Methods of simulating the deployment of implantable medical devices

Abstract

A computer implemented method of simulating the deployment of an implantable medical device at a target deployment position within a patient's vascular structure, the method comprising: receiving a three-dimensional model of the patient's vascular structure; representing the implantable medical device as a centreline of the implantable medical device, the centreline comprising a plurality of nodes, where each node is associated with a two-dimensional shape representing the cross-section of the device at the position of the node; positioning the centreline at the target deployment position within the patient's vascular structure; iteratively (i) calculating a force on each node if the corresponding two-dimensional shape contacts or intersects a wall of the vascular structure within the three-dimensional model, (ii) determining a movement of the node based on the calculated force and (iii) updating the position of the node based on the determined movement; determining the deployed position of the implantable medical device based on the position of the nodes of the centreline when a convergence condition is met.

Description

[0001] METHODS OF SIMULATING THE DEPLOYMENT OF IMPLANTABLE MEDICAL DEVICES

[0002] FIELD OF INVENTION

[0003] This invention relates generally to methods and systems for simulating the deployment of implantable medical devices, such as flow-diverting devices, intrasaccular devices and stents, inside a patient’s anatomy.

[0004] BACKGROUND

[0005] Clinicians use implantable medical devices, such as flow diverters, stents, and intrasaccular devices, to treat a variety of different medical conditions. Such implantable medical devices (IM Ds) are used to treat blood vessels affected by an aneurysm, particularly to avoid the further expansion or rupture of the aneurysm and to avoid blood clots travelling from the aneurysm to block other blood vessels. Clinicians will generally deploy such devices at a target location by inserting them into the patient’s vascular system in a radially compressed state within a catheter, before releasing the IMD such that it expands into a final configuration determined by parameters of the device and the morphology of the deployment location within a patient’s vascular structure.

[0006] It Is important, for effective functioning of a device, that specific final deployed dimensions and positioning of the implantable medical device are achieved. The specific parameters of the final deployed position can be difficult for a clinician to predict, as the dynamics of the deployment of a device through a vascular system are sensitive to the specific details of the patient’s anatomy, parameters of the device, and the deployment technique. In practice, clinicians generally use their experience to select a particular model and size of device that they hope will provide the best results.

[0007] One important category of implantable medical device is intrasaccular devices, which are inserted and deployed inside aneurysms towards the treatment of these diseased vascular structures. These devices are essentially shaped “plugs” composed of braided metal wires which allow the device to collapse when loaded into a catheter prior to deployment and then expand to fill the space of the aneurysm dome. The presence of these devices increases the resistance to incoming blood flow inside the aneurysm dome which prevents the growth and rupture of the aneurysm and allows the tissue to begin to heal. In ideal conditions, a new endothelial layer will be generated along the surface of the device, effectively sealing off the aneurysm and improving patient outcomes.

[0008] Intrasaccular devices come in a variety of sizes from the manufacturer, requiring the surgeon to select the correct device size based on the specific geometry of the aneurysm for a given patient. Correct sizing of the intrasaccular device is of crucial importance to the successful treatment of the aneurysm and improved patient outcome. Devices that are too small will not remain fixed within the dome and will not reduce blood flow as effectively as a well-sized device. Devices that are too large may exert excessive forces against the fragile diseased tissue of the dome wall, risking rupture, or may protrude out of the dome into the parent vasculature, increasing thromboembolic risk. Similar issues arise with other types of devices, notably flow-diverters and stents.

[0009] Not only is device sizing critically important for the treatment of the patient, but it can also be difficult to correctly predict. Each patient’s vasculature is unique, as is the location, shape, size, and orientation, of the aneurysm dome. Surgeons can perform measurements on images, produced by medical scans, to try to characterise the geometry of the dome, however, the common morphological irregularity of these domes often makes it difficult to determine which measurements along the dome are important for sizing as the devices can reposition and reorient during deployment due to mechanical interactions with the dome wall. Additionally, the braided wire composition of these devices causes them to deploy nonlinearly meaning predicting the exact shape and dimensions the device will deploy to can be difficult.

[0010] There is hence a need for improved solutions for assisting the clinician in predicting the deployment of IM Ds within realistic representations of the patient’s specific vasculature before commencing a surgical procedure to deploy the device within the patient. Currently, solutions for simulating implantable medical devices are often prohibitively slow for a surgeon to test a sufficient number of devices, deployment configurations, and techniques. It can take hours, or even days, for state of the art methods, such as finite element analysis, to run one simulation, with a surgeon often having to leave a simulation to run overnight. Therefore, there is an additional need for these improved methods to be computationally efficient, to reduce the time it takes for one simulation to run and allow clinicians to quickly iterate through different device sizes and deployment configurations, to truly determine the best possible plan for treating the patient.

[0011] SUMMARY OF INVENTION

[0012] According to a first aspect of the invention, there is provided a computer implemented method of simulating the deployment of an implantable medical device at a target deployment position within a patient’s vascular structure, the method comprising: receiving a three-dimensional model of the patient’s vascular structure; representing the implantable medical device as a centreline of the implantable medical device, the centreline comprising a plurality of nodes, where each node is associated with a two-dimensional shape representing the crosssection of the device at the position of the node; positioning the centreline at the target deployment position within the patient’s vascular structure; iteratively (i) calculating a force on each node if the corresponding two-dimensional shape contacts or intersects a wall of the vascular structure within the three-dimensional model, (ii) determining a movement of the node based on the calculated force and (iii) updating the position of the node based on the determined movement; determining the final deployed position of the implantable medical device based on the position of the nodes when a convergence condition is met.

[0013] The present method provides improvements for simulating the deployment of an implantable medical device which provides high accuracy in the predicted position, orientation, and length of the deployed device, while also being significantly more computationally efficient than prior art methods. Representing the implantable medical device as nodes along a centreline, with two-dimensional shapes representing the cross-section of the device at the position of the nodes, captures the essential information needed to simulate deployment, but without simulating the full dynamics of each individual wire that forms the outer surface of the device. This is beneficial for a surgeon using this method to find a suitable implantable medical device for a particular patient, as the surgeon can test many different sizes and shapes of devices quickly, reducing the overall time required to plan a procedure. The method of using nodes along a centreline allows for the simulation of the full dynamics of the deployment of the implantable medical device, by iteratively updating the position of each node based on the expanding cross-sections from being deployed from a catheter to the final deployed position. In particular, the constraints may be varied between nodes to simulate the deployment from a catheter. The method is, therefore, usable to determine whether the positioning and orientation of the implantable medical device in its final position are suitable to treat a particular condition in the patient’s vascular structure.

[0014] Preferably the method comprises, at each iteration, determining at each node whether the corresponding two-dimensional shape contacts or intersects a wall of the vascular structure within the three-dimensional model. In particular, the method comprises checking whether there is contact or intersection of the two- dimensional shape and the wall of the vascular structure within the three- dimensional model and calculating a force on each node where contact of intersection is determined. Alternatively stated, the method may comprise: iteratively (i) determining at each node whether the corresponding two- dimensional shape contacts or intersects a wall of the vascular structure within the three-dimensional model, calculating a force on each node if the corresponding two-dimensional shape contacts or intersects a wall of the vascular structure within the three-dimensional model, (ii) determining a movement of the node based on the calculated force and (iii) updating the position of the node based on the determined movement.

[0015] Preferably the convergence condition is met by one or more of: (1) a predetermined number of iterations have occurred, (2) the net force on one or more nodes is reduced below a threshold; (3) a kinetic variable calculated for one or more nodes is below a threshold; (4) the movement of the nodes is below a threshold.

[0016] In preferable examples, the deployed position of the implantable medical device is based on the position of the nodes of the centreline when the movement of the nodes is below a threshold. The movement of the nodes may be determined based on a calculated displacement or velocity of one or more nodes.

