0:01 My name is Dario Carbonaro. I’m an assistant professor at Politecnico di Torino in the biomechanics team. So today I’m going to talk about medical devices and so we will see different approaches and different type of application and I will focus on self-expandable cardiovascular devices. So as you can see here these devices are adopted for minimally invasive procedure. So they are inserted into a catheter and they are delivered and they change in their design.

0:45 So they exist like really different designs in terms of these devices depending on the pathology and I will show you three example today like a transaic valve a femoral stand and a stand graft and why they are called self-expandable. They are called self expandable because they don’t need any additional tool. Some other devices need for example balloon inflation. These devices self expanded after being ex inserted into the delivering catheter and they have in common that they are manufactured by nitinol or nickel titanium and this alloy has several unique properties.

1:31 One of these is the shape memory effect. This means that when we deform below a certain temperature there are some plastic deformation and by heating up we can recover this deformation. I don’t know if you ever see the the trick with a spoon with a magician. This was a nitinal spoon. But this is not the only unique properties of nitinol. Another like a really interesting feature is the super elasticity.

2:06 So basically super elasticity is the capability of the device to sustain elastically like really high strain value up to 15 20% and so here we can see for example a unaxial tensel curve and and this capability is essential for this type of devices. So we can see here for example a stent which is inserted into like a really small catheter of like a couple of millm and by extracting it we we can recover its initial shape without any permanent deformation.

2:42 And some medical devices have like some mechanical characteristics which depends on the geometry and on the material properties and all these mechanical characteristics has an impact on the interaction between the device and the anatomy. So at the end by working like on the geometric feature by adopting computational design and optimization we can also like work on the clinical efficiency and safety of the treatment. So the first advantage of adopting like computation optimization is enabling to have like better devices in terms of treatment performance but I mean it’s medical devices go also through like a really long development phase in which we have like a really high cost and really high time so computational design also enable like to have an initial iteration on the design and on the objectives that we want to meet and accordingly it also enable us to obtain some reduction in term of cost and times associated to the development phase.

So some for example some certification agency FDA finally get like really interested into computational design and they they delivered some guidance for credibility but so we can start with the first example. It’s a self-expandable truss cathetic valve. These devices are adopted for the treatment of arctic stenosis. So the valve is delivered here into the aortic arch and these are usually composed by a frame which is made of nitinol which is laser cut and after its shape setting and here we can see some commercially available devices and you can see that the design changed in the years and this change in design is related like to the increase clinical efficiency that we need for these devices.

7:42 So we can adopt optimization to increase so the clinical efficiency and this was our idea in this study. And first of all we need to build like a biomechanics model because we have some mechanical quantities and we have other quantities which are of clinical interest. And at this point by simulating the implantation procedure we can for example analyze the pullout force which we can like relate to the risk for migration or we can study other quantities such as the stress rate into the vessel in order to study the risk for arctic wall tissue tearing.

13:29 And I mean once that we have like a biomechanics model, our idea was to build like a parametric frame of the of the of the valve and also like some idilized frame and once we have like our biomechanics model we can define an optimization problem and in this optimization problem we can put all the quantities of clinical interest. So we developed like a sampling scheme. We we we changed the geometry implanted like different geometries and at the end we build some surrogate models for the optimization of the frame.

So here for example we we modeled different type of pathology and below we can see like the simulation. So we we crimp the finite simulation we crimp the device and we expand it into the Arctic valve and for the parameterization we directly work like with the mesh without need for the geometry we use like the morphing feature of alire and we can see here like different morphing shape and each morphing shape is associated to a shape factor.

So by by working and by combining the shape factor we can have like a complete parameterization of the valve. So we can change each of the parameters. And here we can see some example for example of the final element simulation we which we study the force that the device exchange or the the stresses the contact pressure of of the device. And so at this point by changing like the geometrical features we we have like a a data set of simulation and from this data set of simulation we build our surrogate models.

And here for example you can see some response surface of our surrogate model. And so at this point we can go through like the optimization from our surrogate model. So our surrogate model basically defines an a relationship between like the geometrical feature of the valve and the objectives which are of clinical interest. So we can adopt for example multi-objective optimization which we have like a family of par optimal solution or we can also work on monobjective optimization by optimizing the the frame for a single anatomy as we can see below and and this open us also like the possibility to develop some optimization for a patient specific scenario.

13:51 This means that we are developing some patient specific devices made and tailored for each anatomy. So another application that I will show you are femoral stands. So the these stands are commonly adopted for the treatment of peripheral artery disease in order like to open up the vessel and prevent reclusion. And there are like a really different type of designs as we can see here and and it’s really small devices as you can see here in this image and the research activity that I’m presenting is part of the reset project and the objective is to develop like innovative unit cell of this this stand in order like to meet some prescribed mechanicalistics and optimize the the treatment efficiency of of these devices.

