or screen any recommendations based on one or more of theA0013365W001 / 1241-330W001 likelihood of stent malapposition, predicted post-procedure fluid flow characteristics, predicted post-procedure turbulence, predicted post-procedure wall-shear stress, and / or the like. For example, processing circuitry 204 may present an ordered list of one or more recommended stents and associated stent locations (which may be shown overlaid on imaging data 216, in some examples).

[0045] Processing circuitry 204 may output for display, e.g., to display 206 and / or display device 110, patient imaging data 216 during a current medical procedure of a current patient to assist the clinician during the current medical procedure. Processing circuitry 204 may also output for display to display 206 and / or display device 110, procedural recommendation(s) 226.

[0046] For example, machine learning model(s) 222 may be trained using data collected from past medical procedures, such as imaging data 214. Thus, in some examples, machine learning model(s) 222 may be trained on actual treatments and actual outcomes from past medical procedures.

[0047] For example, a k-means clustering model may be used having a plurality of clusters: one for each particular stent in a particular location. Each identified lesion may associated with a vector that includes variables for, e.g., type of coronary issue, severity of the coronary issue, complexity of the coronary issue, location of the coronary issue, classification of a lesion, anatomy in the area of the coronary issue, other anatomy, comorbidities of the patient, cholesterol level, blood pressure, blood oxygenation, age, physical exercise level, and / or the like. The location of the vector in a given one of the clusters may be indicative of a particular stent being implanted in a particular location.

[0048] Other potential machine learning or artificial intelligence techniques that may be used include Naive Bayes, k-nearest neighbors, random forest, support vector machines, neural networks, linear regression, logistic regression, etc.

[0049] Processing circuitry 204 may be implemented by one or more processors, which may include any number of fixed-function circuits, programmable circuits, or a combination thereof. In various examples, control of any function by processing circuitry 204 may be implemented directly or in conjunction with any suitable electronic circuitry appropriate for the specified function. Fixed-function circuits refer to circuits that provide particular functionality and are preset on the operations that may be performed. Programmable circuits refer to circuits that may programmed to perform various tasks and provide flexible functionality in the operations that may be performed. For instance, programmable circuits may execute software or firmware that cause the programmable circuits to operate in the manner defined by instructions of the software or firmware. Fixed-function circuits may execute software instructions (e.g., to receive parameters or output parameters), but the types of operations that the fixed-function circuits perform areA0013365W001 / 1241-330W001 generally immutable. In some examples, the one or more of the units may be distinct circuit blocks (fixed-function or programmable), and in some examples, the one or more units may be integrated circuits.

[0050] Instructions may be executed by one or more processors, such as one or more digital signal processors (DSPs), general purpose microprocessors, application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), graphics processing units (GPUs) or other equivalent integrated or discrete logic circuitry. Accordingly, the term processing circuitry 204 as used herein may refer to one or more processors having any of the foregoing processor or processing structure or any other structure suitable for implementation of the techniques described herein. In addition, in some aspects, the functionality described herein may be provided within dedicated hardware or software modules configured for encoding and decoding, or incorporated in a combined codec. Also, the techniques could be fully implemented in one or more circuits or logic elements.

[0051] Display 206 may be touch sensitive or voice activated, enabling display 206 to serve as both an input and output device. Alternatively, a keyboard (not shown), mouse (not shown), or other data input device(s)s (e.g., input device(s) 210) may be employed.

[0052] Network interface 208 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. In some examples, network interface 208 may include one or more application programming interfaces (APIs) for facilitating communication with other devices. For example, computing device 200 may receive imaging data 214 and / or patient imaging data 216 from imager 140 and / or additional imager(s) 170 during medical procedures via network interface 208.

[0053] Input device(s) 210 may be any device that enables a user to interact with computing device 200, such as, for example, a mouse, keyboard, foot pedal, touch screen, augmented-reality input device(s) receiving inputs such as hand gestures or body movements, or voice interface.

[0054] Output device(s) 212 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.

[0055] Applications 217 may be one or more software programs stored in memory 202 and executed by processing circuitry 204 of computing device 200.