[0017] Preferably, the method comprises initiating the simulation with the size of each two-dimensional shape set to a first value representing the initial cross-section of the device at the start of the simulation. The size of the two-dimensional shape may be defined by a function, for example a radius function, dependent on an angle around the centreline, to provide an arbitrary cross-sectional shape. In one example, the simulation is initiated with each two-dimensional shape at its maximal expansion (which may be its total maximal expansion or maximal expansion for a particular configuration, such as a compacted state). A force is then calculated on each shape which contacts or intersects a vessel wall within the three-dimensional model. In another example, the simulation is initiated with each two-dimensional shape at an initial size, with the size of the two-dimensional shape increased through the iterations of the simulation to simulate the expansion of the implantable medical device within the patient’s vascular structure. Having the simulation initiated with each two-dimensional shape at an initial size with the size of the two-dimensional shape increased through the iterations of the simulation may also be beneficial in simulating deployment of the implantable medical device from the catheter.

[0018] In particular, the method may comprise simulating the full deployment dynamics of the implantable medical device by iteratively expanding the two-dimensional shapes from an initial compressed orientation, for example representing containment in a catheter, to a maximally expanded configuration(which may be its total maximal expansion or maximal expansion for a particular configuration, such as a compacted state). Alternatively the method may comprise determining the final deployed position based on the maximally expanded configuration, without considering the deployment dynamics. In particular the method may comprise positioning the centreline at the target deployment position within the patient’s vascular structure, with the size of each two-dimensional shape set to a size corresponding to a maximal expansion or unconstrained configuration of the implantable medical device. The method may comprise initialising the simulation with the size of each two-dimensional shape set according to a maximal expansion of the implantable medical device. The interaction with the device and the vessel walls occurs predominantly at its maximal expansion and so a good approximation of the deployed orientation can be achieved by simulating the forces based only on an expanded configuration. This can provide enhanced computational efficiency since the full deployment need not be simulated.

[0019] In these examples, the iterations of the method comprise, iteratively: (i) calculating a force on each node if the corresponding two-dimensional shape intersects a wall of the vascular structure within the three-dimensional model, (ii) determining a movement of the node based on the calculated force and (iii) updating the position of the node based on the determined movement.

[0020] Preferably, the computer implemented method comprises, at each iteration, calculating a kinematic cost function representing the motion of the nodes and determining that the convergence condition is met when the calculated kinematic cost function reaches a convergence threshold. In this way it is determined when the substantive motion of the device has ended and the device is in its final deployed state. This is beneficial as it is important to know when the implantable medical device has stopped moving, in order to determine when the simulation has ended. The final deployed state of the implantable medical device can be determined from the orientation of the device at the end of the simulation. The kinematic cost function represents a measure of the overall motion of the nodes (and therefore the device) and hence provides a convergence condition for ending the simulation and determining the final position. The kinematic cost function may depend on one or more kinematic variables. The kinematic cost function may be based on the displacement of the nodes (e.g. the displacement between subsequent iterations of the simulation) and / or the velocity of the nodes. The kinematic cost function may be a sum or average of the one or more kinematic variables across the nodes. The convergence threshold can be increased or decreased to change how sensitive the simulation is to the kinematics of the nodes.

[0021] Preferably, the kinematic cost function is dependent on one or both of displacement and velocity of one or more nodes of the centreline. By making the cost function dependent on one or more nodes of the centreline, the final position of the device can depend on one particular section of the implantable medical device, or the entire device.

[0022] Preferably, the kinematic cost function comprises a sum or average of the magnitude of the displacement or velocity calculated across a plurality of nodes. By averaging across a plurality of nodes, the method may find a final deployment position that is more suitable for multiple parts of the device, and not optimised for only one part. In particular, a kinematic cost function defined in this way provides a measure of the overall motion of the device.

[0023] Preferably the method comprises iteratively: expanding each two-dimensional shape from an initial configuration towards a maximal radial expansion of the device at the position of the node; calculating a force when a two-dimensional shape contacts or intersects a wall of the vascular structure; determining a movement of the corresponding node based on the calculated force; wherein the deployed position is determined when the movement of the nodes, based on the calculated forces, is below a threshold or the device has reached its maximal expansion.

[0024] Alternatively phrased, the method may comprise iteratively expanding each two- dimensional shape from an initial configuration towards a maximal radial expansion of the device at the position of the node, wherein each iteration comprises (i) calculating a force on each node if the corresponding two- dimensional shape contacts or intersects a wall of the vascular structure within the three-dimensional model, (ii) determining a movement of the node based on the calculated force and (Hi) updating the position of the node based on the determined movement.

[0025] In this way it is possible to model the expansion of the device and the corresponding movement of the device during the expansion due to the interaction with the surrounding vasculature. In some implementations it is important to model the full expansion of the device to determine the final deployed orientation, including whether or not the implantable medical device is able to fully expand in its final deployment position. In some embodiments, the phrase “when a two- dimensional shape contacts or intersects a wall of the vascular structure” may mean that a distance between a point on the perimeter of the two-dimensional shape and the wall of the patient’s vascular structure is zero, or less than some threshold value. Similarly, it may mean that the distance between a point on the perimeter of the two-dimensional shape and the wall of the patient’s vascular structure is less than zero, i.e. the two-dimensional shape is overlapping with the vessel wall within the model. Preferably at each iteration of the simulation, the cross-sectional size of one or more of the two-dimensional shapes is expanded and it is determined if the cross-section of the one or more two-dimensional shapes contacts or intersects the vessel wall with the model and generates force accordingly.

[0026] Preferably, the method of the second aspect comprises, at each iteration, expanding the two-dimensional shapes and calculating a kinematic cost function representing the motion of the nodes; and determining the deployed position when the calculated kinematic cost function reaches a convergence threshold.

[0027] Preferably expanding a two-dimensional shape comprises expanding the two- dimensional shape perpendicular to the centreline of the implantable medical device at the position of the corresponding node. This is beneficial as expanding the two-dimensional shapes perpendicular to the centreline best represents the expansion of real implantable medical devices.

[0028] Preferably, determining a movement of a node comprises: determining a contact force at each position on each two-dimensional shape that contacts or intersects a wall of the vascular structure; determining a net force on each node by summing all contact forces applied to the corresponding two-dimensional shape; and determining a movement of each node based on the net force. In this way, the movement of the IM D is modelled by determining movement of the nodes of the centreline, providing a particularly computationally efficient method to model the overall motion of the device. By using the net force applied to a node, the method can readily incorporate other types of forces acting on the device, represented by the two-dimensional shapes. By summing all contact forces applied to the corresponding two-dimensional shape, the influence of all of the contact points on each two-dimensional shape is considered, which produces a more accurate simulation.

[0029] Preferably, in addition to determining a contact force, the method comprises determining one or more of the following additional forces: an elastic force that resists a change in length of the centreline; a bending force that limits the curvature of the centreline; frictional forces following contact between the two- dimensional shape and a wall of the vascular structure; where the method comprises including the above one or more additional forces when determining the net force. In this sense, the simulation better represents deployment of a real implantable medical device and, therefore, is more accurate. In some embodiments, forces other than: elastic forces, bending forces, and frictional forces, may be determined and included in determining the net force.

[0030] The method of simulating the deployment of an implantable medical device may further comprise: casting rays radially from the centreline of the implantable medical device to measure the distance from the centreline of the implantable medical device to a wall of the vascular structure. Determining the relevant dimensions of the model of the vascular structure can be challenging and computationally intensive. By using ray casting from the centreline, it is possible to measure the relevant dimensions from the device’s perspective as it deploys. Preferably the rays are cast from the nodes of the centreline in a plurality of directions at regular angular spacing around the centreline. Preferably the rays are cast repeatedly to repeatedly measure the distance of the centreline to the wall of the vascular structure as the nodes move during the simulation. Preferably the rays are cast at a plurality of iterations, preferably at each iteration or at periodic iterations. In this way the relevant dimensions of the vascular structure are continuously measured during simulation of the deployment. A significant challenge in estimating the deployed position of an implantable medical device is determining the important relevant dimensions of the vascular geometry given the device will move during deployment. The current method allows the relevant dimensions to be measured accurately using ray casting from the nodes.

[0031] Preferably the calculated ray distance is used to determine when one of the two- dimensional shapes contacts or intersects a wall of the vascular structure. By determining contact or intersection points using ray casting, the method can apply the contact force at these points, to simulate the interaction between the implantable medical device, and the walls of the patient’s vascular structure. The method may comprise determining a difference between the distance from the centreline to the vessel wall measured by ray-casting and the radial size of the two-dimensional shape (for each of a plurality of measurement angles around the centreline). In some examples the determined contact force may be proportional to the difference between the measurements, when the two-dimensional shape intersects the vessel wall.

