It's my pleasure to introduce our research speaker for the morning, Dr. Anderson from the University of Iowa. He serves as the Richard and Jan Johnson Chair in Orthopedic Biomechanics, as well as the Vice Chair of Research for the department. He has over 120 publications, multiple DOD grants, NIH grants, and he's gonna be challenging us today, this morning, to up our game in foot and ankle research. Please join me in giving a warm welcome, Dr. Don Anderson. Thank you. Holy cow, I've never had music at my introduction. Thank you very much. I'm looking forward to this presentation. I appreciate the invitation from the research committee and AOFAS to come and present some of our research committed at the University of Iowa, as long as some of our colleagues. As my introduction said, I'm really kind of challenging the community to kind of up our game in foot and ankle research. The good thing is there's going to be no test here and there won't be a lot of details. So hopefully you'll bear with me. Here's my disclosures. So a little bit of a roadmap. On the right, there's kind of a nice highway that takes you on a clear corner here. Unfortunately, more often it's something like this where you have to coordinate your way through a riddled path with cleared downed trees. But we're going to talk about some case studies, some research of some other people I brought along whose work I really respect and I hopefully then conclude. Just motivation-wise, I think it's fair to say that present-day orthopedic foot and ankle practice is built on over a century of research, much of it involving anatomical and cadaveric studies that are subject to historical limitations. And state-of-the-art mechanical testing, biology, imaging, modeling capabilities have really greatly expanded the rigor with which research can be done. And these new capabilities really afford opportunities to better understand the complex interplay of biology, mechanics in the foot and ankle, which can lead to evidence-based practice improvements. I love to start with imaging studies like this because it's just really amazing the amount of information that's in an image set like this, this 3D volumetric reconstruction. But we've often turned to the words of Lord Kelvin, who said back in 1883 that when you cannot measure what you are speaking about and express it in numbers, your knowledge is of a meager and unsatisfactory kind, whatever the matter may be. Admittedly, a little bit of a scientist's perspective. But let's work our way through this first case study. I'm going to speak a little bit about the pathomechanics of ankle OA, some work we've done in this area over the last 20 years. And you all well know that ankle osteoarthritis is generally rare, but it's a very high incidence after fractures such as those of the tibial pilum, which are by definition post-traumatic OA. They can occur early, as soon as two years after injury. The etiology is unclear, but there's clearly mechanical factors implicated. But these factors are really pretty difficult to quantify. There's a disabling major economic burden, of course, the predisposing joint injuries we mentioned. And I'm just going to talk today about two pathomechanical risk factors, the intensity of the acute fracture severity, and the chronic contact stress elevation left after this joint is put back together again. And we'll talk a little bit about how that might influence orthopedic treatment. So surgeons have long known that fractures occur over a severity, and we often hear people speak of low energy or high energy fractures. In fact, some of the subjective fracture classification systems try to indicate fracture severity, but they suffer from poor inter-observer reliability. Of course, the great thing is there's a whole branch of science called fracture mechanics, and I'm an engineer by training. But here we have a structure where energy is stored in the structure until the load is exceeded, and then you have failure. And the energy is liberated in this fracturing process, and fracture energy, the quantitative objective measure, is proportional to the fracture-liberated surface area in brittle solids such as bone under the high rates of loading associated with fracture. We've spent 20 years coming up with ways in which to use preoperative CT data routinely acquired for these studies to quantify the fracture energy. Here's an interesting slide. I just want to introduce a basic idea. These are 20 patients whose data I've colored. Two years after their intra-articular fracture, they either did not develop OA, their green columns if they didn't, yellow if they were kind of on the borderline, and red they clearly developed OA. We asked a series of three surgeons to rank order these from the least severe to the most severe by looking at the preoperative CT scans. They agreed pretty well. There's about a 72% concordance between them. And here's one individual's rankings. You can see it's more or less green to the left, their lower severity ones, and red to the right, the higher severity ones. But there's some notable outliers here. I just want to show you then when we use fracture severity in contrast, now the y-axis is no longer an arbitrary number. It's the fracture energy-based measure. And here you can see there's