[0056] In some examples, processing circuitry 204 may provide clinical guidance (e.g., one or more procedural recommendation(s) 226) to a clinician. For example, processing circuitry 204 may execute a surrogate machine learning model of machine learning model(s) 222 to provide a clinician with procedural recommendation(s) 226. Procedural recommendation(s) 226 mayA0013365W001 / 1241-330W001 include a recommended stent to be used for implantation and a recommended implantation location for the recommended stent. In some examples, procedural recommendation(s) 226 may also include one or more of a recommended balloon to be used, a recommended guidewire to be used, a recommended catheter to be used, predicted fluid flow characteristics (e.g., fluid flow map data, FFR data, and / or the like), or a likelihood of stent malapposition using the recommended stent.

[0057] FIG. 3 is a flow diagram illustrating an example technique of using a surrogate model for stenting according to one or more aspect of this disclosure. Processing circuitry 204 may build a database of image data (300). For example, processing circuitry 204 may control imager 140 and / or additional imager(s) 170 to capture image data of patients and store the image data in imaging data 214 of memory 202 and / or of server 160.

[0058] Processing circuitry 204 may build CAD geometry based on imaging data 214 and, in some examples, solve fluid flow, for example, using CFD (302). For example, processing circuitry 204 may convert geometry within image data to CAD geometry. Processing circuitry may use the CAD geometry as input to a simulation of simulation(s) 224 to determine fluid flow within vessels depicted in the image data.

[0059] Processing circuitry 204 may solve for flow characteristic(s) after simulated stenting for various stents and / or stent positions (304). For example, processing circuitry 204 may simulate different stents and / or different stent positions being deployed within the vessels represented by the CAD geometry and generate simulation data including such flow characteristic(s).

[0060] Processing circuitry 204 may train a surrogate model using the simulation data (306). For example, processing circuitry 204 may train one or more machine learning model(s) 222 which may include a surrogate model using the simulation data of step 304 and / or step 306.

[0061] During a medical procedure, processing circuitry 204 may control imager 140 and / or additional imagers 170 to capture image data, such as CINE data from a current patient of the medical procedure. Processing circuitry 204 may attempt to match the image data with previously captured image data in the database (308). For example, processing circuitry 204 may select one or more closest matches from imaging data 214 to the patient imaging data 216 captured of the current patient.

[0062] Processing circuitry 204 may recommend an optimal (e.g., suggested) stent and rank a similarity between the imaging data 214 and the patient imaging data 216 of the current patient (310). For example, processing circuitry 204 may select one or more recommended stents and rank a similarly of the imaging data 214 and patient imaging data 216 captured of the current patient. Processing circuitry 204 may control display 206, display device 110, and / or output device(s) 212 to output, for use by the clinician, the recommendation(s) and / or rankings.A0013365W001 / 1241-330W001

[0063] In some examples, processing circuitry 204 may compare the similarity ranking to a predetermined threshold ranking. If the similarity ranking satisfies the predetermined threshold (e.g., the difference in similarity ranking is greater than, or greater than or equal to a difference threshold, or the similarity is less than, or less than or equal to a similarity threshold), processing circuity 204 may collect more image data and / or add the image data of the current patient (e.g., imaging data 216) to the database (e.g., imaging data 214).

[0064] FIG. 4 is a flow diagram illustrating an example frontend of a technique for the use of a surrogate model for stenting according to one or more aspects of this disclosure. A clinician may capture one or more images of a lesion (400). For example, a clinician may use computing device 200 to control imager 140 and / or additional imager(s) 170, or may use imager 140 and / or additional imager(s) 170 to capture imaging data 214. This imaging data may include, for example, any of, or any combination of, angiography, N-plane angiography, CCTA, MRI, OCT, and / or IVUS image data.

[0065] Processing circuitry 204 may convert the image data to 3D geometry (402). For example, processing circuitry 204 may build 3D geometry 228 based on the image data, such as building a 3D CAD model.

[0066] Processing circuitry 204 or the clinician may sweep the lesion and / or surrounding anatomy using imaging sensor(s) of imager 140 and / or additional imager(s) 170 and processing circuitry 204 may control display 206 and / or display device 110 to visually present the imaging data being collected to the clinician such that the clinician can confirm validity of the 3D geometry (404). For example, the clinician may use input device(s) 210 to enter an indication that the 3D geometry appears valid. In some examples, processing circuitry 204 may determine that the 3D geometry is valid if the clinician does not enter an indication that the 3D geometry appears invalid, for example, within a period of time.