[0032] In some examples the two-dimensional shapes comprise a circle. This may be used to represent a substantially round or cylindrical implantable medical device.

[0033] In some examples the method of simulating the deployment of an implantable medical device further comprises varying a maximum radial dimension of the two- dimensional shapes with time and / or position along the centreline, thereby simulating the unloading of the implantable medical device from a catheter. In this way the dynamics of the deployment of the device from a catheter may be modelled. In particular a first portion of the implantable medical device may have a maximum radial dimension that is greater than a second portion of the implantable medical device, such that the second portion appears to be confined in the catheter, and the first portion is free to expand. In this sense, the device appears to bloom out of the catheter, as it does when physically deployed in a patient’s vascular structure. This is sometimes referred to as “flowering” out of the catheter. State of the art methods for simulating deployment from a catheter are prohibitively computationally expensive. In this sense, the time it takes to run a state of the art simulation of deployment from a catheter is too long for a surgeon to practically perform this step for many devices. The method presented here for simulating deployment from a catheter is computationally efficient, meaning that a surgeon is able to practically perform this step when simulating many implantable medical devices in quick succession according to the method of the present disclosure.

[0034] Preferably the maximum radial dimension of the device is increased in a stepwise manner over the length of the implantable medical device to simulate the implantable medical device emerging from the catheter. In this way, the method more accurately simulates the progressive release (or ‘emerging’) of the implantable medical device from the catheter.

[0035] Preferably the method of simulating the deployment of an implantable medical device further comprises: receiving a user input specifying the target deployment position. In particular, the user is able to select a position within the three- dimensional model of the patient’s vascular structure. This may be a point at which the distal point of the device is positioned. In other examples it may be a line at which the centreline is to be deployed, or a volume or section of the vascular structure. Preferably the method of simulating the deployment of an implantable medical device further comprises: receiving a user input specifying a target deployment direction. This may specify the orientation of the centreline, and / or the position and retraction direction of the catheter.

[0036] In some examples, receiving a user input specifying a target deployment direction comprises: receiving a user input specifying an approach vessel within the patient’s vascular structure along which a catheter is to be inserted to deploy the implantable medical device; determining the path of the catheter approach based on determined forces applied to the catheter by walls of the patient’s vascular structure, thereby determining the orientation of the catheter when at the deployment position. In this way the initial position and / or orientation of the centreline may be determined based on the intended route to insert the catheter.

[0037] Preferably the method of simulating the deployment of an implantable medical device further comprises: calculating the length of the implantable medical device along the centreline, based on the size of one or more two-dimensional shapes representing the cross-section of the IMD. Preferably the method involves using the size of the cross-sectional shapes at each node to determine the length. Since the length of the device at a particular orientation is dependent on the radial expansion of the device, the device length at a given configuration may be determined based on the current state of radial expansion. The length may be calculated at one or more iterations of the method, for example at the particular configuration of the device at the given iteration. The length may be calculated at each iteration of the method for maximum precision or at periodic iterations of the method or only at a final iteration of the method for a reduced computational requirement. In some examples the method comprises: calculating the length of the implantable medical device along the centreline, for a given orientation, based on the size of one or more two-dimensional shapes representing the cross-section of the device at the given orientation.

[0038] Preferably the length of the implantable medical device is calculated using a device model describing the relationship between the radial expansion and a length of the device for a specific implantable medical device. The device model may define the length of the device based on a given radial expansion of the device. It may take as input an average radial expansion across the nodes or more preferably the radial expansion at each node. The device model may be specific to a specific model and size of IMD.

[0039] Preferably the method comprises, iteratively, during expansion of the implantable medical device: expanding each two-dimensional shape; and calculating the length of the implantable medical device based on the radial expansion of the two- dimensional shape at the current stage of expansion. The length may be calculated at each iteration or periodically, for example every N iterations of the simulation.

[0040] Preferably the implantable medical device comprises a plurality of braided wires and the method comprises calculating the height (referred to interchangeably as the “length”) of the device using a braiding model, the braiding model describing the device height as a function of the diameter of the device and one or more parameters of the braided wires of the implantable medical device. There are many known braiding models defining the dimensions of a particular type of device based on the parameters of the braided wires. The braiding model represents the physical wires of an implantable medical device, as it would actually be administered to a patient. The braiding model is preferably an abstraction of the complexity of the real wires, which allows the simulation to be more computationally efficient, compared to simulating the full dynamics of all individual wires. Having the length be a function of the diameter of the device and one or more parameters of the braided wires of the implantable medical device captures the essential dynamics of the wires, and makes the simulation more accurately reflect a physical implantable medical device. The complexity of the braiding model can be tailored to computational resources available, and to balance the time required to run the simulation with the accuracy of the simulation required by the surgeon.

[0041] Preferably the parameters of the braided wires comprise one or more of: the number of wires, the length of each wire, the braiding angle, picks-per-inch (PPI) and the interstrut distance between wire crossings. In certain preferable examples the method uses the number of wires and length of each wire. In this sense, the braiding model better simulates the braiding of a real implantable medical device, by considering each wire. Considering each wire as a parameter in the braiding model improves the accuracy of the simulation. The full dynamics of each individual wire does not need to be simulated, but can be, in some embodiments, if the computational resources available allow for it. Preferably the number of nodes of the centreline is less than 50, preferably between 5 and 30. This range balances computational efficiency of the simulation with its accuracy.

[0042] Preferably the implantable medical device comprises an intrasaccular device. Intrasaccular devices are implantable medical devices that get implanted into saccular structures, such as aneurysms. The present method provides significant improvements over existing methods of simulating the deployment of intrasaccular devices within an aneurysm.

[0043] Preferably the implantable medical device is a flow diverter. In this way, the method may be used to simulate the deployed position of a flow diverter within a vessel, for example in a vessel including an aneurysm.

[0044] Where the implantable medical device comprises a flow diverter the method preferably comprises: determining a vessel centreline of a vessel of the patient’s vascular structure within the three-dimensional model, along which the flow diverter is to be deployed; positioning the centreline of the implantable medical device along a segment of the vessel centreline within the patient’s vascular structure.

[0045] In a further aspect of the invention there is provided a computer implemented method of selecting a best-fit implantable medical for deployment at a target deployment position within a patient’s vascular structure, the method comprising: receiving user-selection of a first candidate implantable medical device and a second candidate implantable medical device; performing simulation of the deployed position of the first candidate implantable medical device and second candidate implantable medical device according to any method defined above or in the appended claims; selecting a best-fit implantable medical device based on the simulations.

[0046] The method preferably further comprises calculating a suitability metric for each of the first candidate implantable medical device and the second candidate implantable medical device based on the simulations; outputting an indication of a best-fit implantable medical device based on the suitability metric. Preferably the suitability metric comprises a measure of the fit of the deployed implantable medical device within the patient’s vascular structure. Preferably the suitability metric comprises a measure of the apposition between the implantable medical device in the deployed configuration and a wall of the vascular structure.

[0047] In a further aspect of the invention there is provided a computer program comprising instructions that when executed by a processor cause the processor to perform the method of any of the appended claims.

[0048] In a further aspect of the invention there is provided a computer-readable storage medium comprising instructions that when executed by a processor cause the processor to perform the method of any of the appended claims.

[0049] In a further aspect of the invention there is provided a system comprising one or more processors configured to perform the method of any of the appended claims.