an 88% concordance with whether somebody developed OA or not and a perfect agreement with whether they had a KL grade of two or higher in their fracture energy. And there appears to be a threshold here, and that threshold may well be a point of intervention. So we've taken this methodology outside of the ankle and used it in a variety of other joints. And it's kind of fascinating to think people have long said that different joints respond to fracture, inter-articular fractures differently, and their risk of developing post-traumatic OA vary accordingly. Well, it turns out if you take this fracture energy metric, which I've here normalized to the local area because you have an impact over an area, and sometimes the contact area is smaller, other times it's larger, that's the x-axis. And I'm showing here five different parts of the body, five different articular joints. And then the historical PTOA rates, and you get this really high correlation. This is wonderful work of Kevin Dibbern, a PhD student in my lab who's since moved on. So those are all based on PTOA being classified by the KL grade. I think we could all agree that a grading scheme from 1957 may be a little long in the tooth, but it's still routinely used. I will say this about the KL grade. It's stamina. It's withstood the test of time. But it's a subjective grading scheme that's strictly categorical. It's weight-bearing radiographs, but they're 2D in nature. They're projection of a 3D structure, and they poorly capture early changes. And they're not as useful in joints of the foot and ankle because it's just hard to visualize those articular surfaces. So this is an area where we've explored weight-bearing CT, which is an emerging means to analyze the 3D joint structure under a loaded weight-bearing configuration. And the 3D joint space width, a key element of the traditional KL grade, is able to be characterized with this. And just as a quick aside, of course, there's also interesting information about the bone. And we just recently, Cesar De Cesar Neto and our group just recently published a paper in FAO showing how you may well be able to characterize changes in the bone density in the vicinity of that surface as well. So turning back to the 3D joint space width measure, we did a study here of, I'm showing you, of 40 patients. This was originally funded by the Orthopedic Trauma Association and then the Arthritis Foundation. We evaluate changes in the tibial tailor joint using weight-bearing CT after tibial pillon fracture. 18-month follow-up time period at six-month intervals, we have weight-bearing CT. And I'll just draw your attention, of course, here's a series of radiographs of the same ankle that there's a CT slice on the bottom. And if you look at those radiographs, it's awfully hard to see any kind of incongruity. But clearly on that CT scan, there's a substantial posterior step-off. So we use these measures, the intact contralateral provides a basis to assume what the joint space width was prior to the injury. Then if we kind of overlay them, we can develop these maps of the joint space width, the distance between those bones. And we can look at how that changes over time using something like the talus as a fiducial. So here we have four different groups among these cases. These were the first 26 cases here. And I'm just showing you that some of them see pretty substantial change in joint space width. This at the 12-month, that'd be the far right for you. But some of the others really look indistinguishable from the intact. And that's consistent with our experience that kind of different people have a more higher or lower likelihood of developing post-traumatic OA. And this is just some data to show that the fractured ankles averaged greater joint space narrowing, but it really varied among the fractured ankles. And this is then taking that fracture severity idea I introduced to you earlier and looking at the joint space narrowing at 12 months. And we have this correlation here between the fracture severity metric and the joint space narrowing. So now I'm going to shift gears slightly and talk a little bit about what we all understand. The reason people try to reduce these fractures is there's chronic contact stress elevation afterwards. And we're trying to minimize that to increase the likelihood that they don't develop post-traumatic OA. Well, patient-specific computational modeling is here to help. It enables prospective evaluation of surgically obtained fracture reductions. Here again, I've got CT volumetric renderings of those same 20 ankles after the surgeon put them together. And they did not do a perfect job in all those cases. But the modeling we do allows us to look at this is a simplified methodology called discrete element analysis. With the bone and the cartilage identified, we're able to look in a weight-bearing apposition and see what kind of contact stress develops. And this simplification allows us to be able to do quick measures and to simulate loading across an entire gait cycle. So we're not just looking at a maximum contact stress at one loading point. We're actually looking at how the habitual loading accumulates. In fact, just take it a step further. We all know that contact stress is not on its own harmful, that there are