[0067] Processing circuitry 204, executing one or more machine learning model(s) 222 may 406 interpret the 3D geometry (406). Because the surrogate machine learning model may be trained on imaging data 214 and output of simulation(s) 224, the surrogate machine learning model may be configured to interpret the 3D geometry. For example, processing circuitry 204, executing one or more machine learning model(s) 222 may determine one or more of 1) a flow map and / or FFR data of the patient pre-procedure, 2) a post-procedure flow map and / or FFR data, 3) a malapposition count and identification of malapposition region(s), 4) an identification of turbulent (e.g., thrombogenic) regions, and / or 5) wall shear stress (408). In some examples, processing circuitry 204 may identify any malapposition regions and / or turbulent regions on the 3D geometry or on displayed image data by highlighting or otherwise marking the malapposition regions and / orA0013365W001 / 1241-330W001 turbulent regions. In some examples, the identifications of any malapposition regions and turbulent regions may be different, so as to enable a clinician to differentiate therebetween.

[0068] FIG. 5 is a flow diagram illustrating an example technique for collection and / or generation of training data according to one or more aspects of this disclosure. Processing circuitry 204 may control imager 140 and / or additional imager(s) 170 to capture image data 214 of real coronary anatomy of a plurality of patients (500). This imaging data may include, for example, any of, or any combination of, angiography, N-plane angiography, CCTA, MRI, OCT, and / or IVUS image data.

[0069] Processing circuitry 204 or a clinician, may control imager 140 and / or additional imager(s) 170 to perform a 3D sweep of the anatomy (502). Processing circuitry 204 may use the captured image data, including the image data captured during the 3D sweep, to determine real geometry and lesion composition data (504). For example, processing circuitry 204 may use any applicable techniques to determine lesion composition based on the collected image data.

[0070] Processing circuitry 204 may perform a large deformation diffeomorphic matrix mapping on the determined real geometry (506) to generate synthetic training data. Processing circuitry 204 may automatically assign mesh and boundary conditions for the real and synthetic training data 508 using, for example, batch processing (510).

[0071] Processing circuitry 204 may thereby generate a relatively large initialized coronary geometry and lesion model library (512). For example, because the training data includes both real and synthesized training data, the initialized coronary geometry and lesion model library may be greater in size than if just real data were used.

[0072] Processing circuitry 204 may execute a simulation of simulation(s) 224, such as a CFD simulation, on the model library (514) to generate pre-procedural FFR data and / or pre-procedural fluid distribution data 516.

[0073] Processing circuitry 204 may also execute one or more simulations of simulation(s) 224, such as a CFD simulation, an FEA simulation, and / or an FSI simulation for a plurality of possible stents and / or stent positions (518) to generate post-procedural predictions of one or more of FFR data, fluid distribution data, stent malapposition data (e.g., number and / or locations(s)), turbulence data, and / or wall shear stress data 520.

[0074] In some examples, processing circuitry 204 may use a device finite element analysis model and material library of stents, balloons, catheters, and / or guidewires 522, to down select the stent pool 524 for which processing circuity 204 executes the one or more simulations 518. In this manner, the number of simulations may be limited by the inventory of medical devices available, thereby avoiding additional simulations which may not be feasible to implement or replicate in aA0013365W001 / 1241-330W001 patient. In some examples, processing circuitry 204 may down-select the stent pool 524 further based on lesion length and / or arterial diameter.

[0075] FIG. 6 is a conceptual diagram illustrating an example of training model(s) according to one or more aspects of this disclosure. Processing circuitry 204 may train the machine learning model (e.g., of machine learning model(s) 222) based on pre-procedure geometry data 600. The pre-procedure geometry data 600 may include radial position (r), angular position (0), and angular position (z). By training the machine learning model on pre-procedure geometry data 600, processing circuitry 204 may generate a trained model that may be configured to predict preprocedural fluid flow field data and / or pre-procedural FFR data

[0076] Processing circuitry 204 may also train the machine learning model based on pre- procedural data 604. Pre-procedural data 604 may include radial position (r), angular position (0), and angular position (z), stent position, balloon selected, and / or stent selected. By training the machine learning model on pre-procedure geometry data 604, processing circuitry 204 may generate a trained model that may be configured to predict prost-procedural fluid flow field data and / or pre-procedural FFR data, updated anatomy geometry, malapposition, turbulence, and / or wall shear stress 606.