[0050] BRIEF DESCRIPTION OF DRAWINGS

[0051] Embodiments of the invention are now described, by way of example, with reference to the drawings, in which:

[0052] Figures 1 A-1 D illustrate a method of simulating the deployment of an implantable medical device (IMD) at a target position within a patient’s vascular structure;

[0053] Figure 2 shows the method used to simulate the deployment of IM Ds;

[0054] Figure 3 shows a physical embodiment of the IMD, that would be administered to a patient, with the centreline, nodes, two-dimensional cross-sectional shapes, and radial direction, all shown;

[0055] Figures 4A-4F show how a user may select the initial target position and / or direction of the IMD;

[0056] Figure 5 shows how a user can select the approach vessel for the catheter; Figure 6A shows an intrasaccular device, which is one particular embodiment of the IMD, deployed at a deployment point in an aneurysm dome, from a blood vessel;

[0057] Figure 6B shows the intrasaccular device beginning to expand inside of the aneurysm dome, while the centreline of the intrasaccular device begins to shrink due to the braiding model being applied;

[0058] Figure 6C shows the intrasaccular device in a final position within the aneurysm dome, where the convergence condition has been reached;

[0059] Figure 7 shows the implantable medical device, being simulated during deployment from a catheter at the deployment point, in one embodiment of the invention;

[0060] Figures 8A-8F, show how the deployment of the intrasaccular device shown in Figures 6A-6C looks to a user of the simulation, as the intrasaccular device is released from the catheter of Figure 7;

[0061] Figure 9A shows a side-on view of another embodiment of the implantable medical device, a flow diverter, which is initially deployed in a blood vessel with an aneurysm dome;

[0062] Figure 9B shows the flow diverter when it has reached a final position inside of the blood vessel.

[0063] DETAILED DESCRIPTION

[0064] Figures 1A, 1 B and 1C, and 1 D illustrate a method of simulating the deployment of an IMD 2 at a target position 3 within a patient’s vascular structure 19. As will be described, the method involves receiving a three-dimensional model of the patient’s vascular structure 19, representing the implantable medical device 2 (IMD) as a centreline 60 of the IMD 2 where the centreline 60 comprises a plurality of nodes 200-204 positioned along the centreline 60. Each node 200-204, together representing a centreline 60 of the IMD 2, is associated with a two- dimensional shape 100-104 representing the cross-section of the IM D 2 at the position of the node 200-204.

[0065] The simulation method involves positioning the centreline 60 at the target deployment position 3, which may be a target area, volume or, as in the case of Figure 1A, a point at which the distal end of the device (i.e. a distal end node) is positioned before being deployed. The simulation proceeds by iteratively (i) calculating a force on each node 200-204 if the corresponding two-dimensional shape 100-104 contacts or intersects a wall 4 of the vascular structure 19 within the three-dimensional model, (ii) determining a movement of the node 200-204 based on the calculated force and (iii) updating the position of the node 200-204 based on the determined movement. In this way, the simulation involves iteratively, at each time step of the simulation, calculating forces based on interaction between the two-dimensional shapes 100-104 and the vessel wall 4 based on the orientation of the nodes 200-204 and shapes 100-104 at the configuration at that time step, determining the motion of a node 200-204 based on the calculated forces and updating the position of the nodes 200-204. The simulation proceeds until a convergence condition is met. This may be a predetermined number of iterations or when a determined parameter associated with one or more nodes meets a threshold condition, for example when a calculated kinematic variable falls below a threshold. In some examples this may be the net force on the plurality of nodes. In the preferred examples described below the convergence condition is met when there is no further motion of the nodes 200-204, or the motion (i.e. the calculated displacement or velocity) is below a threshold. The deployed position of the IMD 2 is determined based on the position of the nodes 200-204 when this convergence condition is met.

[0066] Figure 3 illustrates the representation of an IMD 2, in this case an intrasaccular device, usable within the present simulation method. Figure 3 shows the full outer surface of the device comprising a plurality of braided wires, forming a cage-like structure. While illustrated in Figure 3, this full outer wire structure is not directly included within the representation used in the present method. Instead the device is represented by a plurality of nodes 230-239 along a centreline 63 of the device, where the centreline extends between a distal point 21 and a proximal point 23. Each node 230-239 is associated with a two-dimensional shape 130-139 representing the cross-section of the device at that point. In this case, the device is a broadly cylindrical device, such that each cross-sectional shape comprises a circle, but the shape can vary to model devices with arbitrary cross-sectional shape. As will be described, the size of the cross-section can be varied to set the size of the device within the simulation, and simulate the expansion of the device and interaction with a vessel wall. Contact forces may be calculated depending on the interaction between the cross-sectional shape and vessel wall 4, with the net force on the node calculated to determine movements of the node during the simulation of the deployment.

[0067] According to the present method, the simulation of the deployment of the IMD 2 is simplified to determining the movement of the nodes 200-204 of the centreline 60, based on forces calculated by determining the relative position of the outer surface of the device, represented by the two-dimensional shapes 100-104, and the vessel wall 4. This method provides accurate determination of the deployed configuration with significantly reduced computational complexity compared to prior art methods.

[0068] There are a number of possible implementations of this method, which have differing advantages for different applications and requirements. In one important implementation, the two-dimensional shapes 100-104 are expanded within the vascular structure 19 from an initial configuration towards a maximal expansion of the IMD 2, thereby simulating the expansion of the IMD 2 within the vascular structure 19. In particular, the two-dimensional shapes 100-104 have an initial size and the size of the two dimensional shapes 100-104 is expanded as the iterations of the simulation progress to represent the expansion of the IMD 2. The forces applied to the IMD 2 related to the interaction with the vessel wall 4 are calculated when the expanding shapes 100-104 contact (touch, intersect or overlap with) the vessel wall 4. This implementation is illustrated in Figures 1A - 1 D and described in more detail below. In an alternative implementation, the expansion of the IMD 2 may not be directly simulated and instead only the maximum radial expansion of a particular IM D 2 configuration is used to determine the forces applied. This can be modelled by placing the two-dimensional shapes 100-104, representing the maximum or optimum expansion, at the nodes 200-204 to represent the expanded IMD 2 immediately within the vessel structure 19 at a target deployment position. In this case the forces may be calculated based on the intersection (or equivalently “overlap”) of the shapes 100-104 with the vessel wall 4. This alternative is described in more detail below, particularly in reference to Figures 9A and 9B. The two-dimensional shape 100-104 representing the maximal or optimum expansion may be different for each node 200-204.

[0069] Figures 1 A - 1 D illustrate the simulation of an expanding implantable medical IMD 2 to be deployed at a target deployment point 3, and simulated as a series of expanding circles 100-104 about nodes 200-204, representing the centreline 60 of the IMD 2.

[0070] Figure 1A shows the IMD 2 in an initial configuration, representing the initial deployment position of the IMD 2. As will be described in more detail below, the initial deployment of the IMD 2 can be simulated in a number of ways. In particular the initial maximum size of the two-dimensional shapes 100-104 may be set to simulate the deployment from a catheter of the IMD 2, for example this size may initially be limited to simulate the containment within a catheter and then increased to allow expansion, in one example progressively increasing the maximum diameter of the nodes 200-204 in sequence along the centreline 60 to simulate the retraction of a catheter. In the example of Figure 1A, the two dimensional shapes 100-104 are initiated at a uniform initial size, with the centreline 60 positioned aligned along an elongate axis 5 of the vessel 19 and slightly off-axis towards the vessel wall 4.

[0071] The simulation may use certain set parameters to determine the conditions for the simulation. Preferably, certain initial conditions may be set by a user. In particular, a key advantage of the method is that it allows clinicians to readily simulate different device parameters and deployment parameters, by running the simulation with different selections of such parameters, thereby determining the best approach for a surgery. The initial conditions include:

[0072] - Model of the device: these may include devices of varying geometries, for example: the WEB™-SL device (MicroVention Inc., CA, US), which has a cylindrical geometry; the WEB™-SLS device (MicroVention Inc., CA, US), which has an elliptical geometry; and the Contour™ (Stryker, Ml, US) device, which has a hemispherical geometry. These devices can be classified into being intravascular devices, or intrasaccular devices, depending on whether they are intended to be positioned within a vessel, or a saccular structure (such as an aneurysm). Many models of device have a circular cross-sectional shape, and as such a circle is used as the two-dimensional shape associated with each node (at least until contact with a vessel wall when, in some examples, the cross-sectional shape may deviate from the unconstrained cross-sectional shape). However, the method may use any two-dimensional shape to simulate a cross-section with any arbitrary shape.