probably reasonable levels of contact stress that are healthy for cartilage. But when we exceed a damage threshold, here we've put it together as a summation over the gait cycle. And we look at the time at which those elevated contact stresses are exposed. We can come up with what we call a contact stress time overexposure. So this is a per gait cycle exposure of potentially harmful contact stress to the ankle joint. And when we take that on the x-axis here as an overexposed area, which percentage of the area has this elevated contact stress? And on your right, I guess it would be your left, you'd see that the intact contralaterals look relatively green, the low contact stress over time exposure, but the reds show up on the fractured cases. In fact, it appears to be a tolerance threshold that delineates between those that developed arthritis and those that did not. So that obviously offers an opportunity. If there are other ways beyond just doing a great job of putting this back together where we can reduce the contact stress exposure, we may be able to decrease the risk of PTOA. And in fact, we've developed methodologies then to be able to take a post-op CT scan all the way through to contact stress predictions. Of course, we all know that post-op CT scans are not exactly the standard of care. So one of the challenges in the post-op CT scan is actually the ankle is not in a loaded app position when it's scanned. So the modeling has to come up with a way to put it in a load-bearing pose. Well, enter a weight-bearing CT scan, of course, that automatically is in a weight-bearing pose and we're able to move much more quickly to contact stress predictions based on something like a six-month post-op weight-bearing CT scan, which might be comparable to roughly three radiographs that you might obtain at six months to see how the ankle is doing. So then here's a metric then of that overexposed area. We're still early in this series, but overexposed area versus changes in the joint space width over time. So moving forward, with mechanical concepts developed that help us to understand and predict PTOA risk and development, how might that information be used? Well, clearly it's useful in terms of if you're trying to do a clinical study, you kind of have a basis to say what risk each ankle had to develop post-traumatic away. But what about providing perhaps intraoperative feedback to surgeons or alternative treatment strategies? And I'll just touch on a couple of those ideas. This is a system we developed, courtesy of NIH funding, that we call a biomechanical guidance system. You're all familiar with intraoperative navigation systems. Well, this goes a step further and tries to give you an idea of the contact stress in the middle of a fracture reduction. So when you're trying to decide if you have done a good enough job, you're able to use this information to make that kind of a decision. And this is based on fluoroscopy. So if you're doing this in some kind of a limited approach for fluoroscopy, on the fly we can perhaps help you understand how good of a job you've done. And after developing that technology, here's a kind of an idea. This is an intermediate stage in the procedure, and based off of that fluoroscopy we're able to superimpose that 3D geometry in a nice aligned pose and tell you where those fracture fragments are at an intermediate stage. And that then allows you, here we did a proof of concept study to essentially the top row on your right shows ankles that were reduced using the guidance system. The bottom row shows you fractures that were reduced without the guidance system. I think you can get the appreciation that you may well be able to reduce the contact stress using an assist like this in the operating room. The second alternative might be to use some kind of a bracing strategy. And I tip my hat here to the military in particular, a good colleague, Jason Wilkin, who worked at the Center for the Intrepid developing these fancy custom carbon fiber braces. If you watch this video, one of the things that really strikes you is that struck behind the tibia is flexing. So it's storing energy, returning energy, so it actually bypasses some of the load at the ankle and reduces some of the moments at the ankle. So this may well be a strategy, but how would you predict what brace might work well or not? So we've developed a methodology here, OpenSim is a product that allows you to do musculoskeletal dynamic modeling based on gait lab data. And then you can take that all the way through to being able to introduce a brace, vary some of the properties of the brace, and see if you might well be able to reduce the contact stress using our discrete element analysis methodology. And just to show you, here's one case where we show on your far left would be no CDO, no brace. And on the far right, you can see under the right configuration with the right stiffness of the brace, you can actually reduce the contact stress. This is a prediction that needs to be validated in the laboratory, but we're very encouraged by this result. And it's fundamentally because the brace is able to reduce bare part of the moments across the ankle, thereby reducing the joint reaction force. So I'm going to shift gears slightly now and talk about some of