[0077] FIG. 7 is a flow diagram illustrating example surrogate model stenting techniques according to one or more aspects of this disclosure. Processing circuitry 204 may receive imaging data of at least a portion of a vasculature of a patient (700). For example, processing circuitry 204 may obtain patient imaging data 216 from imager 140 and / or additional imager(s) 170 during a current medical procedure of the patient.

[0078] Processing circuitry 204 may execute the surrogate machine learning model to determine, based on the imaging data of the at least the portion of the vasculature of the patient, a recommended stent and a recommended stent location within the vasculature of the patient (702). For example, processing circuitry 204 may execute the surrogate machine learning model on patient imaging data 216 to generate procedural recommendation(s) 226. The surrogate machine learning model may be trained on imaging data of a plurality of patients and simulation output data based on at least a portion of the imaging data of the plurality of patients.

[0079] Processing circuitry 204 may output an indication of the recommended stent and the recommended stent location for display. For example, processing circuitry 204 may output procedural recommendations to display 206 and / or display device 110 during the stent implantation procedure in near real-time.

[0080] In some examples, processing circuitry 204 may execute the surrogate machine learning model to determine at least one of a recommended balloon, a recommended guidewire, a recommended catheter, predicted fluid flow characteristics, or a likelihood of stent malapposition,A0013365W001 / 1241-330W001 and wherein the output further comprises at least one of an indication of a recommended balloon, an indication of the recommended guidewire, an indication of the recommended catheter, the predicted fluid flow characteristics, or the likelihood of stent malapposition. In some examples, the simulation output data includes output data of at least one simulation, the at least one simulation comprising a computational fluid dynamics simulation, a finite element analysis simulation, or a fluid- structure interaction simulation.

[0081] In some examples, the simulation output data includes simulated pre-procedure data, the simulated pre-procedure data at least one of pre-procedural fluid flow field data or preprocedural fractional flow reserve data. In some examples, the simulation output data includes simulated post-procedural data, the simulated post-procedural data including at least one of postprocedural fluid flow field data, post-procedural fractional flow reserve data, post-procedural anatomy geometry, an indication of post-procedural malapposition, post-procedural turbulence data, or post-procedural vessel wall shear stress data. In some examples, the simulated postprocedural data includes respective simulated post-procedural for each of a plurality of potential stents.

[0082] In some examples, processing circuitry 204 may be configured to train the surrogate machine learning model. In some examples, to train the surrogate machine learning model, processing circuitry 204 may be configured to convert the imaging data of the plurality of patients to geometry data, and to execute the one or more simulations on the geometry data to generate the simulation output data. In some examples, to train the surrogate machine learning model, processing circuitry 204 may be configured to perform a large deformation diffeomorphic matrix mapping on the geometry data to generate synthetic training data, and train the surrogate machine learning model based on geometry data and the synthetic training data.

[0083] In some examples, processing circuitry 204 may refrain from executing a simulation on the imaging data of at least the portion of the vasculature of the patient. For example, processing circuitry 204 may not execute any of simulation(s) 224 on patient imaging data 216 and instead execute the surrogate machine learning model on patient imaging data 216.

[0084] FIG. 8 is a conceptual diagram illustrating an example machine learning model according to one or more aspects of this disclosure. Machine learning model 800 may be an example of machine learning model(s) 222. Machine learning model 800 may be an example of a deep learning model, or deep learning algorithm, trained to determine procedural recommendation(s) 226. Procedural recommendation(s) 226 may include a recommended stent and a recommended stent implantation location. In some examples, procedural recommendation(s) 226 may further include one or more of a recommended balloon, a recommended guidewire, a recommended catheter, predicted fluid flow characteristics, or a likelihood of stent malapposition.A0013365W001 / 1241-330W001Computing device 200 may train, store, and / or utilize machine learning model 800, but other devices of system 100 may apply inputs to machine learning model 800 in some examples. In some examples, various types of machine learning and deep learning models or algorithms may be utilized. For examples, a convolutional neural network (CNN), e.g., ResNet-18, may be used. Some non-limiting examples of machine learning techniques include Support Vector Machines, K- Nearest Neighbor algorithm, and Multi-layer Perceptron.

