Thanks for everyone for having me here. It's Provich and I'm here to talk about a serendipitous discovery perhaps of a new way to describe critical humeral bone loss. So you know the clinical problem that we know is that about a quarter of primary arthroscopic stabilizations fail, and we know that there's a dose-dependent response between recurrence and bipolar bone loss. Dr. Lin has told us a lot about the near-track lesion, so now we understand that it is in fact a spectrum, but we don't still know how to stratify those patients into different procedures. Do we do Bancart only? Do we do Bancart plus or omplissage or open Bancart, or do we go straight to a bony procedure for these on-track patients? We think there's really two simple approaches. You can go all the way to a machine learning algorithm and have a computer tell you what surgery to do. That we haven't figured out yet, although we are working on that in Pittsburgh. Or we can go back to the drawing board and pay a little more attention to the bipolar aspect of bone loss. We pay a lot of attention to the glenoid. We know everything we need to know about glenoid bone loss, but we never really think of humerus in the same way. Deep down as orthopedic surgeons, we understand that bone loss is bipolar, and therefore the humerus should matter just as much as the glenoid. Here's the problem with the track concept. This goes along with DTD as well. It's one-dimensional. It just measures effectively in the medial lateral dimension. Nobody ever talks about the vertical dimension of the humerus bone loss, and we wanted to look at that as potentially an independent predictor for instability. That was our hypothesis, and if you had patients who were apprehensive with the arm and adduction, you would understand that perhaps inferior glenoid, excuse me, humeral bone loss may actually matter, and those are perhaps the patients that have a big Hill-Sachs lesion, and that's exactly what we went out to do. So it was a retrospective cohort study with minimum two-year follow-up. These were on-track only patients. These were anterior instability only patients. They were patients that underwent only a Bankart repair without a remplissage, and we looked at whether inferior Hill-Sachs extension correlated with recurrence. How do we define inferior extension? We looked at sagittal MRI. We looked at whether these lesions extended below the equator of the humerus. And here are results about 24% recurrence. Again, matching literature, about a quarter of these patients fail. Sixty-nine percent of these patients were contact sport athletes, so this was a very young and very active cohort. We looked at their univariate variables, which demonstrated that some of our measures were perhaps quite predictive even on their own without controlling for confounders. However, once you feed this data into a multivariate analysis, you start to realize just how predictive inferior humeral bone loss extension may matter. So if we define the threshold at 90 degrees, that odds ratio for recurrent dislocation specifically can go as high as 4.8. If you were to plot this again on a heat map and you put lower edge angle, which is the inferior extension, that black line is the humeral equator against percent glenoid bone loss, this actually becomes a really good illustration of what bipolar bone loss could look like. There's this corner in the bottom right with patients who are having these huge heel-sax lesions who have a very, very high, about almost 100% risk of recurrent dislocation. On average, happening at about 1.7 years from primary surgery. So was this a serendipitous discovery of critical humeral bone loss? Maybe. This is really the only study that looked into this. 22% of our patients fell into this critical bone loss category. So in summary, we think that this is quite a simple screening method. Any heel-sax lesions that extend below the equator, just pay attention to them when you scroll through an MRI. This is extraordinarily clinically facile. It doesn't waste your time in the clinic. You don't have to crunch any numbers. Eyeball it. The next time you scroll through the sagittals, we do need to validate this and see if it could potentially be a stratifier for surgery. We are looking into machine learning. Quick plug for this team at Pittsburgh. Pay attention and stay tuned. Thanks very much.

humeral bone loss

arthroscopic stabilizations

recurrent dislocation

heel-sax lesions

surgery stratification