the work Paragon has funded in our lab, looking at total ankle replacement. And this is really a computational modeling approach to look at press fit fixation design features. And you all know about uncemented fixation considerations. There's a variety of design features that are intended to provide early stability while supporting later in growth. We're going to be trying, obviously, to reduce micromotions so we can get bone to go across interfaces, since micromotions is a known contributor to implant loosening and influences survivorship. But we also need adequate load transfers to the surrounding bone to avoid either bone under or overloading. So some strategies might be to develop a model that looks at putting this implant in, whatever the implant might be here, a pegged design, without sidewalls, with sidewalls, with some kind of an interference fit here. We've modeled an interference fit both on the pegs as well at the edge of the implant. This is modern finite element analysis capabilities for bone are pretty impressive and formed off of CT scan data. So we can certainly model the stiffness of the bone appropriately to look at the CT scan data. But we can also use that to predict elastic plastic bone behavior, which means as you implant that pegged implant, you're going to cause some plastic irrecoverable deformation of the bone. And that rebound of that is going to be what holds that implant in. So we've modeled that off of the CT scan data. And this is just a way to show you. So on your left, you'll see the displacement of this implant. It goes in the stress that develops at the interface. And on the far right, anywhere you're seeing red, you're seeing high plastic strain. So you see bone deformation. And in fact, maybe you'll see if I can run that a second time, because you might appreciate watching rebound. So as we bottom out and we relieve the load, we're going to see a little rebound or recoil, as you might expect. So then what does this do? So we've leaned on some data from Hospital for Special Surgery, where they implanted and had a cadaveric robotic simulation to come up with these force and moment distributions across the gait cycle and implanted those. So again, I'd say a simple finite element model, I could apply a single load. I might get some information. If I don't include the plastic strain of the bone, I'm not really going to be able to represent very well how the bone holds the implant. And these are all important factors. But of the right considerations, hopefully you'll be able to see this on the video, you're far left, of course, without any sidewalls, without any interference fit. These implants are actually pretty unstable, right? Kind of makes sense. They really rely on that interference fit early. Introducing sidewalls helps some, but really the press fit is where you see the real benefit here. And just to show you that data, here's for the Apex implant. It's a one-pegged design. You can see how when I don't have sidewalls or press fit, there's pretty high micro motion actually early in the stance phase of gait. But as you start to introduce press fit and sidewalls, you can see that you greatly reduce. So whether you model this accurately or not is really going to change the result. In fact, this puts you in a position to be able to look at different implant design strategies. I will caution you. I'm showing here the Infinity and Salto, but we have not yet introduced the press fit in those models. But anyway, it's really interesting data to look at the ability of these implants to hold themselves in place. And here's a look at the micro motion. Again, now we've got the ability to look at the motion between that plate and the bone behind it throughout the gait cycle. And you see relatively high micro motions without any press fit. And as you move from left to right in the Apex and add more press fit, you see a great reduction in the amount of micro motion. And in fact, the far right, if you really want to look at the variation, you've got to zoom in quite a bit on a scale to have zero to 20 microns instead of zero to 40. So let me shift gears a little bit out of my work now. And I'm going to ask you to bear with me a little bit because some of this work you'd need to talk to the actual scientists who did it. But I just want to give you an idea of some work that's out there. And Amy Lenz, who I think is in the audience, was kind enough to share some of her really wonderful work she's doing at the University of Utah using statistical shape modeling. And I just want to say we've long known that there are measurements that can be made off of radiographs that tell you about changes in bone, changes in the relative position of bone. And traditionally, those measurements have been made off of radiographs. Maybe they've been made with calipers, maybe they've been made with some kind of software. But we now have capabilities using methods called statistical shape modeling to look at this in a different way. And here's a case. On the top row, you've got a healthy ankle. On the bottom, you've got one with ankle OA. And clearly, there are signs here that you could easily see. You'd see off of your radiographs to be able to help you. But measuring those is a little difficult. Here's one strategy. You