[0085] As shown in the example of FIG. 8, machine learning model 800 may include three types of layers. These three types of layers include input layer 802, hidden layers 804, and output layer 806. Output layer 806 comprises the output from the transfer function 805 of output layer 806. Input layer 802 represents each of the input values XI through X4 provided to machine learning model 800. In some examples, the input values may include any of the values input into the machine learning model, as described above. For example, the input values may include patient imaging data 216, as described above. In addition, in some examples input values of machine learning model 800 may include additional data, such as other data that may be collected by or stored in system 100.

[0086] Each of the input values for each node in the input layer 802 is provided to each node of a first layer of hidden layers 804. In the example of FIG. 8, hidden layers 804 include two layers, one layer having four nodes and the other layer having three nodes, but fewer or greater number of nodes may be used in other examples. Each input from input layer 802 is multiplied by a weight and then summed at each node of hidden layers 804. During training of machine learning model 800, the weights for each input are adjusted to establish a relationship between imaging data 214, output of simulation(s) 224, various stents, various stent locations, and / or a likelihood of stent malapposition. In some examples, one hidden layer may be incorporated into machine learning model 800, or three or more hidden layers may be incorporated into machine learning model 800, where each layer includes the same or different number of nodes.

[0087] The result of each node within hidden layers 804 is applied to the transfer function of output layer 806. The transfer function may be linear or non-linear, depending on the number of layers within machine learning model 800. Example non-linear transfer functions may be a sigmoid function or a rectifier function. The output 807 of the transfer function may be a classification of a model of stent, a stent implantation location, and a likelihood of stent malapposition.

[0088] As shown in the example above, by applying machine learning model 800 to input data such as patient imaging data 216, processing circuitry 204 is able to determine or estimate a likelihood of stent malappostion for a plurality of different stents at different locations within a particular patient. This may also improve the ability of a clinician to successfully treat a medical condition, thereby improving patient outcomes.A0013365W001 / 1241-330W001

[0089] FIG. 9 is a conceptual diagram illustrating an example training process for a machine learning model according to one or more aspects of this disclosure. Process 900 may be used to train machine learning model(s) 222 or machine learning model 800. A machine learning model 974 (which may be an example of machine learning model 800 and / or machine learning model(s) 222) may be implemented using any number of models for semi-supervised, supervised, and / or reinforcement learning, such as but not limited to, an artificial neural network, a decision tree, naive Bayes network, support vector machine, or k-nearest neighbor model, CNN, recursive neural network (RNN), long short-term memory (LSTM), ensemble network, to name only a few examples.

[0090] In some examples, one or more of computing device 150 and / or server 160 initially trains machine learning model 974 based on a corpus of training data 972. Training data 972 may include, for example, imaging data 214, 3D geometry 228, and / or output of simulation(s) 224. For example, training data 972 may include imaging data 214, output of simulation(s) 224, simulated training data, and / or the like.

[0091] In some examples, training data 972 may include images of anatomy of the current patient, images of anatomy of other patients, and / or the like. For example, training data 972 may include data from past medical procedures performed on a plurality of patients having different anatomy, different prior medical procedures, annotations or tags, other training data mentioned herein, and / or the like. In some examples, with the subsequent medical follow up of patients, treatment and outcome data may be used to train machine learning model (s) 222 or machine learning model 974 to recommend a stent to be implanted in a current patient and a stent location for the recommended stent to be implanted.

[0092] While training machine learning model 974, processing circuitry 204 may compare 976 a prediction or classification with a target output 978. Processing circuitry 204 may utilize an error signal from the comparison to train (leaming / training 980) machine learning model 974. Processing circuitry 204 may generate machine learning model weights or other modifications which processing circuitry 204 may use to modify machine learning model 974. For examples, processing circuitry 204 may modify the weights of machine learning model 974 based on the leaming / training 980. For example, computing device 200 may, for each training instance in training data 972, modify, based on training data 972, the manner in which a recommended stent and / or recommended stent location is determined.