[0073] - Dimensions of the device: the dimensions of the IMD 2 include: the radial extension of the device in the radial direction 14, discussed later, which may also be called the “width” of the device; and the length of the IMD 2, equivalent to the length of the centreline 60, which is equal to the path length of the centreline between the top node 200 and the bottom node 204. It is important for a clinician to choose the right dimensions for an IMD 2 because IM Ds that are too large (in length and / or width) may exert excessive forces against fragile, diseased tissue of the vessel wall 4, risking further complications. If the device is too small (in length and / or width) the device may not work effectively in helping the patient’s condition. By simulating the IMD 2 according to the current method, a clinician can trial many different IMD 2 to ensure they find a best fit.

[0074] Target position / direction: Figures 4A-4F show how a user may select the initial target position and / or direction of the IMD 2. In this embodiment, the IMD 2 is being deployed in an aneurysm dome 15 at a target deployment position 3. An additional target direction vector 5 has also been specified. In this way, the topmost node 200 will be positioned at the deployment point 3, and the centreline 60 will originate along the target deployment direction vector. That is, the centreline 60 will be parallel to the direction vector 5, and share common points. Figure 4E and Figure 4F show the IMD 2 aligned with the target deployment position 3 and the target direction vector 5.

[0075] - Approach vessel for catheter: Figure 5 shows how, in some examples of the invention, a user can select the approach vessel 17 for the catheter. Once the approach vessel 17 has been specified, along with a target deployment point 3 in the aneurysm dome 15, an algorithm can be applied to determine the approach path of the catheter. The approach of the catheter can then be simulated, which better reflects the real-world scenario of deploying IM Ds 2.

[0076] - An option to simulate catheter exit dynamics: the user may decide whether or not they would like the blooming effect, discussed later, of the IMD 2 from the catheter to be simulated. Having the blooming effect better reflects the real-world scenario of deploying IM Ds 2, but may not always be necessary for a surgeon, who may wish to solely focus on a later stage of the deployment process.

[0077] - An option to simulate full-dynamics or final deployment: in certain scenarios, the surgeon may just wish to focus on the final deployment of the IMD 2. In this case, the surgeon can run the simulation such that the IMD 2 is fully expanded at the deployment point 3 from the start of the simulation. Then, the simulation runs to adjust the position of the centreline 60.

[0078] In the example of Figure 1A, a cylindrical IMD 2 with a circular cross section is being simulated leading to a choice of circles 100-104 as the two-dimensional shapes 100-104. The device has been deployed, in a vessel 19, from an initial deployment position 3 along an initial direction vector 5, which has been depicted away from the centreline 60 for clarity, but which would be aligned with the centreline 60. The IMD 2 has been deployed off-centre in the vessel 19, which has been depicted, in a simplified view, having a gradual tapering of diameter along its length. The vascular structure could be part of a blood vessel or an aneurysm dome. The initial deployment of the IMD 2 in Figure lAcould represent a device being held at a certain width inside of a catheter of the same size. Dynamics of the release for the catheter have been opted, by a user, to not be simulated, resulting in all of the circles 100-104 being free to expand uniformly.

[0079] During iterations of the simulation the circles 100-104, representing the crosssection of the device, are progressively expanded. For examples the two- dimensional shape (in this case circles) 100-104 may be expanded by a predetermined radial increment in each iteration (time step) of the simulation. At each time step it may be determined whether there is contact or overlap between each circle 100-104 and the vessel wall 4. This is determined by casting rays, at a range of angles, from each node in the radial direction 12 at each iteration to continuously measure the distance between the node and the vessel wall. When the radial size of the cross-sectional shape is greater or equal to the distance to the vessel wall, a contact is determined.

[0080] Figure 1 B shows a subsequent iteration in which the expanding circles 100-104 of the IMD 2 in Figure 1 A have expanded during iterations of the simulation, and how the topmost circle 100 is in contact with the vessel wall 4 at a contact point 16. As described above, the method involves calculating a contact force when a two dimensional shape (in this case one of the expanding circles 100-104) contacts the vessel wall 4. The contact forces are calculated at each point on each two-dimensional shape that contacts the vessel wall 4. These contact forces may then be summed to determine the net force on the corresponding node 200- 204. In the case of Figure 1 B the uppermost circle 100 contacts the vessel wall 4 at a single contact point, resulting in the net force on the node 200 being equal to the contact force determined at the contact point 16 on the two-dimensional shape 100. Figure 1 C shows the I M D 2 of Figure 1 A and Figure 1 B with the position of topmost node 200 on the centreline 60 updated, as a result of the topmost circle 100 making contact, at a contact point 16, with the vessel wall 4. In particular a displacement of the node 200 is calculated, based on the net force on the node 200, and the position of the node 200 is then updated, such that it moves further towards the centre of the vessel 19 to allow the uppermost circle 100 to continue to expand.

[0081] In the model, the centreline 60, nodes 200-204 and associated two-dimensional shapes 100-104 may be defined in a number of different ways, for example using a chosen coordinate system. Preferably the two dimensional shapes 100-104 may be defined using polar coordinates. In this case, the normal vector of the plane containing the surface of each expanding circle 100-104 is aligned with an infinitesimal section of the centreline 60 at the corresponding node 200-204. The expanding circles 100-104 are then defined in terms of polar coordinates on their respective plane, with the origin being the respective node 200-204 corresponding to a particular expanding circle 100-104. A point on the perimeter of one particular expanding circle 100-104 is defined in terms of its distance r from the corresponding node 200-204 along a radial axis 12, and its angular distance 0 from an angular axis 14, and is, therefore, defined by a unique point (r, 0) in the plane containing the one particular expanding circle 100-104. The distance r may be dependent on the angular distance 0. The skilled person would readily understand that in the case of IM Ds with a circular cross-section, represented by expanding circles 100-104 r is constant with respect to 0, however, in the case of a general two-dimensional shape, it could vary to define devices with different cross-sectional shapes, or to simulate deformation of the surface of the device under contact with the vessel wall 4.

[0082] Once the initial conditions have been specified, the simulation begins to expand the circles 100-104 by increasing the r value for each point (r, 0) on the perimeter of each expanding circle 100-104 by some Ar at each time step of the iteration, such that the new position of the point is (r+Ar, 0). Ar may be the same for all values of 0, or might vary based on the value of 0. Ar may be different for each expanding circle 100-104. Using the definition of polar coordinates, the term “expand” means to set the coordinate of one or more points (r, 0) on the perimeter of one particular expanding circle 100-104 to (r+Ar, 0). In some cases, points may be expanded straight from (r, 0) to (Rmax, 0). Rmax may depend on 0. Rmaxmay be different for each expanding circles 100-104. After the expansion of an expanding circle 100-104, an updated position of the corresponding node 200-204 is calculated based on the net force. For example, in Figure 1 B, the expanding circle 100 makes contact with the vessel wall 4 at a contact point 16.

[0083] A contact with the vessel wall 4 may be determined using ray-casting, where rays 50 are cast at a subset of angles with a constant angular spacing, with respect to the angular axis 14, from each node 200-204. The distance the ray travels along in the radial axis 12, from the nodes 200-204 before intersecting with the vessel wall 4, is then measured for each given angle 0 and subtracted from the value of r on the perimeter of the expanding circles 100-104 at that same angle 0. This gives a measure of the distance between each point on the perimeter of the expanding circles 100-104, and the vessel wall 4. Ray casting may be performed repeatedly during the simulation, for example at every iteration, to maintain an accurate measure of the relevant dimensions of the vascular structure 19 as the device moves. Where this distance is sufficiently close to zero (that is, below a threshold value), there is a contact point. In the case where contact point 16 has been determined, in Figure 1 B, a contact force 18 is calculated and applied to node 200. In the absence of other forces or contact points, this contact force 18 would be equal to the net force acting on the node 200, and result in a velocity v according to Equation 1

[0084] Equation 1 where eta (q) represents a time stepping parameter that scales how much velocity is produced per unit of net force, and f is the net force acting on the node 200. In other embodiments, the net force will be a sum of the contact forces 18 from multiple contact points, in addition to other forces such as: frictional forces which prevent the amount of motion the IMD 2 can undergo once it has made contact with the vessel wall 4; pushing forces which better represent real-life deployment of the IMD 2 from a device such as a catheter; and boundary forces which restores points that are moved out of the vessel wall 4, during expansion, towards the interior of the vessel wall 4. In some embodiments, where the device is expanded straight to Rmax, points may start outside of the vessel wall 4. The boundary force at each point (r, 0) is proportional to the difference between r and the radial distance from the centreline 60 to the vessel wall 4 at angle 0, determined by raycasting. The skilled person would understand that other forces could also be simulated and included in the expression for net forces at a node 200-204. Velocity of a node 200-204 is the time derivative of the displacement u so, using Equation 1 , the displacement of a node 200-204 can be written in the form shown in Equation 2 below.

