It is time for another blog post and with this I want to introduce Edward Fält, Vice President Business Development New Initiatives at Mentice. During his time at Mentice, Edward has had the opportunity to work with both core development activities, clinical interactions to define details of a product as well as marketing activities to appropriately position the use of simulation to maximize value for clients. In this blog Edward discusses the link between technology and medicine and the blog is intended to introduce the importance of disruptive technologies in medical simulation and more specifically full physics simulation with respect to endovascular procedures.
Göran Malmberg, CEO/President
Working at Mentice means living in the borderland between medicine and technology. As such, we are regularly exposed to the most recent and innovative techniques of both worlds, and the task of understanding and translating everyday medical language into the technical requirements needed for an optimal simulation experience can sometimes be a challenge.
Nevertheless, it is indeed most of the time a very rewarding task and the satisfaction you feel when you see the smile on a physician’s face as he or she starts using the system for the first time, only to be followed by a nervous grin a few seconds later as the simulated patient suddenly goes into cardiac arrest, definitely makes it all worthwhile.
This “translation” process is for natural reasons mostly a one-way street, going from medicine to technology, and therefore I thought it could be interesting to do it the other way around for a change, i.e. to attempt to explain the cornerstones of simulation technology to the general medical community. This will be the theme in my upcoming blog posts and we will start with the part of a high fidelity simulation that makes everything tick, the so-called full physics engine.
But before we start, let’s take a small step back and get acquainted with what endovascular simulation is all about. The concept of medical simulation originated in the wake of a vast expansion of flight simulation training in the 1960s, mainly driven by major advances in technology and computing possibilities. At first, the medical simulators were simple training mannequins used to practice cardiopulmonary resuscitation, but the field soon started growing into other areas of medicine. During the same time period, popularity for minimally invasive surgery gained ground, and in conjunction with this so did the need to train physicians on these new types of procedures. Mentice develops training simulators for this particular field and mainly on the endovascular side, i.e. for minimally invasive procedures that can be carried out through the patient’s blood vessels. New endovascular procedures are invented constantly, and there are various ways to train students and physicians on them, either through traditional mentoring, lectures, glass models, low-end simulators or high fidelity simulators like the ones we develop.
So, full physics simulation sounds nice enough but what does it really mean? Well, from a user perspective, the difference is most apparent when you experiment with different techniques or different types of clinical devices inside of the simulated patient. For example, if you use a coronary catheter with a shape that is poorly suited for a particular patient anatomy, you might encounter problems to engage the coronary ostium in the way you are used to. Or, if you engage it too forcefully you might cause a dissection of the arterial wall. This is different from less advanced simulators, where you are typically not able to move the catheter around freely to train on alternative cannulation techniques, or from glass models, in which you will never cause a dissection no matter how careless you are. In other words, a full-physics simulator will try to predict what will happen in the virtual patient as a result of your actions, regardless of which technique or device you choose to use.
From a technical perspective though, the main task of the physics engine is to solve the following problem: given the forces acting on a system, what is the motion of the system? In our case the “system” is the human body and the forces and motions we want to calculate are those between a catheter or other clinical devices on the one hand, and the interior vessel wall that is part of a patient’s vascular anatomy on the other. In other words, when we move a catheter with a certain shape through a blood vessel of a certain size and curvature, what forces would we expect the device and vessel wall to exhibit and how will that affect which way the catheter will pass through the anatomy? If the shape is adequate for the anatomy, it will be easy to engage the coronary ostium. If it is not, it might prove difficult. If the forces on the arterial wall are large, we might cause a dissection to occur, and so on.
For the physics engine to be able to solve this problem in the most realistic way possible, a few key components are needed. First, we need a good model of the vascular anatomy. Nowadays, with the significant advances of medical imaging, this is typically not a problem. Segmentation workstations are commonplace and accurate 3D data can be obtained from a range of modalities such as CT, MRI and even ultrasound. Modern scanners are also able to take high-resolution images of moving vessels such as the coronary arteries of a beating heart. Second, the physical properties of the clinical devices that are used have to be measured. This is something we typically do with each simulated device at Mentice. We measure actual lengths and sizes (which do not always correspond to the specified ones even in the tightly regulated world of medicine) and perform tests of the surface material, friction, stiffness, and for more complex devices a lot of other properties. This information is then fed into a model describing all the physical properties of the complete clinical device.
At this point we have at our disposal some very exact computer models of both what the vasculature and the clinical devices look like and how they can be expected to behave. What’s missing from the equation now is something that can fill in the “gap” between the two, or more precisely put, a way to predict how the two models will interact with one another. This is exactly the job of the full physics engine! At every second during the course of the simulation, it goes through millions of interactions between points along the clinical device and on the vessel surface, so-called collisions points, and calculates what forces are generated as a result of these interactions. The forces are then used to update the velocity with which different parts of the clinical device move and in doing so causes it both to alter its original shape (e.g. when going through a tortuous vessel) and to take a realistic route through the vasculature.
There is a very simple rule of thumb for these physics engines: the faster it can count, the more realistic the resulting predictions. Without sufficient speed, the simulation will appear slow and its predicted behaviors will appear unrealistic. That is the reason why only modern high-end computers are used with our simulators, and it’s also the reason why we use our own proprietary physics engine, designed to be extremely efficient for exactly this type of medical simulation, and capable of performing many millions of collision calculations per second. Everything to provide you, the end user, with the world’s most realistic endovascular simulator.
It’s amazing how fast things change. Only a few years ago a whole room of high-end computers would have been needed for this type of simulation. The development of parallel multi-core computing plays a large part in this. Computers with up to 16 processor cores are commonly commercially available today, and it’s been estimated that within five year’s time they will feature as many as 128 cores. As a result, endovascular simulation training gets more and more realistic by the day. Currently our simulators are able to use up to 32 cores, and can even simulate multiple clinical devices used in parallel and how they in turn will interact with each other, e.g. in the case of an EVAR procedure when the contralateral gate is being cannulated, or predict how the vessels themselves might react, such as when straightening out loops in the radial artery.
The model of the patient is also becoming better. Nowadays we have the possibility to create patient-specific virtual anatomies within a matter of minutes and with higher accuracy than ever before. Eric Topol in his recent book, “The Creative Destruction of Medicine”, dedicates an entire chapter to the advances of medical imaging and calls this development the “virtualization of the real world”. But as the models improve and get easier to import into the virtual world, won’t the full physics engine and its prediction of what results you can expect when working on these models need to improve as well? Long before the rise of the internet and multi-core computers, David Gelernter wrote in his book “Mirror Worlds” that in the future “You will look into a computer screen and see the real world”. But how real would such a world seem if it didn’t also behave in the way we expected? Surely, filling in the “gap” between all the new available information in the best way possible will become increasing important. This is really where the fundamental benefit of using full physics simulation lies; in contrast to other alternatives it will only continue to improve as technology progresses, and offer an ever more realistic and valuable training experience to medical students, residents, and physicians.