[0093] The techniques described in this disclosure may be implemented, at least in part, in hardware, software, firmware or any combination thereof. For example, various aspects of the described techniques may be implemented within one or more processors or processing circuitry, including one or more microprocessors, digital signal processors (DSPs), application specificA0013365W001 / 1241-330W001 integrated circuits (ASICs), field programmable gate arrays (FPGAs), or any other equivalent integrated or discrete logic circuitry, as well as any combinations of such components. The terms “controller”, “processor”, or “processing circuitry” may generally refer to any of the foregoing logic circuitry, alone or in combination with other logic circuitry, or any other equivalent circuitry. A control unit comprising hardware may also perform one or more of the techniques of this disclosure. Such hardware, software, and firmware may be implemented within the same device or within separate devices to support the various operations and functions described in this disclosure. In addition, any of the described units, circuits or components may be implemented together or separately as discrete but interoperable logic devices. Depiction of different features as circuits or units is intended to highlight different functional aspects and does not necessarily imply that such circuits or units must be realized by separate hardware or software components. Rather, functionality associated with one or more circuits or units may be performed by separate hardware or software components or integrated within common or separate hardware or software components.

[0094] The techniques described in this disclosure may also be embodied or encoded in a computer-readable medium, such as a computer-readable storage medium, containing instructions. Instructions embedded or encoded in a computer-readable storage medium may cause a programmable processor, or other processor, to perform the method, e.g., when the instructions are executed. Computer readable storage media may include random access memory (RAM), read only memory (ROM), programmable read only memory (PROM), erasable programmable read only memory (EPROM), or electronically erasable programmable read only memory (EEPROM), or other computer readable media.

[0095] The techniques of this disclosure include the following non-limiting examples.

[0096] Example 1. A medical system comprising: one or more memories configured to store a surrogate machine learning model, the surrogate machine learning model being trained on imaging data of a plurality of patients and simulation output data based on at least a portion of the imaging data of the plurality of patients; and processing circuitry communicatively coupled to the one or more memories, the processing circuitry being configured to: receive imaging data of at least a portion of a vasculature of a patient; execute the surrogate machine learning model to determine, based on the imaging data of the at least the portion of the vasculature of the patient, a recommended stent and a recommended stent location within the vasculature of the patient; and output an indication of the recommended stent and the recommended stent location for display.

[0097] Example 2. The medical system of example 1, wherein the processing circuitry is further configured to execute to the surrogate machine learning model to determine at least one of a recommended balloon, a recommended guidewire, a recommended catheter, predicted fluid flowA0013365W001 / 1241-330W001 characteristics, or a likelihood of stent malapposition, and wherein the output further comprises at least one of an indication of the recommended balloon, an indication of the recommended guidewire, an indication of the recommended catheter, the predicted fluid flow characteristics, or the likelihood of stent malapposition.

[0098] Example 3. The medical system of example 1 or example 2, wherein the simulation output data comprises output data of at least one simulation, the at least one simulation comprising a computational fluid dynamics simulation, a finite element analysis simulation, or a fluid- structure interaction simulation.

[0099] Example 4. The medical system of any of examples 1-3, wherein the simulation output data comprises simulated pre-procedure data, the simulated pre-procedure data at least one of preprocedural fluid flow field data or pre-procedural fractional flow reserve data.

[0100] Example 5. The medical system of any of examples 1-4, wherein the simulation output data comprises simulated post-procedural data, the simulated post-procedural data including at least one of post-procedural fluid flow field data, post-procedural fractional flow reserve data, postprocedural anatomy geometry, an indication of post-procedural malapposition, post-procedural turbulence data, or post-procedural vessel wall shear stress data.

[0101] Example 6. The medical system of example 5, wherein the simulated post-procedural data comprises respective simulated post-procedural for each of a plurality of potential stents.

[0102] Example 7. The medical system of any of examples 1-6, wherein the processing circuitry is further configured to train the surrogate machine learning model.

[0103] Example 8. The medical system of example 7, wherein to train the surrogate machine learning model, the processing circuitry is further configured to: convert the imaging data of the plurality of patients to geometry data; and execute one or more simulations on the geometry data to generate the simulation output data.