[0085] Equation 2

[0086] A forward Euler scheme can then be used to calculate an updated position

[0087] Equation 3 to the first-order. A higher order expansion may be used, in some embodiments. To evaluate a convergence condition, utotai is evaluated, which is the sum of the term u for all nodes 200-204. Equally, this can be expressed as

[0088] Equation 4 where, Xcurrent describes the position of the node after it has been updated, and xprev describes the previous position of the node, before it was updated. Each is a vector quantity. The convergence condition is described in Equation 5 evaluates whether utotai is less than some value £disP.

[0089] Equation 5

[0090] In other embodiments, the specific form of Equations 1 -5 may differ, as long as a calculation of the motion of the device may be determined. In some embodiments, Equation 3 may be expressed in terms of velocity.

[0091] Figures 1 A-1 D illustrate the method steps of the simulation being applied. These method steps are shown in the flowchart of Figure 2. Firstly, in step 300, a three- dimensional model of the patient’s vascular structure 19 is received. The three- dimensional model is received from a separate process and is constructed from medical imaging of the patient. In other embodiments, the three-dimensional model may be received from other sources. Next, in step 302, the implantable medical device is represented as a centreline 60 comprising a plurality of nodes 200-204 where each node is associated with a two-dimensional shape 100-104 representing the cross section of the device at the position of the node. The centreline 60 is then positioned at the target deployment position 3 within the patient’s vascular structure, in step 304. Steps 306, 308, 310 are then iteratively applied to the simulation of the implantable medical device 2. Aforce is calculated on each node if the corresponding two-dimensional shape 100-104 contacts or intersects a wall 4 of the vascular structure 19 within the three-dimensional model. The intersection or contacting with the wall 4 is described previously. The movement 20 of the nodes 200-204 based on the force is then determined, and the position of the node subsequently updated based on this movement 20. Finally, after N iterations of steps 306, 308, and 310, step 312 determines the deployed position of the IM D 2 based on the position of the nodes 200-204 of the centreline 60 when a convergence condition is met, for example when the movement of the nodes 200-204 is below a threshold. The threshold, in some embodiments, is the convergence condition of Equation 5. In some embodiments, the centreline 60 may converge to a shape other than a straight line, for example, in Figure 9A and Figure 9B.

[0092] Figure 1 D shows a final position of the IM D 2 where the movement of the nodes is below a threshold, and the threshold is the convergence condition of Equation 5. In this case, the expanding circles 100-104 are contacting the vessel wall 4 at all points around their perimeter, meaning that there are many contact points exerting contact forces on nodes 200-204. Because there are contact forces (which in the case shown are of equal magnitude) all around the perimeter of each expanding circle 100-104, the contact forces at the nodes 200-204 cancel, meaning that a net force is only produced from other forces such as: frictional forces, pushing forces, and boundary forces. These other forces, in this case, are small enough that the sum of the displacement of each node 200-204 is less than Edisp. Therefore, the position of the centreline 60 is said to have converged, and the final orientation of the IMD 2 is determined. In alternative embodiments (not shown), the expanding circles 100-104 may reach a maximum expansion in the radial direction 12, Rmax, before being able to contact, or intersect, the vessel wall. In this case, all points on the perimeter of each expanding circle 100-104 respectively are defined by (Rmax, 0), meaning that r can increase no further. The device is too small to contact the vessel wall 4, or extend past it, and so there are no contact or boundary forces, and the convergence condition is reached, as the net force is zero.

[0093] The maximum expansion is considered for each expanding circle 100-104 individually, and the expanding circles 100-104 do not all have to be expanded at the same rate, meaning that one particular expanding circle 100-104 may stop expanding, while the others continue to do so. In embodiments where the expanding circles 100-104 are instead general two-dimensional shape, maximum expansions may be considered when one point on the perimeter of the general two-dimensional shape reaches (Rmax, 0). In other embodiments, Rmax depends on 0 and so maximum expansion is reached, fora general two-dimensional shape, when Rmax is reached for all 0. In some embodiments Rmax cannot be reached because the device is contained within the boundaries of the vessel wall 4, and is not allowed to extend past it. In this case, the IM D 2 is in contact with the wall 4 on all sides, and the net force will be zero. This means that the movement of the nodes 200-204 will be below the threshold value, and the simulation will stop. In some embodiments, the IMD 2 is modelled, with the size of the expanding circles 100-104 initialised immediately at Rmax, such that just the final orientation is determined. In this case, the IMD 2 will start fully expanded, and the contact forces applied will be proportional to the amount by which the expanding circles 100-104 extend past the vessel wall 4 in the radial direction 12.

[0094] The above method describes how the position of the nodes 200-204 of the centreline 60 can be determined based on the interaction of the cross-section with the vessel wall 15 as the device expands. The radial expansion of the device may also influence the length of the device and this can further be incorporated into the simulation to accurately determine the final length of the deployed device. In particular, for many expandable implantable medical devices the length of the device has a defined relationship with the radial expansion, and so the length of the device in the simulation can be updated as the device expands. For example, many examples of IM Ds 2, such as the one shown in Figure 3, comprise a plurality of braided wires forming the outer surface of the device, which can stretch and bend during deployment. As the radius of the device, represented by the circles 130-139, expands, the wires cause the length of the centreline 63 to reduce. The effect of the expanding crosssection of the device on the length of the device, is included in the simulation of the IMD 2 using a ‘device model, which relates the expansion of the expanding circles 130-139 along the radial axis 12, with the length of the device along centreline 63. The device model may be a “braiding model”, that relates the expansion of the device to the length of the device by considering parameters of the wires for a specific device model. The length of the centreline 63 may be updated at each time step of the iteration, or periodically after a certain number of timesteps. One example of a suitable braiding model is described in Simulation of intra-saccular devices for pre-operative device size selection: Method and validation for sizing and porosity simulation, Munoz. R et al (2022), Computers in Biology and Medicine Volume 147. Figure 6A shows an intrasaccular device 7, which is one particular example of an IMD 2, deployed at a deployment point 3 inside of an aneurysm dome 15, from a blood vessel 17. The intrasaccular device 7 has four expanding circles 110-113. Figure 6B shows the intrasaccular device 7 beginning to expand inside of the aneurysm dome 15, while the centreline 61 of the intrasaccular device 7 begins to shrink due to the braiding model being applied. Figure 6C shows the intrasaccular device 7 in a final position within the aneurysm dome 15, where the convergence condition has been reached.

[0095] The geometry of the expanding circles 110-113 relative to the intrasaccular device 7 is defined in exactly the same way as the geometry of the expanding circles 100- 104 relative to the centreline 60 of the IMD 2, with the same radial axis 12 and angular axis 14. The centreline 61 of the intrasaccular device 7 behaves identically to the centreline 60 of the IMD 2. The intrasaccular device 7 starts at a deployment point 3, and with a pushing force 32 to simulate the pushing of the device out of a catheter. As a result, the intrasaccular device 7 moves up inside of the aneurysm dome 15, while also expanding. During this process, the braiding model is also applied, reducing the height of the centreline 61. By the time the intrasaccular device 7 reaches its final position in Figure 6C, the length of the centreline 61 has sufficiently reduced such that the IMD 7 fits fully inside of the aneurysm dome 15. Similarly to the embodiment shown in Figure 1 D, the expanding circles 110-113 are in contact with the wall of the aneurysm dome 15 in a number of places around their respective circumferences, such that the net force from all of the contact points is equal to zero. In this case, the movement of the nodes 210-213 is below the threshold value, and the centreline 61 is in its final position.

