My name is Stu Kutazogl, and I will be presenting Developing Objective Computational Methods for Quantifying Ankle Arthritis Using Low-Dose Weight-Bearing CT. Here are our disclosures. So ankle osteoarthritis is typically post-traumatic and requires conventional radiographic findings to determine the severity of the disease. That severity then dictates the treatment. And the most commonly used classification system to classify stages of ankle osteoarthritis is Kellgren-Lawrence. And as I look through these images here, we can see that mild, moderate, and severe joint space narrowing is really what drives the classification system forward. Unfortunately, these images are done on subjective two-dimensional radiographs, and there's no clear demarcation between grades, making this quite a subjective measure of arthritis severity. The goal of our study was to develop a software system to assess ankle osteoarthritis using weight-bearing CT and hopefully lay a foundation for a more objective ankle osteoarthritis classification system. We hypothesized that we would find quantitative differences in two measurements between arthritic ankles and controls. So our study was a retrospective case control study. We took two patients with ankle osteoarthritis and one control patient. And we had two foot and ankle orthopedic surgeons grade those ankles using conventional radiographs on the Kellgren-Lawrence scale. We then took the weight-bearing CT imaging of those three patients and opened them up in the proprietary software, used bony landmarks to place the center of the ankle joint, as demonstrated by this yellow severe here. And the software then automatically takes a cubic volume of interest and places the center of the cube at the center of that ankle joint. Now this volume of interest allows us to analyze the imaging data within that cube and also contains four linear projections that allows us to look at the image intensity directly along those lines. Now here's what that volume of interest looks like superimposed on top of the weight-bearing CT image. And each of those four projections thus corresponds to these four graphical plots that we can see on the right side of the screen. So what kind of information can we get from these graphical plots? Well, you can see as we move from the tibia to the talus, you see the image intensity corresponding to cancels bone, then the subchondral bone as marked by those two peaks on the graph, and the joint space as the valley in between the two peaks. This then allows us to calculate the joint space width as the difference between those two peaks, as well as the contrast using this equation here, which gives us a measure of the ability to differentiate between two different image intensities. So here's just an example of the control ankle that we used. Kellgren-Lawrence grade zero, we can see a nice wide and joint space width in the weight-bearing CT images on the right hand side of the screen. And we can see in each of the four projections, this two peaks corresponding to the subcortical bone, as well as the joint space width represented by the valley in between. If we look at a mildly arthritic case, we can see some joint space narrowing in the posterior section of the joint as seen on the weight-bearing CT. And when we look at the graphical representations, we can see the difference in the narrowing here as the sharp valley in between, as well as just the narrow joint space width as marked by the shaded section. And then finally, we have an ankle with severe arthritis, complete joint space loss posteriorly as seen on the imaging. And we can see in projection three, that corresponds to a single peak without a valley in between on the graphical plot. Looking at the quantitative data as a whole, we can see that the average joint space width between control mild arthritis and severe arthritis ankles were different. We can also see the similar pattern with the contrast calculations, but however, due to the sample size, we can't draw any statistical conclusions from these results. So talking about limitations in our study, this was a retrospective study, and the small sample size in this presentation limits it to a simple technique introduction. As with any sort of reader study, we would ideally have an inter- or inter-observer reliability performed, and that's what we hope to do in the future. And then we also did not correlate these findings, quantitative or imaging, to the clinical presentation. But we believe that this methodology is a good step towards a more robust ankle arthritis assessment system. We're going to need to perform additional studies to assess the wide variety of radiographic presentations of arthritis, including subchondral sclerosis, any sort of Taylor tilt subluxations that are associated with arthritis. And finally, if we want to use this to improve clinical practice, to be able to correlate these findings to clinical outcomes. With that, thank you for your attention.

Stu Kutazogl

ankle arthritis

low-dose weight-bearing CT

objective classification system

joint space width