[0104] Example 9. The medical system of example 8, where to train the surrogate machine learning model, the processing circuitry is configured to: perform a large deformation diffeomorphic matrix mapping on the geometry data to generate synthetic training data; and train the surrogate machine learning model based on geometry data and the synthetic training data.

[0105] Example 10. The medical system of any of examples 1-9, wherein the processing circuitry is further configured to refrain from executing a simulation on the imaging data of at least the portion of the vasculature of the patient.

[0106] Example 11. A method comprising: receiving, by processing circuitry, imaging data of at least a portion of a vasculature of a patient; executing, by the processing circuitry, a surrogate machine learning model to determine, based on the imaging data of the at least the portion of the vasculature of the patient, a recommended stent and a recommended stent location within theA0013365W001 / 1241-330W001 vasculature of the patient, wherein the surrogate machine learning model is trained on imaging data of a plurality of patients and simulation output data based on at least a portion of the imaging data of the plurality of patients; and outputting, by the processing circuitry, an indication of the recommended stent and the recommended stent location for display.

[0107] Example 12. The method of example 11, further comprising: executing, by the processing circuitry, the surrogate machine learning model to determine at least one of a recommended balloon, a recommended guidewire, a recommended catheter, predicted fluid flow characteristics, or a likelihood of stent malapposition; and outputting, at least one of an indication of a recommended balloon, an indication of the recommended guidewire, an indication of the recommended catheter, the predicted fluid flow characteristics, or the likelihood of stent malapposition.

[0108] Example 13. The method of example 11 or example 12, wherein the simulation output data comprises output data of at least one simulation, the at least one simulation comprising a computational fluid dynamics simulation, a finite element analysis simulation, or a fluid-structure interaction simulation.

[0109] Example 14. The method of any of examples 11-13, wherein the simulation output data comprises simulated pre-procedure data, the simulated pre-procedure data at least one of preprocedural fluid flow field data or pre-procedural fractional flow reserve data.

[0110] Example 15. The method of any of examples 11-14, wherein the simulation output data comprises simulated post-procedural data, the simulated post-procedural data including at least one of post-procedural fluid flow field data, post-procedural fractional flow reserve data, postprocedural anatomy geometry, an indication of post-procedural malapposition, post-procedural turbulence data, or post-procedural vessel wall shear stress data.

[0111] Example 16. The method of example 15, wherein the simulated post-procedural data comprises respective simulated post-procedural for each of a plurality of potential stents.

[0112] Example 17. The method of any of examples 11-16, further comprising training, by the processing circuitry, the surrogate machine learning model.

[0113] Example 18. The method of example 17, where training the surrogate machine learning model comprises: converting the imaging data of the plurality of patients to geometry data; and executing one or more simulations on the geometry data to generate the simulation output data.

[0114] Example 19. The method of example 18, wherein training the surrogate machine learning model comprises: performing a large deformation diffeomorphic matrix mapping on the geometry data to generate synthetic training data; and training the surrogate machine learning model based on geometry data and the synthetic training data.A0013365W001 / 1241-330W001

[0115] Example 20. A non-transitory computer-readable storage medium storing instructions, which, when executed, cause processing circuitry to: receive imaging data of at least a portion of a vasculature of a patient; execute a surrogate machine learning model to determine, based on the imaging data of the at least the portion of the vasculature of the patient, a recommended stent and a recommended stent location within the vasculature of the patient, wherein the surrogate machine learning model is trained on imaging data of a plurality of patients and simulation output data based on at least a portion of the imaging data of the plurality of patients; and output an indication of the recommended stent and the recommended stent location for display.

[0116] Various examples have been described. These and other examples are within the scope of the following claims.

Claims

A0013365W001 / 1241-330W001What is claimed is:

1. A medical system comprising: one or more memories configured to store a surrogate machine learning model, the surrogate machine learning model being trained on imaging data of a plurality of patients and simulation output data based on at least a portion of the imaging data of the plurality of patients; and processing circuitry communicatively coupled to the one or more memories, the processing circuitry being configured to: receive imaging data of at least a portion of a vasculature of a patient; execute the surrogate machine learning model to determine, based on the imaging data of the at least the portion of the vasculature of the patient, a recommended stent and a recommended stent location within the vasculature of the patient; and output an indication of the recommended stent and the recommended stent location for display.