So I go fast here. So basically the the stand is can be considered like a latalatic structure. So a cylindrical latice structure which is made by the repetition of the unit cell. So our workflow it’s based on homogenization theory. In this case we model it as a homogeneous holo cylinder which is composed of a homogenized an anizotropic material. So in this case like the really the micro scale which we can consider here like the the stand unit cell can be considered as like the the stiffness tensor.

But so at this point we can we adopted some solid mechanics equation and we developed like the relationship between like the homogenized material and the mechanical performance. So at this point we can develop some unit cells with topology optimization in order to meet some mechanical characteristics. So we here we see like the stiffness or like the change in length of the stand. So and so in this slide I’m I’m showing like the general workflow that would be adopted.

So at this point when we have homogenization theory we can use inverse homogenization topology optimization in order like to create and to develop some unit cells which are which satisfy some prescribed characteristics in terms of mechanical or geometrical characteristics. And so at this point we can automatically generate some new unit cells and by repeating this unit cells in a cylindrical latis structure we can have our final design.

So the the project is still going on but we tested this this workflow in this case with three design scenarios. We we changed like the objectives and we work with like volume fraction stiffness and we also like considering some constraints on the manufacturing because these devices need to be manufactured by lesser cutting. And so by inserting input these different u design scenarios we obtained this three unit cells and this unit cell they satisfy the prescribed mechanical characteristics and you can see here I mean the the final design of the stand.

So once we have our unit cell at this point we we needed to reposition in order to have like a realistic design but so we work with adaptive meshes. So we we don’t have a smooth contour which can create some troubles in the laser cutting process. So we automatically like smooth the the contour and and we can repeat all our like pattern of of the stand design.

And this that you see here in blue is actually like the input of the laser cutting machine. So we don’t need to work with the cylindrical latis structure. We we can work just in the two dimension. So in this input we passed it to the to the laser cutting machine and we also performed some prototypes some electropolishing and and this is our first prototype. It’s like a really small one.

It’s 6 mm in diame diameter and I mean the project is still going on. So we we are working like on prototyping more designs. So we are working also some experimental testing and like to validate our framework with some experimental data. So the last example that I will we will show is atoric arctic stand graft. This this device is commonly adopted for the treatment of arctic anorism and it’s composed of a nitinal filament which is sutred into a graft and as you can see like in in the video is delivered to a catheter with a minimally invasive procedure and I mean as as we saw before we can optimize the geometry to increase the efficiency of the treatment.

But we can also work on the material and as you can see here on the right there are different type of nitenol cardiovascular devices and these are made of nitinol and are shape seted. This means that we we deform the device we apply some sha shape setting and to fix like desired shape. But something that it’s u it’s shown in literature that when you apply a heat treatment during the shape setting the mechanical properties of the material change and and this is usually quite a problem for the industry because they need to understand how it change.

14:19 But for our point of view, it’s it’s also like an interesting point because we we asked to ourself, can we really control the the mechanical properties of nitinol by just controlling temperature and time of the heat treatment. And so this is was the the basically answer that we want to the the question that we want to answer in this project. So to understand how we can adopt the heat treatment in order to change the mechanical properties of the material and accordingly also on the device.

14:54 So we took some specimen, we applied like different heat treatment with different type and time and temperature and we analyze 15 different samples. And at this point we we characterize mechanically this specimen and we input this mechanical properties into a finite element model. And so the fin element model we have like a pre-stress procedure and after like a a crimping procedure and then in the crimping procedure we analyze the radial force.

15:29 So each point that you can see here like in this in this box it’s represented by the radial force of one specimen. This means that we can analyze the radial force of the device by changing by time and temperature of the heat treatment. And like the first conclusion that we can can see here it’s that the radial force change quite a lot from 55 to 75. And the other interesting aspect that we developed some multilinear regression model in order to control the time and temperature of of the treatment.

16:03 So basically it’s like like a recipe. So we can change time and and temperature of the heat treatment and we can control the response of the device. So this is just a preliminary study. So but I hope that we have the possibility to to continue. So it’s u it’s time to to conclude. So in in my research groups in biomechanics we are working on different type of topics but from I think from this audience point of view there are some interesting topics which is for example like the topology optimization or generative design and we are working also metam material for biomedical application and also on presical planning like we are developing some simulation in order like to guide the clinician to to have some some planning of the treatment advice also like on digital twin and on bio resorbable material.

17:12 Okay so I I finished. Thank you. Thank you. Thank you. To learn more about the CDFM computational design symposium series, to see the archives of previous presentations, and to learn about future events, visit CDFAM.com.