might be able to use some kind of a tomographic approach where you try to make some measurements of compensated or uncompensated joints dealing with ankle osteoarthritis. Under the calcaneus and its relative position under the load. Or you might use DZR's radiographic measurements where you're looking at the degree of tailored tilt, the hindfoot alignment angles. These are the traditional methodologies. They're kind of prone a little bit to some user bias and they're kind of difficult measures sometimes to make. One statistical shape modeling is a methodology whereby you can describe in mathematical sense this shape of an object. This is an automated process here. You're looking at two different tali. These dots are finding these balls are finding their positions landmark wise on the joint in a way that's able to define subtle differences in the geometry. And those can be tracked over time and studied. In fact, those are looking at a single bone. You may want to look at the relative position of different bones as you have collapse of the foot and ankle. And Amy and her group have developed this capability to put the dots, the balls on different bones and look at the relative position between them. So if you will think about the shape of the whole structure and these articulated bones and how they deviate in their positions. So this then allows you to do things like look at a population, a statistical population of variation. This is fundamentally for those of you who don't know. This is how you do things like morph faces, morph structures into different population based metrics. And she's been able to show some of the early work has shown some of these differences between compensated non-compensated ankles looking at the range of variation within those groups. So for instance, then you're able to do things like identify and measure some of these variables that historically have been identified as important indicators of progress in treatment or advancement of disease. And I just want to say again, Amy, thank you very much for sharing these data with us. The last work I'm going to do is show you some stuff I had a good friend, Bill Ledoux, who works with Dr. St. George's and out at the University of Washington in the VA center there. They have a center for limb loss and mobility that over the last decade has really ramped up the great work that they had done there. And this is just one location and I just kind of, this is even a little faster run through their data than Amy's because I don't understand it as well, but they've concerted effort. VA funding has supported this NIH funding. They've been able to take work where they better understand through mechanical measurement of tissue properties, the properties of the diabetic plantar soft tissues. They're able to develop potentially non-invasive methodologies to be able to characterize using ultrasound the soft tissue properties and then developing mechanical systems to be able to look at 3D scans of plantar soft tissues in live subjects. And this is kind of really wonderful work that I think has potential to really change the way we understand those tissues and evaluate them. They've also done work with cadaveric gait simulation. They're not the only lab in the country that's done this or the world, but I've got to zoom in a little bit so you can see it of course, but there's somewhere in the middle of that structure, there's a poor little foot and ankle that's being taken through a range of motion that would simulate gait or other input wave forms. And this robotic gait simulator can take cadaveric feet through a walking cycle. And they can use that to compare different implants for treating different conditions and see how that changes the motion of that joint. And then of course, if you haven't seen the biplane video fluoroscopy or radiography methodologies that are out there now, you can now, using these video fluoroscopy techniques, quantify the motion of the bone as a person walks through a space here on a treadmill. And they've developed a biplane system out there that allows them to have people walk across the treadmill and they can track the motion of those bones. And again, that's kind of a really interesting way in this human setting. I think one of the things I hope you can understand is we've moved away from relying as heavily on cadaveric studies. And in fact, if you're looking for NIH or Department of Defense funding, it's harder and harder to get projects funded that involve cadaveric tissues. But fortunately, the technologies give us capabilities to study these in live subjects. So I'd like to thank the Arthritis Foundation, the AOFAS, and Paragon 28 for supporting the ankle OA think tank. And I think you'll be hearing more about this. You've already heard a little bit, but this is really an opportunity to marshal some research forces along with some clinical forces to look at some really lingering problems that haven't been studied as rigorously as, say, knee OA. I'd like to acknowledge my other funding sources. Thank you for your attention. I'd be happy to take any questions if that's allowed.

foot and ankle biomechanics

ankle osteoarthritis

fracture severity

chronic contact stress elevation

preoperative CT data analysis

total ankle replacements