2. The medical system of claim 1, wherein the processing circuitry is further configured to execute to the surrogate machine learning model to determine at least one of a recommended balloon, a recommended guidewire, a recommended catheter, predicted fluid flow characteristics, or a likelihood of stent malapposition, and wherein the output further comprises at least one of an indication of the recommended balloon, an indication of the recommended guidewire, an indication of the recommended catheter, the predicted fluid flow characteristics, or the likelihood of stent malapposition.

3. The medical system of claim 1 or claim 2, wherein the simulation output data comprises output data of at least one simulation, the at least one simulation comprising a computational fluid dynamics simulation, a finite element analysis simulation, or a fluid- structure interaction simulation.

4. The medical system of any of claims 1-3, wherein the simulation output data comprises simulated pre-procedure data, the simulated pre-procedure data at least one of pre-procedural fluid flow field data or pre-procedural fractional flow reserve data.

5. The medical system of any of claims 1-4, wherein the simulation output data comprises simulated post-procedural data, the simulated post-procedural data including at least one of post-A0013365W001 / 1241-330W001 procedural fluid flow field data, post-procedural fractional flow reserve data, post-procedural anatomy geometry, an indication of post-procedural malapposition, post-procedural turbulence data, or post-procedural vessel wall shear stress data.

6. The medical system of claim 5, wherein the simulated post-procedural data comprises respective simulated post-procedural for each of a plurality of potential stents.

7. The medical system of any of claims 1-6, wherein the processing circuitry is further configured to train the surrogate machine learning model.

8. The medical system of claim 7, wherein to train the surrogate machine learning model, the processing circuitry is further configured to: convert the imaging data of the plurality of patients to geometry data; and execute one or more simulations on the geometry data to generate the simulation output data.

9. The medical system of claim 8, where to train the surrogate machine learning model, the processing circuitry is configured to: perform a large deformation diffeomorphic matrix mapping on the geometry data to generate synthetic training data; and train the surrogate machine learning model based on geometry data and the synthetic training data.

10. The medical system of any of claims 1-9, wherein the processing circuitry is further configured to refrain from executing a simulation on the imaging data of at least the portion of the vasculature of the patient.

11. A method comprising: receiving, by processing circuitry, imaging data of at least a portion of a vasculature of a patient; executing, by the processing circuitry, a surrogate machine learning model to determine, based on the imaging data of the at least the portion of the vasculature of the patient, a recommended stent and a recommended stent location within the vasculature of the patient, wherein the surrogate machine learning model is trained on imaging data of a plurality of patientsA0013365W001 / 1241-330W001 and simulation output data based on at least a portion of the imaging data of the plurality of patients; and outputting, by the processing circuitry, an indication of the recommended stent and the recommended stent location for display.

12. The method of claim 11, further comprising: executing, by the processing circuitry, the surrogate machine learning model to determine at least one of a recommended balloon, a recommended guidewire, a recommended catheter, predicted fluid flow characteristics, or a likelihood of stent malapposition; and outputting, at least one of an indication of a recommended balloon, an indication of the recommended guidewire, an indication of the recommended catheter, the predicted fluid flow characteristics, or the likelihood of stent malapposition.

13. The method of claim 11 or claim 12, wherein the simulation output data comprises output data of at least one simulation, the at least one simulation comprising a computational fluid dynamics simulation, a finite element analysis simulation, or a fluid- structure interaction simulation.

14. The method of any of claims 11-13, wherein the simulation output data comprises simulated pre-procedure data, the simulated pre-procedure data at least one of pre-procedural fluid flow field data or pre-procedural fractional flow reserve data.

15. A non-transitory computer-readable storage medium storing instructions, which, when executed, cause processing circuitry to: receive imaging data of at least a portion of a vasculature of a patient; execute a surrogate machine learning model to determine, based on the imaging data of the at least the portion of the vasculature of the patient, a recommended stent and a recommended stent location within the vasculature of the patient, wherein the surrogate machine learning model is trained on imaging data of a plurality of patients and simulation output data based on at least a portion of the imaging data of the plurality of patients; and output an indication of the recommended stent and the recommended stent location for display.

Images