[0096] A further extension of the method involves the calculation of the final deployed radius and, optionally, the final deployed cross-sectional shape at each node 210- 213. As described above, the cross-section of the device is modelled as a two- dimensional shape 110-113 at each node 210-213. At each node 210-213, the two dimensional shape 110-113 attempts to expand to a corresponding maximum unconstrained radius of the device Rmaxat the position of the node. Rmaxmay be constant across nodes 210-213 or may vary along the length of the centreline 61 , with a different maximum radius at one or more nodes 210-213. The cross- sectional shape of the (unconstrained) device at each node 210-213 may also be constant along the centreline 61 or it may vary. In many cases this unconstrained shape is circular, but devices with other cross-sectional shapes may also be modelled by appropriate selection of the two-dimensional shape 110-113. In many examples, the final deployed radius of the device will be limited by the vascular structure to a value less than Rmax. The cross-sectional shape of the device can also deviate from the unconstrained shape under contact from the vessel wall 15. The final deployed radius (at each node 210-213) and optionally shape (at each node 210-213) can be determined in a number of ways.

[0097] As described above, the proximity of the device to the vessel wall 15 is determined using the ray tracing algorithm, which casts rays 50 in all directions and measures the distance to the vessel or dome wall 15. This results in a contact surface, which can be defined by a general function of radius versus theta, describing any arbitrary cross-sectional shape of the vessel. In some examples the deployed radius of the device (i.e. the final radius of the cross-sectional shape) at each node 210-213 may simply be determined as the radius when the convergence condition is met. In particular, the vessel wall 15 and device may be considered as entirely rigid. The deployed radius of the device may then be determined based on the minimal contact distance, i.e. smallest radius value of the contact surface.

[0098] In other examples, the deployed radius may be determined in other ways to reflect a more physically accurate situation in which there is some elasticity to the device and / or contact surface. In a first example, the final deployed radius (at a particular node 210-213) may be determined as the average contact distance, determined by taking the mean of all radius values in the contact surface. In this example, the final deployed shape of the two- dimensional shape 110-113 is not altered but the radius is increased beyond the maximum radius given by a fully rigid device and vessel wall, to reflect some deformation to the contact surface. An alternative would be to determine the final radius to provide a two-dimensional shape 11Q- 113 having the same perimeter distance, or the same cross-sectional area, as the contact surface of the vessel. Again, the two dimensional shape 110-113 is not altered from the unconstrained shape of the device (for example a circle) but the radius is selected to give a perimeter or area that would be consistent with conformity of the device to the vessel wall 15. In both of these examples, the device would breach the vessel wall 15 in the simulation, but may reflect a more physically accurate deployment radius value, without requiring altering of the shape of the two-dimensional shape 110-113 representing the cross section.

[0099] In an alternative example, the shape of the two dimensional shapes 110-113 representing the device cross-section can be altered to model a conforming of the device to the contact surface of the vessel wall 15. This can prevent breaching of the vessel wall 15 by the cross-section of the device in the simulation. In this case, the contact surface itself can be used as the cross-sectional shape of the device in its maximally expanded state. Some degree of smoothing of the contact surface can be performed, for example using an averaging window, before it being assigned as the device cross-section.

[0100] The latter method can be expanded by simulating deformation of the device cross- sectional shape under contact with the vessel wall, prior to reaching the final deployed position. In particular, when a contact point of a two-dimensional shape 110-113 is determined, a deformation of the two-dimensional shape 110-113 may be calculated and taken into account when determining the net force and movement of the node 210-213, and the length of the device centreline 63. A contact force may be determined as described in the method above and a deformation of the cross-sectional shape may be determined based on the contact force, resulting in a change from the unconstrained cross-sectional shape. The net force on the node may then be adjusted accordingly and used to calculate the displacement of the node 210-213.

[0101] Figure 7 is a schematic illustration showing a cross section of the IMD 2, being simulated during deployment from a catheter 24 at the deployment point 3, in one embodiment of the invention. In other embodiments, the intrasaccular device 7, or another implantable medical device, may be used. The cross-section of the catheter 24 may vary to accommodate implantable medical devices that do not have a circular cross-section. During this deployment, the implantable medical device has a top section 28 and a bottom section 26. In the bottom section 26, the expanding circles (not shown) are being constrained by a maximum radial extension smaller than the maximum radial extension of the IMD 2. In other words, the expanding circles of the bottom section 26, expand to a value less than Rmax for the IMD 2. This value is suitably equal to the internal radius of the catheter, Rc. In this way, the expanding circles of the bottom section 26 are prevented from expanding past the boundaries of the catheter 24. The nodes (not shown), of the IMD 2, are given a pushing force, to simulate movement out of the catheter. The position of the nodes is updated, according to Equations 1-3. In other embodiments, different equations may be used. When the simulation detects that a node 200-204 from the bottom section 26 has moved past an opening in the catheter 32, the maximum radial expansion of the corresponding expanding circle (not shown) is set from Rcto Rmax. In this way, the device expands out of the catheter 24 as it would during real-life deployment.

[0102] Figures 8A-8F, show how the deployment of the intrasaccular device 7 shown in Figures 6A-6C looks to a user of the simulation, as the intrasaccular device 7 is released from the catheter of Figure 7. Visualisations have been extracted from a number of selected timesteps of the simulation. Starting from Figure 8A, the intrasaccular device 7 can be seen entering the aneurysm dome 15 through a blood vessel 17. It is an initial configuration that represents the confinement of the device within a catheter. Figures 8B-8D then show the blooming of the device from the catheter, with the expanding circles towards the top of the device expanding while the expanding circles towards the bottom of the device are still confined within the catheter. The remaining Figures show the expanding of the device, and the shrinking of the centreline, as the device fills the aneurysm dome 15. In Figures 8E and 8F, a bottom surface of the intrasaccular device 7 has expanded more than the upper portion, to create a flat base covering the blood vessel 17. Although a braiding model is used, instead of simulating the full dynamics of each individual wire, the user sees a rendered mesh over the outer surface of the intrasaccular device 7. Figure 9A shows a side-on view of another example of the present invention, applied to the deployment of another type of IM D 2, a flow diverter 46, which is initially deployed in a blood vessel 38 within an aneurysm dome 36. This illustrates the application of the method to a different situation, namely the deployment of a flow diverter 46 in a vessel 38 containing an aneurysm 36. Here the method can be applied to determine how the device centreline moves away from the centreline of the vessel to permit expansion of the device around the aneurysm. Figure 9B shows the flow diverter 46 when it has reached a final position inside of the blood vessel 38. The flow diverter 46 has a centreline 62 which is similar to the centreline 60 of the IMD 2. Nodes 220-228 are positioned on the centreline 62, and are similar to the nodes 200-204 of the IMD 2.

[0103] In one example, the circles 120 - 128 representing the cross-section may be initialised at a size representing the maximal expansion of the device. The maximal expansion may be the unconstrained cross-sectional dimension of the device or it may be the maximal expansion, based on the local vessel geometry. The maximal expansion value Rmax may therefore vary between nodes. For example, two dimensional shapes in the vicinity of the aneurysm, such as circles 122 - 125 may have a larger maximal expansion value to those circles 120, 121 , 127, 128 within the sections 34a and 34c of the vessel around the aneurysm. When the IMD is initially deployed at the target location, in some areas (such as circles 123 - 125 within the aneurysm_ the two dimensional shapes will initially overlap with the vessel wall, as shown in Figure 9A. Through the iterations of the method, the forces on these shapes may be calculated and the movement of the nodes determined based on these forces. In other words, the approximate final cross-sectional dimension of each two dimensional shape may be derived from the local vessel dimensions and therefore it may not be necessary to simulate the full deployment dynamics by simulating the expansion of the two-dimensional shapes. A less computationally intensive simulation may therefore be performed in which only the (approximate) final cross-sectional size, at maximal expansion of the device, may be used to calculate the forces on the device and move the nodes accordingly. As described below, the simulation may only be performed for sections of the vessel in which the shape departs from a regular tubular shape (i.e. section 34b). In sections 34a and 34c the forces may be assumed to be in balance with the position of the device roughly constant.

[0104] In other examples, to achieve maximum accuracy in the simulation, the full deployment of the IMD from a catheter may be simulated. In this case, the radial dimensions of the two dimensional shapes are initialised at an initial size, for example representing the dimensions when contained within a catheter. The size of the shapes are then increased through the iterations to simulate the deployment, as described above. In this example, the radial direction 12 and the angular direction 14, as shown in Figure 1 , are defined identically relative to the centreline 62 as they are to the centreline 60, and the geometry of expanding circles 120-128 is defined identically to the expanding circles 100-104 of Figure 1.

[0105] The nodes 220-228 are initially placed along the centreline of the vessel, which can be calculated via any one of several routine methods known in the art, for example as described in Antiga, L., & Remuzzi, A. (2002); Patient-Specific Modelling of Geometry and Blood Flow in Large Arteries. In this example of the method, applied to flow diverters, the simulation method may only be applied to certain sections of the vessel. In particular, the simulation may only be applied where the vessel departs from a regular tubular shape, for example in the vicinity of the aneurysm to determine the displacement of the centreline of the flow diverter within these sections. As shown in Figure 9A vessel sections 34a and 34c are determined as having a cross-sectional shape that is broadly tubular and therefore the method of the present invention is only applied in the central section 34b around the aneurysm 36. The sections can be determined by performing a geometrical analysis of the shape of the vessel, for example calculating a parameter defining how much the vessel 38 differs from a regular cylindrical shape.

[0106] In sections 34a and 34c where the vessel has a regular tubular shape, there is no need to perform the method since the flow diverter 46 is arranged along the vessel centreline and will not move from this arrangement (i.e. the forces applied by the walls of the vessel will balance, meaning there is no movement of the centreline). In contrast, where the shape of the vessel departs from this tubular shape near the aneurysm the device centreline 62 will move to accommodate further expansion of the device.

[0107] In particular, the radius calculated for the expanding circles 122-126 overextends the vessel wall, as can be seen in Figure 9A. As a result of this overextending, a force is calculated which is proportional to the amount by which the expanding circles 122-126 overextend the vessel wall. Just as with any previous embodiment, this then acts as the net forces acting on each of the corresponding nodes 222-226, causing these nodes 222-226 to move away from the vessel wall. The position of the centreline 62 is then updated according to the new position of these nodes 222-226, until the net force acting on the nodes 222-226 is zero, and, as a result, their movement is below a threshold value. At this point, the centreline is in a final position, shown in Figure 9B. In some embodiments, the movement of the nodes 222-226 is determined to be below a threshold using Equation 5. In some embodiments, the net force may be generated from forces other than the force due to the overextension of the expanding circles 122-126. In some embodiments, the expanding circles 120-128 are generic two-dimensional shapes, and not circles.

[0108] Application of the Simulation

[0109] The method of simulating an IM D 2 described above can be used in isolation but is ideally integrated into a larger computer-implemented method for simulating IMDs 2. In particular, the method of simulating an IMD described herein may be run for a plurality of devices and the “best-fit” device selected. The selection may be made automatically by the software by calculating a suitability metric providing a measure of the fit of the device. An example of such a method of automatically selecting a best fit device based on a calculated metric is disclosed in EP20210218506. In this method, the three-dimensional model of the patient’s vascular structure is created from image data, received from medical scanners, and the deployed configuration of the IMD 2 is simulated to determine a suitability metric for each simulated IMD 2. The suitability metric provides a measure of the fit of the IMD in the patient’s vascular structure, which can be outputted to a surgeon in order for them to make a determination about which IMD 2 to administer to a patient. In this method the step of simulating the IMDs 2 deployed configuration can be performed according to the method of the present disclosure.

Claims

CLAIMS1. A computer implemented method of simulating the deployment of an implantable medical device at a target deployment position within a patient’s vascular structure, the method comprising: receiving a three-dimensional model of the patient’s vascular structure; representing the implantable medical device as a centreline of the implantable medical device, the centreline comprising a plurality of nodes, where each node is associated with a two-dimensional shape representing the crosssection of the device at the position of the node; positioning the centreline at the target deployment position within the patient’s vascular structure; iteratively (i) calculating a force on each node if the corresponding two- dimensional shape contacts or intersects a wall of the vascular structure within the three-dimensional model, (ii) determining a movement of the node based on the calculated force and (iii) updating the position of the node based on the determined movement; determining the deployed position of the implantable medical device based on the position of the nodes of the centreline when a convergence condition is met.

2. The computer implemented method of claim 1 wherein the convergence condition is met when the net force on the nodes or the movement of the nodes is below a threshold.

3. The computer implemented method of claim 1 or claim 2 wherein the method comprises, at each iteration, calculating a kinematic cost function representing the motion of the nodes and determining that the convergence condition is met when the calculated kinematic cost function reaches a convergence threshold.

4. The computer implemented method of claim 3 wherein the kinematic cost function is dependent on one or both of displacement and velocity of one or more nodes of the centreline.

5. The computer implemented method of claim 3 or claim 4 wherein the kinematic cost function comprises a sum or average of the magnitude of the displacement or velocity calculated across a plurality of nodes.

6. The computer implemented method of any preceding claim wherein the method comprises iteratively: expanding each two-dimensional shape from an initial configuration towards a maximal radial expansion of the device at the position of the node; calculating a force when a two-dimensional shape contacts or intersects a wall of the vascular structure; determining a movement of the corresponding node based on the calculated force; wherein the deployed position is determined when the movement of the nodes, based on the calculated forces, is below a threshold or the device has reached its maximal expansion.

7. The computer implemented method of claims 6 further comprising varying a maximum radial dimension of the two-dimensional shapes with time and / or position along the centreline, thereby simulating the unloading of the implantable medical device from a catheter.

8. The computer implemented method of any preceding claim wherein the method comprises iteratively expanding each two-dimensional shape from an initial configuration towards a maximal radial expansion of the device at the position of the node, the method further comprising: calculating the length of the implantable medical device along the centreline, based on the size of each two-dimensional shape representing the cross-section of the device.

9. The computer implemented method of claim 8 wherein the length of the implantable medical device is calculated using a device model describing the relationship between the radial expansion and a length of the device for a specific implantable medical device.

10. The computer implemented method of claim 8 or claim 9 wherein the length of the device is calculated at a plurality of iterations of the simulation or only when the convergence condition is met.

11. The computer-implemented method of any of claims 1 to 5, wherein the method comprises initialising the simulation with the size of each two- dimensional shape set according to a maximal expansion of the implantable medical device.

12. The computer-implemented method of claim 11 wherein the iterations of the method comprise, iteratively:(i) calculating a force on each node if the corresponding two-dimensional shape intersects a wall of the vascular structure within the three-dimensional model, (ii) determining a movement of the node based on the calculated force and (iii) updating the position of the node based on the determined movement.

13. The computer implemented method of any preceding claim, wherein determining a movement of a node comprises: determining a contact force at each position on each two-dimensional shape that contacts or intersects a wall of the vascular structure; determining a net force on each node by summing all contact forces applied to the corresponding two-dimensional shape; determining a movement of each node based on the determined net force.

14. The computer implemented method of claim 13, wherein in addition to determining a contact force, the method comprises determining one or more of the following additional forces: an elastic force that resists a change in length of the centreline; a bending force that limits the curvature of the centreline; frictional forces following contact between the two-dimensional shape and a wall of the vascular structure; where the method comprises including the above one or more additional forces when determining the net force.

15. The computer implemented method of any preceding claims, further comprising: casting rays radially from the centreline of the implantable medical device to measure the distance from the centreline of the implantable medical device to a wall of the vascular structure; and using the determined distance to determine when one of the two- dimensional shapes contacts or intersects a wall of the vascular structure.

16. The computer implemented method of any preceding claim wherein the implantable medical device comprises an intrasaccular device.

17. The computer implemented method of any of claims 1 to 15 wherein the implantable medical device comprises a flow diverter device and the method comprises: determining a vessel centreline of a vessel of the patient’s vascular structure within the three-dimensional model, along which the flow diverter is to be deployed; positioning the centreline of the implantable medical device along a segment of the vessel centreline within the patient’s vascular structure.

18. A computer program comprising instructions that when executed by a processor cause the processor to perform the method of any preceding claim.

Images